id
large_string | submitter
large_string | authors
large_string | title
large_string | comments
large_string | journal-ref
large_string | doi
large_string | report-no
large_string | categories
large_string | license
large_string | abstract
large_string | versions
large list | update_date
timestamp[ms] | authors_parsed
large list | prompt
large_string | label
string | prob
float64 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2204.10779 | Xunguang Wang | Xunguang Wang, Yiqun Lin, Xiaomeng Li | CgAT: Center-Guided Adversarial Training for Deep Hashing-Based
Retrieval | null | null | null | null | cs.CV cs.AI cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep hashing has been extensively utilized in massive image retrieval because
of its efficiency and effectiveness. However, deep hashing models are
vulnerable to adversarial examples, making it essential to develop adversarial
defense methods for image retrieval. Existing solutions achieved limited
defense performance because of using weak adversarial samples for training and
lacking discriminative optimization objectives to learn robust features. In
this paper, we present a min-max based Center-guided Adversarial Training,
namely CgAT, to improve the robustness of deep hashing networks through worst
adversarial examples. Specifically, we first formulate the center code as a
semantically-discriminative representative of the input image content, which
preserves the semantic similarity with positive samples and dissimilarity with
negative examples. We prove that a mathematical formula can calculate the
center code immediately. After obtaining the center codes in each optimization
iteration of the deep hashing network, they are adopted to guide the
adversarial training process. On the one hand, CgAT generates the worst
adversarial examples as augmented data by maximizing the Hamming distance
between the hash codes of the adversarial examples and the center codes. On the
other hand, CgAT learns to mitigate the effects of adversarial samples by
minimizing the Hamming distance to the center codes. Extensive experiments on
the benchmark datasets demonstrate the effectiveness of our adversarial
training algorithm in defending against adversarial attacks for deep
hashing-based retrieval. Compared with the current state-of-the-art defense
method, we significantly improve the defense performance by an average of
18.61\%, 12.35\%, and 11.56\% on FLICKR-25K, NUS-WIDE, and MS-COCO,
respectively. The code is available at https://github.com/xunguangwang/CgAT.
| [
{
"version": "v1",
"created": "Mon, 18 Apr 2022 04:51:08 GMT"
},
{
"version": "v2",
"created": "Mon, 7 Nov 2022 03:01:23 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Nov 2022 02:58:21 GMT"
},
{
"version": "v4",
"created": "Fri, 16 Dec 2022 09:04:36 GMT"
},
{
"version": "v5",
"created": "Wed, 22 Mar 2023 14:57:41 GMT"
},
{
"version": "v6",
"created": "Sat, 25 Mar 2023 13:58:36 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Xunguang",
""
],
[
"Lin",
"Yiqun",
""
],
[
"Li",
"Xiaomeng",
""
]
] | TITLE: CgAT: Center-Guided Adversarial Training for Deep Hashing-Based
Retrieval
ABSTRACT: Deep hashing has been extensively utilized in massive image retrieval because
of its efficiency and effectiveness. However, deep hashing models are
vulnerable to adversarial examples, making it essential to develop adversarial
defense methods for image retrieval. Existing solutions achieved limited
defense performance because of using weak adversarial samples for training and
lacking discriminative optimization objectives to learn robust features. In
this paper, we present a min-max based Center-guided Adversarial Training,
namely CgAT, to improve the robustness of deep hashing networks through worst
adversarial examples. Specifically, we first formulate the center code as a
semantically-discriminative representative of the input image content, which
preserves the semantic similarity with positive samples and dissimilarity with
negative examples. We prove that a mathematical formula can calculate the
center code immediately. After obtaining the center codes in each optimization
iteration of the deep hashing network, they are adopted to guide the
adversarial training process. On the one hand, CgAT generates the worst
adversarial examples as augmented data by maximizing the Hamming distance
between the hash codes of the adversarial examples and the center codes. On the
other hand, CgAT learns to mitigate the effects of adversarial samples by
minimizing the Hamming distance to the center codes. Extensive experiments on
the benchmark datasets demonstrate the effectiveness of our adversarial
training algorithm in defending against adversarial attacks for deep
hashing-based retrieval. Compared with the current state-of-the-art defense
method, we significantly improve the defense performance by an average of
18.61\%, 12.35\%, and 11.56\% on FLICKR-25K, NUS-WIDE, and MS-COCO,
respectively. The code is available at https://github.com/xunguangwang/CgAT.
| no_new_dataset | 0.949435 |
2205.13673 | Nicolas Pr\"ollochs | Chiara Drolsbach, Nicolas Pr\"ollochs | Diffusion of Community Fact-Checked Misinformation on Twitter | Accepted at CSCW 23 | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The spread of misinformation on social media is a pressing societal problem
that platforms, policymakers, and researchers continue to grapple with. As a
countermeasure, recent works have proposed to employ non-expert fact-checkers
in the crowd to fact-check social media content. While experimental studies
suggest that crowds might be able to accurately assess the veracity of social
media content, an understanding of how crowd fact-checked (mis-)information
spreads is missing. In this work, we empirically analyze the spread of
misleading vs. not misleading community fact-checked posts on social media. For
this purpose, we employ a dataset of community-created fact-checks from
Twitter's Birdwatch pilot and map them to resharing cascades on Twitter.
Different from earlier studies analyzing the spread of misinformation listed on
third-party fact-checking websites (e.g., Snopes), we find that community
fact-checked misinformation is less viral. Specifically, misleading posts are
estimated to receive 36.62% fewer retweets than not misleading posts. A partial
explanation may lie in differences in the fact-checking targets: community
fact-checkers tend to fact-check posts from influential user accounts with many
followers, while expert fact-checks tend to target posts that are shared by
less influential users. We further find that there are significant differences
in virality across different sub-types of misinformation (e.g., factual errors,
missing context, manipulated media). Moreover, we conduct a user study to
assess the perceived reliability of (real-world) community-created fact-checks.
Here, we find that users, to a large extent, agree with community-created
fact-checks. Altogether, our findings offer insights into how misleading vs.
not misleading posts spread and highlight the crucial role of sample selection
when studying misinformation on social media.
| [
{
"version": "v1",
"created": "Thu, 26 May 2022 23:27:59 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Nov 2022 01:07:09 GMT"
},
{
"version": "v3",
"created": "Sun, 1 Jan 2023 23:24:14 GMT"
},
{
"version": "v4",
"created": "Fri, 24 Mar 2023 19:03:20 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Drolsbach",
"Chiara",
""
],
[
"Pröllochs",
"Nicolas",
""
]
] | TITLE: Diffusion of Community Fact-Checked Misinformation on Twitter
ABSTRACT: The spread of misinformation on social media is a pressing societal problem
that platforms, policymakers, and researchers continue to grapple with. As a
countermeasure, recent works have proposed to employ non-expert fact-checkers
in the crowd to fact-check social media content. While experimental studies
suggest that crowds might be able to accurately assess the veracity of social
media content, an understanding of how crowd fact-checked (mis-)information
spreads is missing. In this work, we empirically analyze the spread of
misleading vs. not misleading community fact-checked posts on social media. For
this purpose, we employ a dataset of community-created fact-checks from
Twitter's Birdwatch pilot and map them to resharing cascades on Twitter.
Different from earlier studies analyzing the spread of misinformation listed on
third-party fact-checking websites (e.g., Snopes), we find that community
fact-checked misinformation is less viral. Specifically, misleading posts are
estimated to receive 36.62% fewer retweets than not misleading posts. A partial
explanation may lie in differences in the fact-checking targets: community
fact-checkers tend to fact-check posts from influential user accounts with many
followers, while expert fact-checks tend to target posts that are shared by
less influential users. We further find that there are significant differences
in virality across different sub-types of misinformation (e.g., factual errors,
missing context, manipulated media). Moreover, we conduct a user study to
assess the perceived reliability of (real-world) community-created fact-checks.
Here, we find that users, to a large extent, agree with community-created
fact-checks. Altogether, our findings offer insights into how misleading vs.
not misleading posts spread and highlight the crucial role of sample selection
when studying misinformation on social media.
| no_new_dataset | 0.794225 |
2205.15531 | Hyeon Cho | Hyeon Cho, Junyong Choi, Geonwoo Baek, Wonjun Hwang | itKD: Interchange Transfer-based Knowledge Distillation for 3D Object
Detection | Accepted at CVPR 2023 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Point-cloud based 3D object detectors recently have achieved remarkable
progress. However, most studies are limited to the development of network
architectures for improving only their accuracy without consideration of the
computational efficiency. In this paper, we first propose an autoencoder-style
framework comprising channel-wise compression and decompression via interchange
transfer-based knowledge distillation. To learn the map-view feature of a
teacher network, the features from teacher and student networks are
independently passed through the shared autoencoder; here, we use a compressed
representation loss that binds the channel-wised compression knowledge from
both student and teacher networks as a kind of regularization. The decompressed
features are transferred in opposite directions to reduce the gap in the
interchange reconstructions. Lastly, we present an head attention loss to match
the 3D object detection information drawn by the multi-head self-attention
mechanism. Through extensive experiments, we verify that our method can train
the lightweight model that is well-aligned with the 3D point cloud detection
task and we demonstrate its superiority using the well-known public datasets;
e.g., Waymo and nuScenes.
| [
{
"version": "v1",
"created": "Tue, 31 May 2022 04:25:37 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 04:30:25 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Cho",
"Hyeon",
""
],
[
"Choi",
"Junyong",
""
],
[
"Baek",
"Geonwoo",
""
],
[
"Hwang",
"Wonjun",
""
]
] | TITLE: itKD: Interchange Transfer-based Knowledge Distillation for 3D Object
Detection
ABSTRACT: Point-cloud based 3D object detectors recently have achieved remarkable
progress. However, most studies are limited to the development of network
architectures for improving only their accuracy without consideration of the
computational efficiency. In this paper, we first propose an autoencoder-style
framework comprising channel-wise compression and decompression via interchange
transfer-based knowledge distillation. To learn the map-view feature of a
teacher network, the features from teacher and student networks are
independently passed through the shared autoencoder; here, we use a compressed
representation loss that binds the channel-wised compression knowledge from
both student and teacher networks as a kind of regularization. The decompressed
features are transferred in opposite directions to reduce the gap in the
interchange reconstructions. Lastly, we present an head attention loss to match
the 3D object detection information drawn by the multi-head self-attention
mechanism. Through extensive experiments, we verify that our method can train
the lightweight model that is well-aligned with the 3D point cloud detection
task and we demonstrate its superiority using the well-known public datasets;
e.g., Waymo and nuScenes.
| no_new_dataset | 0.942295 |
2206.05751 | Chengyang Ying | Chengyang Ying, You Qiaoben, Xinning Zhou, Hang Su, Wenbo Ding,
Jianyong Ai | Consistent Attack: Universal Adversarial Perturbation on Embodied Vision
Navigation | null | Pattern Recognition Letters (PRL), 2023 | 10.1016/j.patrec.2023.03.001 | null | cs.LG cs.CR cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Embodied agents in vision navigation coupled with deep neural networks have
attracted increasing attention. However, deep neural networks have been shown
vulnerable to malicious adversarial noises, which may potentially cause
catastrophic failures in Embodied Vision Navigation. Among different
adversarial noises, universal adversarial perturbations (UAP), i.e., a constant
image-agnostic perturbation applied on every input frame of the agent, play a
critical role in Embodied Vision Navigation since they are
computation-efficient and application-practical during the attack. However,
existing UAP methods ignore the system dynamics of Embodied Vision Navigation
and might be sub-optimal. In order to extend UAP to the sequential decision
setting, we formulate the disturbed environment under the universal noise
$\delta$, as a $\delta$-disturbed Markov Decision Process ($\delta$-MDP). Based
on the formulation, we analyze the properties of $\delta$-MDP and propose two
novel Consistent Attack methods, named Reward UAP and Trajectory UAP, for
attacking Embodied agents, which consider the dynamic of the MDP and calculate
universal noises by estimating the disturbed distribution and the disturbed Q
function. For various victim models, our Consistent Attack can cause a
significant drop in their performance in the PointGoal task in Habitat with
different datasets and different scenes. Extensive experimental results
indicate that there exist serious potential risks for applying Embodied Vision
Navigation methods to the real world.
| [
{
"version": "v1",
"created": "Sun, 12 Jun 2022 14:45:11 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 03:02:02 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Mar 2023 11:18:59 GMT"
},
{
"version": "v4",
"created": "Sat, 25 Mar 2023 08:43:16 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ying",
"Chengyang",
""
],
[
"Qiaoben",
"You",
""
],
[
"Zhou",
"Xinning",
""
],
[
"Su",
"Hang",
""
],
[
"Ding",
"Wenbo",
""
],
[
"Ai",
"Jianyong",
""
]
] | TITLE: Consistent Attack: Universal Adversarial Perturbation on Embodied Vision
Navigation
ABSTRACT: Embodied agents in vision navigation coupled with deep neural networks have
attracted increasing attention. However, deep neural networks have been shown
vulnerable to malicious adversarial noises, which may potentially cause
catastrophic failures in Embodied Vision Navigation. Among different
adversarial noises, universal adversarial perturbations (UAP), i.e., a constant
image-agnostic perturbation applied on every input frame of the agent, play a
critical role in Embodied Vision Navigation since they are
computation-efficient and application-practical during the attack. However,
existing UAP methods ignore the system dynamics of Embodied Vision Navigation
and might be sub-optimal. In order to extend UAP to the sequential decision
setting, we formulate the disturbed environment under the universal noise
$\delta$, as a $\delta$-disturbed Markov Decision Process ($\delta$-MDP). Based
on the formulation, we analyze the properties of $\delta$-MDP and propose two
novel Consistent Attack methods, named Reward UAP and Trajectory UAP, for
attacking Embodied agents, which consider the dynamic of the MDP and calculate
universal noises by estimating the disturbed distribution and the disturbed Q
function. For various victim models, our Consistent Attack can cause a
significant drop in their performance in the PointGoal task in Habitat with
different datasets and different scenes. Extensive experimental results
indicate that there exist serious potential risks for applying Embodied Vision
Navigation methods to the real world.
| no_new_dataset | 0.942188 |
2206.06561 | Kaituo Feng | Kaituo Feng, Changsheng Li, Ye Yuan, Guoren Wang | FreeKD: Free-direction Knowledge Distillation for Graph Neural Networks | Accepted to KDD 2022 | null | 10.1145/3534678.3539320 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge distillation (KD) has demonstrated its effectiveness to boost the
performance of graph neural networks (GNNs), where its goal is to distill
knowledge from a deeper teacher GNN into a shallower student GNN. However, it
is actually difficult to train a satisfactory teacher GNN due to the well-known
over-parametrized and over-smoothing issues, leading to invalid knowledge
transfer in practical applications. In this paper, we propose the first
Free-direction Knowledge Distillation framework via Reinforcement learning for
GNNs, called FreeKD, which is no longer required to provide a deeper
well-optimized teacher GNN. The core idea of our work is to collaboratively
build two shallower GNNs in an effort to exchange knowledge between them via
reinforcement learning in a hierarchical way. As we observe that one typical
GNN model often has better and worse performances at different nodes during
training, we devise a dynamic and free-direction knowledge transfer strategy
that consists of two levels of actions: 1) node-level action determines the
directions of knowledge transfer between the corresponding nodes of two
networks; and then 2) structure-level action determines which of the local
structures generated by the node-level actions to be propagated. In essence,
our FreeKD is a general and principled framework which can be naturally
compatible with GNNs of different architectures. Extensive experiments on five
benchmark datasets demonstrate our FreeKD outperforms two base GNNs in a large
margin, and shows its efficacy to various GNNs. More surprisingly, our FreeKD
has comparable or even better performance than traditional KD algorithms that
distill knowledge from a deeper and stronger teacher GNN.
| [
{
"version": "v1",
"created": "Tue, 14 Jun 2022 02:24:38 GMT"
},
{
"version": "v2",
"created": "Tue, 21 Jun 2022 03:10:14 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Jun 2022 03:58:39 GMT"
},
{
"version": "v4",
"created": "Mon, 27 Mar 2023 05:59:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Feng",
"Kaituo",
""
],
[
"Li",
"Changsheng",
""
],
[
"Yuan",
"Ye",
""
],
[
"Wang",
"Guoren",
""
]
] | TITLE: FreeKD: Free-direction Knowledge Distillation for Graph Neural Networks
ABSTRACT: Knowledge distillation (KD) has demonstrated its effectiveness to boost the
performance of graph neural networks (GNNs), where its goal is to distill
knowledge from a deeper teacher GNN into a shallower student GNN. However, it
is actually difficult to train a satisfactory teacher GNN due to the well-known
over-parametrized and over-smoothing issues, leading to invalid knowledge
transfer in practical applications. In this paper, we propose the first
Free-direction Knowledge Distillation framework via Reinforcement learning for
GNNs, called FreeKD, which is no longer required to provide a deeper
well-optimized teacher GNN. The core idea of our work is to collaboratively
build two shallower GNNs in an effort to exchange knowledge between them via
reinforcement learning in a hierarchical way. As we observe that one typical
GNN model often has better and worse performances at different nodes during
training, we devise a dynamic and free-direction knowledge transfer strategy
that consists of two levels of actions: 1) node-level action determines the
directions of knowledge transfer between the corresponding nodes of two
networks; and then 2) structure-level action determines which of the local
structures generated by the node-level actions to be propagated. In essence,
our FreeKD is a general and principled framework which can be naturally
compatible with GNNs of different architectures. Extensive experiments on five
benchmark datasets demonstrate our FreeKD outperforms two base GNNs in a large
margin, and shows its efficacy to various GNNs. More surprisingly, our FreeKD
has comparable or even better performance than traditional KD algorithms that
distill knowledge from a deeper and stronger teacher GNN.
| no_new_dataset | 0.947672 |
2206.07817 | Jennifer Volk | Jennifer Volk, George Tzimpragos, Alex Wynn, Evan Golden and Timothy
Sherwood | Low-Cost Superconducting Fan-Out with Cell I$_\text{C}$ Ranking | 12 pages, 20 figures, accepted at IEEE TAS | null | 10.1109/TASC.2023.3256797 | null | cs.ET | http://creativecommons.org/licenses/by/4.0/ | Superconductor electronics (SCE) promise computer systems with orders of
magnitude higher speeds and lower energy consumption than their complementary
metal-oxide semiconductor (CMOS) counterparts. At the same time, the
scalability and resource utilization of superconducting systems are major
concerns. Some of these concerns come from device-level challenges and the gap
between SCE and CMOS technology nodes, and others come from the way Josephson
Junctions (JJs) are used. Towards this end, we notice that a considerable
fraction of hardware resources are not involved in logic operations, but rather
are used for fan-out and buffering purposes. In this paper, we ask if there is
a way to reduce these overheads, propose the use of JJs at the cell boundaries
to increase the number of outputs that a single stage can drive, and establish
a set of rules to discretize critical currents in a way that is conducive to
this assignment. Finally, we explore the design trade-offs that the presented
approach opens up and demonstrate its promise through detailed analog
simulations and modeling analyses. Our experiments indicate that the introduced
method leads to a 48% savings in the JJ count for a tree with a fan-out of
1024, as well as an average of 43% of the JJ count for signal splitting and 32%
for clock splitting in ISCAS'85 benchmarks.
| [
{
"version": "v1",
"created": "Wed, 15 Jun 2022 21:08:16 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Oct 2022 19:47:27 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Oct 2022 07:29:24 GMT"
},
{
"version": "v4",
"created": "Mon, 27 Mar 2023 16:26:47 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Volk",
"Jennifer",
""
],
[
"Tzimpragos",
"George",
""
],
[
"Wynn",
"Alex",
""
],
[
"Golden",
"Evan",
""
],
[
"Sherwood",
"Timothy",
""
]
] | TITLE: Low-Cost Superconducting Fan-Out with Cell I$_\text{C}$ Ranking
ABSTRACT: Superconductor electronics (SCE) promise computer systems with orders of
magnitude higher speeds and lower energy consumption than their complementary
metal-oxide semiconductor (CMOS) counterparts. At the same time, the
scalability and resource utilization of superconducting systems are major
concerns. Some of these concerns come from device-level challenges and the gap
between SCE and CMOS technology nodes, and others come from the way Josephson
Junctions (JJs) are used. Towards this end, we notice that a considerable
fraction of hardware resources are not involved in logic operations, but rather
are used for fan-out and buffering purposes. In this paper, we ask if there is
a way to reduce these overheads, propose the use of JJs at the cell boundaries
to increase the number of outputs that a single stage can drive, and establish
a set of rules to discretize critical currents in a way that is conducive to
this assignment. Finally, we explore the design trade-offs that the presented
approach opens up and demonstrate its promise through detailed analog
simulations and modeling analyses. Our experiments indicate that the introduced
method leads to a 48% savings in the JJ count for a tree with a fan-out of
1024, as well as an average of 43% of the JJ count for signal splitting and 32%
for clock splitting in ISCAS'85 benchmarks.
| no_new_dataset | 0.945096 |
2206.15387 | John Nguyen | John Nguyen, Jianyu Wang, Kshitiz Malik, Maziar Sanjabi, Michael
Rabbat | Where to Begin? On the Impact of Pre-Training and Initialization in
Federated Learning | Accepted at ICLR | International Conference on Learning Representations 2023 | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | An oft-cited challenge of federated learning is the presence of
heterogeneity. \emph{Data heterogeneity} refers to the fact that data from
different clients may follow very different distributions. \emph{System
heterogeneity} refers to client devices having different system capabilities. A
considerable number of federated optimization methods address this challenge.
In the literature, empirical evaluations usually start federated training from
random initialization. However, in many practical applications of federated
learning, the server has access to proxy data for the training task that can be
used to pre-train a model before starting federated training. Using four
standard federated learning benchmark datasets, we empirically study the impact
of starting from a pre-trained model in federated learning. Unsurprisingly,
starting from a pre-trained model reduces the training time required to reach a
target error rate and enables the training of more accurate models (up to 40\%)
than is possible when starting from random initialization. Surprisingly, we
also find that starting federated learning from a pre-trained initialization
reduces the effect of both data and system heterogeneity. We recommend future
work proposing and evaluating federated optimization methods to evaluate the
performance when starting from random and pre-trained initializations. This
study raises several questions for further work on understanding the role of
heterogeneity in federated optimization. \footnote{Our code is available at:
\url{https://github.com/facebookresearch/where_to_begin}}
| [
{
"version": "v1",
"created": "Thu, 30 Jun 2022 16:18:21 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 01:44:58 GMT"
},
{
"version": "v3",
"created": "Fri, 24 Mar 2023 19:09:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Nguyen",
"John",
""
],
[
"Wang",
"Jianyu",
""
],
[
"Malik",
"Kshitiz",
""
],
[
"Sanjabi",
"Maziar",
""
],
[
"Rabbat",
"Michael",
""
]
] | TITLE: Where to Begin? On the Impact of Pre-Training and Initialization in
Federated Learning
ABSTRACT: An oft-cited challenge of federated learning is the presence of
heterogeneity. \emph{Data heterogeneity} refers to the fact that data from
different clients may follow very different distributions. \emph{System
heterogeneity} refers to client devices having different system capabilities. A
considerable number of federated optimization methods address this challenge.
In the literature, empirical evaluations usually start federated training from
random initialization. However, in many practical applications of federated
learning, the server has access to proxy data for the training task that can be
used to pre-train a model before starting federated training. Using four
standard federated learning benchmark datasets, we empirically study the impact
of starting from a pre-trained model in federated learning. Unsurprisingly,
starting from a pre-trained model reduces the training time required to reach a
target error rate and enables the training of more accurate models (up to 40\%)
than is possible when starting from random initialization. Surprisingly, we
also find that starting federated learning from a pre-trained initialization
reduces the effect of both data and system heterogeneity. We recommend future
work proposing and evaluating federated optimization methods to evaluate the
performance when starting from random and pre-trained initializations. This
study raises several questions for further work on understanding the role of
heterogeneity in federated optimization. \footnote{Our code is available at:
\url{https://github.com/facebookresearch/where_to_begin}}
| no_new_dataset | 0.950549 |
2207.01297 | Wenhao Wu | Wenhao Wu, Zhun Sun, Wanli Ouyang | Revisiting Classifier: Transferring Vision-Language Models for Video
Recognition | Accepted by AAAI-2023. Camera Ready Version | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transferring knowledge from task-agnostic pre-trained deep models for
downstream tasks is an important topic in computer vision research. Along with
the growth of computational capacity, we now have open-source vision-language
pre-trained models in large scales of the model architecture and amount of
data. In this study, we focus on transferring knowledge for video
classification tasks. Conventional methods randomly initialize the linear
classifier head for vision classification, but they leave the usage of the text
encoder for downstream visual recognition tasks undiscovered. In this paper, we
revise the role of the linear classifier and replace the classifier with the
different knowledge from pre-trained model. We utilize the well-pretrained
language model to generate good semantic target for efficient transferring
learning. The empirical study shows that our method improves both the
performance and the training speed of video classification, with a negligible
change in the model. Our simple yet effective tuning paradigm achieves
state-of-the-art performance and efficient training on various video
recognition scenarios, i.e., zero-shot, few-shot, general recognition. In
particular, our paradigm achieves the state-of-the-art accuracy of 87.8% on
Kinetics-400, and also surpasses previous methods by 20~50% absolute top-1
accuracy under zero-shot, few-shot settings on five popular video datasets.
Code and models can be found at https://github.com/whwu95/Text4Vis .
| [
{
"version": "v1",
"created": "Mon, 4 Jul 2022 10:00:47 GMT"
},
{
"version": "v2",
"created": "Sat, 6 Aug 2022 08:50:30 GMT"
},
{
"version": "v3",
"created": "Thu, 24 Nov 2022 17:16:45 GMT"
},
{
"version": "v4",
"created": "Sun, 26 Mar 2023 16:28:26 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wu",
"Wenhao",
""
],
[
"Sun",
"Zhun",
""
],
[
"Ouyang",
"Wanli",
""
]
] | TITLE: Revisiting Classifier: Transferring Vision-Language Models for Video
Recognition
ABSTRACT: Transferring knowledge from task-agnostic pre-trained deep models for
downstream tasks is an important topic in computer vision research. Along with
the growth of computational capacity, we now have open-source vision-language
pre-trained models in large scales of the model architecture and amount of
data. In this study, we focus on transferring knowledge for video
classification tasks. Conventional methods randomly initialize the linear
classifier head for vision classification, but they leave the usage of the text
encoder for downstream visual recognition tasks undiscovered. In this paper, we
revise the role of the linear classifier and replace the classifier with the
different knowledge from pre-trained model. We utilize the well-pretrained
language model to generate good semantic target for efficient transferring
learning. The empirical study shows that our method improves both the
performance and the training speed of video classification, with a negligible
change in the model. Our simple yet effective tuning paradigm achieves
state-of-the-art performance and efficient training on various video
recognition scenarios, i.e., zero-shot, few-shot, general recognition. In
particular, our paradigm achieves the state-of-the-art accuracy of 87.8% on
Kinetics-400, and also surpasses previous methods by 20~50% absolute top-1
accuracy under zero-shot, few-shot settings on five popular video datasets.
Code and models can be found at https://github.com/whwu95/Text4Vis .
| no_new_dataset | 0.950824 |
2207.01971 | Yan Zhao | Yan Zhao, Ruihai Wu, Zhehuan Chen, Yourong Zhang, Qingnan Fan, Kaichun
Mo, Hao Dong | DualAfford: Learning Collaborative Visual Affordance for Dual-gripper
Manipulation | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It is essential yet challenging for future home-assistant robots to
understand and manipulate diverse 3D objects in daily human environments.
Towards building scalable systems that can perform diverse manipulation tasks
over various 3D shapes, recent works have advocated and demonstrated promising
results learning visual actionable affordance, which labels every point over
the input 3D geometry with an action likelihood of accomplishing the downstream
task (e.g., pushing or picking-up). However, these works only studied
single-gripper manipulation tasks, yet many real-world tasks require two hands
to achieve collaboratively. In this work, we propose a novel learning
framework, DualAfford, to learn collaborative affordance for dual-gripper
manipulation tasks. The core design of the approach is to reduce the quadratic
problem for two grippers into two disentangled yet interconnected subtasks for
efficient learning. Using the large-scale PartNet-Mobility and ShapeNet
datasets, we set up four benchmark tasks for dual-gripper manipulation.
Experiments prove the effectiveness and superiority of our method over three
baselines.
| [
{
"version": "v1",
"created": "Tue, 5 Jul 2022 11:30:37 GMT"
},
{
"version": "v2",
"created": "Fri, 30 Sep 2022 11:36:26 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Oct 2022 04:20:44 GMT"
},
{
"version": "v4",
"created": "Mon, 27 Feb 2023 17:58:11 GMT"
},
{
"version": "v5",
"created": "Fri, 24 Mar 2023 02:50:08 GMT"
},
{
"version": "v6",
"created": "Mon, 27 Mar 2023 04:10:14 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhao",
"Yan",
""
],
[
"Wu",
"Ruihai",
""
],
[
"Chen",
"Zhehuan",
""
],
[
"Zhang",
"Yourong",
""
],
[
"Fan",
"Qingnan",
""
],
[
"Mo",
"Kaichun",
""
],
[
"Dong",
"Hao",
""
]
] | TITLE: DualAfford: Learning Collaborative Visual Affordance for Dual-gripper
Manipulation
ABSTRACT: It is essential yet challenging for future home-assistant robots to
understand and manipulate diverse 3D objects in daily human environments.
Towards building scalable systems that can perform diverse manipulation tasks
over various 3D shapes, recent works have advocated and demonstrated promising
results learning visual actionable affordance, which labels every point over
the input 3D geometry with an action likelihood of accomplishing the downstream
task (e.g., pushing or picking-up). However, these works only studied
single-gripper manipulation tasks, yet many real-world tasks require two hands
to achieve collaboratively. In this work, we propose a novel learning
framework, DualAfford, to learn collaborative affordance for dual-gripper
manipulation tasks. The core design of the approach is to reduce the quadratic
problem for two grippers into two disentangled yet interconnected subtasks for
efficient learning. Using the large-scale PartNet-Mobility and ShapeNet
datasets, we set up four benchmark tasks for dual-gripper manipulation.
Experiments prove the effectiveness and superiority of our method over three
baselines.
| no_new_dataset | 0.939913 |
2207.04183 | Haoxuan Che | Haoxuan Che and Haibo Jin and Hao Chen | Learning Robust Representation for Joint Grading of Ophthalmic Diseases
via Adaptive Curriculum and Feature Disentanglement | Accepted by MICCAI22 | null | 10.1007/978-3-031-16437-8_50 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diabetic retinopathy (DR) and diabetic macular edema (DME) are leading causes
of permanent blindness worldwide. Designing an automatic grading system with
good generalization ability for DR and DME is vital in clinical practice.
However, prior works either grade DR or DME independently, without considering
internal correlations between them, or grade them jointly by shared feature
representation, yet ignoring potential generalization issues caused by
difficult samples and data bias. Aiming to address these problems, we propose a
framework for joint grading with the dynamic difficulty-aware weighted loss
(DAW) and the dual-stream disentangled learning architecture (DETACH). Inspired
by curriculum learning, DAW learns from simple samples to difficult samples
dynamically via measuring difficulty adaptively. DETACH separates features of
grading tasks to avoid potential emphasis on the bias. With the addition of DAW
and DETACH, the model learns robust disentangled feature representations to
explore internal correlations between DR and DME and achieve better grading
performance. Experiments on three benchmarks show the effectiveness and
robustness of our framework under both the intra-dataset and cross-dataset
tests.
| [
{
"version": "v1",
"created": "Sat, 9 Jul 2022 03:02:36 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 18:05:48 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Che",
"Haoxuan",
""
],
[
"Jin",
"Haibo",
""
],
[
"Chen",
"Hao",
""
]
] | TITLE: Learning Robust Representation for Joint Grading of Ophthalmic Diseases
via Adaptive Curriculum and Feature Disentanglement
ABSTRACT: Diabetic retinopathy (DR) and diabetic macular edema (DME) are leading causes
of permanent blindness worldwide. Designing an automatic grading system with
good generalization ability for DR and DME is vital in clinical practice.
However, prior works either grade DR or DME independently, without considering
internal correlations between them, or grade them jointly by shared feature
representation, yet ignoring potential generalization issues caused by
difficult samples and data bias. Aiming to address these problems, we propose a
framework for joint grading with the dynamic difficulty-aware weighted loss
(DAW) and the dual-stream disentangled learning architecture (DETACH). Inspired
by curriculum learning, DAW learns from simple samples to difficult samples
dynamically via measuring difficulty adaptively. DETACH separates features of
grading tasks to avoid potential emphasis on the bias. With the addition of DAW
and DETACH, the model learns robust disentangled feature representations to
explore internal correlations between DR and DME and achieve better grading
performance. Experiments on three benchmarks show the effectiveness and
robustness of our framework under both the intra-dataset and cross-dataset
tests.
| no_new_dataset | 0.947721 |
2208.01647 | Vipul Mann | Vipul Mann and Venkat Venkatasubramanian | AI-driven Hypergraph Network of Organic Chemistry: Network Statistics
and Applications in Reaction Classification | null | null | 10.1039/D2RE00309K | null | q-bio.MN cs.AI cs.LG q-bio.QM stat.CO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Rapid discovery of new reactions and molecules in recent years has been
facilitated by the advancements in high throughput screening, accessibility to
a much more complex chemical design space, and the development of accurate
molecular modeling frameworks. A holistic study of the growing chemistry
literature is, therefore, required that focuses on understanding the recent
trends and extrapolating them into possible future trajectories. To this end,
several network theory-based studies have been reported that use a directed
graph representation of chemical reactions. Here, we perform a study based on
representing chemical reactions as hypergraphs where the hyperedges represent
chemical reactions and nodes represent the participating molecules. We use a
standard reactions dataset to construct a hypernetwork and report its
statistics such as degree distributions, average path length, assortativity or
degree correlations, PageRank centrality, and graph-based clusters (or
communities). We also compute each statistic for an equivalent directed graph
representation of reactions to draw parallels and highlight differences between
the two. To demonstrate the AI applicability of hypergraph reaction
representation, we generate dense hypergraph embeddings and use them in the
reaction classification problem. We conclude that the hypernetwork
representation is flexible, preserves reaction context, and uncovers hidden
insights that are otherwise not apparent in a traditional directed graph
representation of chemical reactions.
| [
{
"version": "v1",
"created": "Tue, 2 Aug 2022 14:12:03 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 15:43:43 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Mann",
"Vipul",
""
],
[
"Venkatasubramanian",
"Venkat",
""
]
] | TITLE: AI-driven Hypergraph Network of Organic Chemistry: Network Statistics
and Applications in Reaction Classification
ABSTRACT: Rapid discovery of new reactions and molecules in recent years has been
facilitated by the advancements in high throughput screening, accessibility to
a much more complex chemical design space, and the development of accurate
molecular modeling frameworks. A holistic study of the growing chemistry
literature is, therefore, required that focuses on understanding the recent
trends and extrapolating them into possible future trajectories. To this end,
several network theory-based studies have been reported that use a directed
graph representation of chemical reactions. Here, we perform a study based on
representing chemical reactions as hypergraphs where the hyperedges represent
chemical reactions and nodes represent the participating molecules. We use a
standard reactions dataset to construct a hypernetwork and report its
statistics such as degree distributions, average path length, assortativity or
degree correlations, PageRank centrality, and graph-based clusters (or
communities). We also compute each statistic for an equivalent directed graph
representation of reactions to draw parallels and highlight differences between
the two. To demonstrate the AI applicability of hypergraph reaction
representation, we generate dense hypergraph embeddings and use them in the
reaction classification problem. We conclude that the hypernetwork
representation is flexible, preserves reaction context, and uncovers hidden
insights that are otherwise not apparent in a traditional directed graph
representation of chemical reactions.
| no_new_dataset | 0.948632 |
2208.09910 | Lihe Yang | Lihe Yang, Lei Qi, Litong Feng, Wayne Zhang, Yinghuan Shi | Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic
Segmentation | Accepted by CVPR 2023. Code and logs:
https://github.com/LiheYoung/UniMatch | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we revisit the weak-to-strong consistency framework,
popularized by FixMatch from semi-supervised classification, where the
prediction of a weakly perturbed image serves as supervision for its strongly
perturbed version. Intriguingly, we observe that such a simple pipeline already
achieves competitive results against recent advanced works, when transferred to
our segmentation scenario. Its success heavily relies on the manual design of
strong data augmentations, however, which may be limited and inadequate to
explore a broader perturbation space. Motivated by this, we propose an
auxiliary feature perturbation stream as a supplement, leading to an expanded
perturbation space. On the other, to sufficiently probe original image-level
augmentations, we present a dual-stream perturbation technique, enabling two
strong views to be simultaneously guided by a common weak view. Consequently,
our overall Unified Dual-Stream Perturbations approach (UniMatch) surpasses all
existing methods significantly across all evaluation protocols on the Pascal,
Cityscapes, and COCO benchmarks. Its superiority is also demonstrated in remote
sensing interpretation and medical image analysis. We hope our reproduced
FixMatch and our results can inspire more future works. Code and logs are
available at https://github.com/LiheYoung/UniMatch.
| [
{
"version": "v1",
"created": "Sun, 21 Aug 2022 15:32:43 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 07:10:13 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yang",
"Lihe",
""
],
[
"Qi",
"Lei",
""
],
[
"Feng",
"Litong",
""
],
[
"Zhang",
"Wayne",
""
],
[
"Shi",
"Yinghuan",
""
]
] | TITLE: Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic
Segmentation
ABSTRACT: In this work, we revisit the weak-to-strong consistency framework,
popularized by FixMatch from semi-supervised classification, where the
prediction of a weakly perturbed image serves as supervision for its strongly
perturbed version. Intriguingly, we observe that such a simple pipeline already
achieves competitive results against recent advanced works, when transferred to
our segmentation scenario. Its success heavily relies on the manual design of
strong data augmentations, however, which may be limited and inadequate to
explore a broader perturbation space. Motivated by this, we propose an
auxiliary feature perturbation stream as a supplement, leading to an expanded
perturbation space. On the other, to sufficiently probe original image-level
augmentations, we present a dual-stream perturbation technique, enabling two
strong views to be simultaneously guided by a common weak view. Consequently,
our overall Unified Dual-Stream Perturbations approach (UniMatch) surpasses all
existing methods significantly across all evaluation protocols on the Pascal,
Cityscapes, and COCO benchmarks. Its superiority is also demonstrated in remote
sensing interpretation and medical image analysis. We hope our reproduced
FixMatch and our results can inspire more future works. Code and logs are
available at https://github.com/LiheYoung/UniMatch.
| no_new_dataset | 0.944331 |
2208.14558 | Stefan Larson | Alexander Groleau, Kok Wei Chee, Stefan Larson, Samay Maini, Jonathan
Boarman | Augraphy: A Data Augmentation Library for Document Images | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper introduces Augraphy, a Python library for constructing data
augmentation pipelines which produce distortions commonly seen in real-world
document image datasets. Augraphy stands apart from other data augmentation
tools by providing many different strategies to produce augmented versions of
clean document images that appear as if they have been altered by standard
office operations, such as printing, scanning, and faxing through old or dirty
machines, degradation of ink over time, and handwritten markings. This paper
discusses the Augraphy tool, and shows how it can be used both as a data
augmentation tool for producing diverse training data for tasks such as
document denoising, and also for generating challenging test data to evaluate
model robustness on document image modeling tasks.
| [
{
"version": "v1",
"created": "Tue, 30 Aug 2022 22:36:19 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 21:49:21 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Groleau",
"Alexander",
""
],
[
"Chee",
"Kok Wei",
""
],
[
"Larson",
"Stefan",
""
],
[
"Maini",
"Samay",
""
],
[
"Boarman",
"Jonathan",
""
]
] | TITLE: Augraphy: A Data Augmentation Library for Document Images
ABSTRACT: This paper introduces Augraphy, a Python library for constructing data
augmentation pipelines which produce distortions commonly seen in real-world
document image datasets. Augraphy stands apart from other data augmentation
tools by providing many different strategies to produce augmented versions of
clean document images that appear as if they have been altered by standard
office operations, such as printing, scanning, and faxing through old or dirty
machines, degradation of ink over time, and handwritten markings. This paper
discusses the Augraphy tool, and shows how it can be used both as a data
augmentation tool for producing diverse training data for tasks such as
document denoising, and also for generating challenging test data to evaluate
model robustness on document image modeling tasks.
| no_new_dataset | 0.946547 |
2209.04870 | Marcin Bienkowski | Marcin Bienkowski, Marcin Mucha | An Improved Algorithm For Online Min-Sum Set Cover | Presented at AAAI 2023 | null | null | null | cs.DS cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study a fundamental model of online preference aggregation, where an
algorithm maintains an ordered list of $n$ elements. An input is a stream of
preferred sets $R_1, R_2, \dots, R_t, \dots$. Upon seeing $R_t$ and without
knowledge of any future sets, an algorithm has to rerank elements (change the
list ordering), so that at least one element of $R_t$ is found near the list
front. The incurred cost is a sum of the list update costs (the number of swaps
of neighboring list elements) and access costs (position of the first element
of $R_t$ on the list). This scenario occurs naturally in applications such as
ordering items in an online shop using aggregated preferences of shop
customers. The theoretical underpinning of this problem is known as Min-Sum Set
Cover.
Unlike previous work (Fotakis et al., ICALP 2020, NIPS 2020) that mostly
studied the performance of an online algorithm ALG against the static optimal
solution (a single optimal list ordering), in this paper, we study an arguably
harder variant where the benchmark is the provably stronger optimal dynamic
solution OPT (that may also modify the list ordering). In terms of an online
shop, this means that the aggregated preferences of its user base evolve with
time. We construct a computationally efficient randomized algorithm whose
competitive ratio (ALG-to-OPT cost ratio) is $O(r^2)$ and prove the existence
of a deterministic $O(r^4)$-competitive algorithm. Here, $r$ is the maximum
cardinality of sets $R_t$. This is the first algorithm whose ratio does not
depend on $n$: the previously best algorithm for this problem was $O(r^{3/2}
\cdot \sqrt{n})$-competitive and $\Omega(r)$ is a lower bound on the
performance of any deterministic online algorithm.
| [
{
"version": "v1",
"created": "Sun, 11 Sep 2022 14:03:51 GMT"
},
{
"version": "v2",
"created": "Sun, 4 Dec 2022 22:39:35 GMT"
},
{
"version": "v3",
"created": "Fri, 24 Mar 2023 23:44:49 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Bienkowski",
"Marcin",
""
],
[
"Mucha",
"Marcin",
""
]
] | TITLE: An Improved Algorithm For Online Min-Sum Set Cover
ABSTRACT: We study a fundamental model of online preference aggregation, where an
algorithm maintains an ordered list of $n$ elements. An input is a stream of
preferred sets $R_1, R_2, \dots, R_t, \dots$. Upon seeing $R_t$ and without
knowledge of any future sets, an algorithm has to rerank elements (change the
list ordering), so that at least one element of $R_t$ is found near the list
front. The incurred cost is a sum of the list update costs (the number of swaps
of neighboring list elements) and access costs (position of the first element
of $R_t$ on the list). This scenario occurs naturally in applications such as
ordering items in an online shop using aggregated preferences of shop
customers. The theoretical underpinning of this problem is known as Min-Sum Set
Cover.
Unlike previous work (Fotakis et al., ICALP 2020, NIPS 2020) that mostly
studied the performance of an online algorithm ALG against the static optimal
solution (a single optimal list ordering), in this paper, we study an arguably
harder variant where the benchmark is the provably stronger optimal dynamic
solution OPT (that may also modify the list ordering). In terms of an online
shop, this means that the aggregated preferences of its user base evolve with
time. We construct a computationally efficient randomized algorithm whose
competitive ratio (ALG-to-OPT cost ratio) is $O(r^2)$ and prove the existence
of a deterministic $O(r^4)$-competitive algorithm. Here, $r$ is the maximum
cardinality of sets $R_t$. This is the first algorithm whose ratio does not
depend on $n$: the previously best algorithm for this problem was $O(r^{3/2}
\cdot \sqrt{n})$-competitive and $\Omega(r)$ is a lower bound on the
performance of any deterministic online algorithm.
| no_new_dataset | 0.948822 |
2209.06278 | Shanyin Tong | Shanyin Tong and Georg Stadler | Large deviation theory-based adaptive importance sampling for rare
events in high dimensions | null | null | null | null | stat.CO cs.NA math.NA math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a method for the accurate estimation of rare event or failure
probabilities for expensive-to-evaluate numerical models in high dimensions.
The proposed approach combines ideas from large deviation theory and adaptive
importance sampling. The importance sampler uses a cross-entropy method to find
an optimal Gaussian biasing distribution, and reuses all samples made
throughout the process for both, the target probability estimation and for
updating the biasing distributions. Large deviation theory is used to find a
good initial biasing distribution through the solution of an optimization
problem. Additionally, it is used to identify a low-dimensional subspace that
is most informative of the rare event probability. This subspace is used for
the cross-entropy method, which is known to lose efficiency in higher
dimensions. The proposed method does not require smoothing of indicator
functions nor does it involve numerical tuning parameters. We compare the
method with a state-of-the-art cross-entropy-based importance sampling scheme
using three examples: a high-dimensional failure probability estimation
benchmark, a problem governed by a diffusion equation, and a tsunami problem
governed by the time-dependent shallow water system in one spatial dimension.
| [
{
"version": "v1",
"created": "Tue, 13 Sep 2022 19:52:39 GMT"
},
{
"version": "v2",
"created": "Mon, 26 Sep 2022 02:38:46 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 00:12:54 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Tong",
"Shanyin",
""
],
[
"Stadler",
"Georg",
""
]
] | TITLE: Large deviation theory-based adaptive importance sampling for rare
events in high dimensions
ABSTRACT: We propose a method for the accurate estimation of rare event or failure
probabilities for expensive-to-evaluate numerical models in high dimensions.
The proposed approach combines ideas from large deviation theory and adaptive
importance sampling. The importance sampler uses a cross-entropy method to find
an optimal Gaussian biasing distribution, and reuses all samples made
throughout the process for both, the target probability estimation and for
updating the biasing distributions. Large deviation theory is used to find a
good initial biasing distribution through the solution of an optimization
problem. Additionally, it is used to identify a low-dimensional subspace that
is most informative of the rare event probability. This subspace is used for
the cross-entropy method, which is known to lose efficiency in higher
dimensions. The proposed method does not require smoothing of indicator
functions nor does it involve numerical tuning parameters. We compare the
method with a state-of-the-art cross-entropy-based importance sampling scheme
using three examples: a high-dimensional failure probability estimation
benchmark, a problem governed by a diffusion equation, and a tsunami problem
governed by the time-dependent shallow water system in one spatial dimension.
| no_new_dataset | 0.951051 |
2209.12152 | Fan Bao | Fan Bao, Shen Nie, Kaiwen Xue, Yue Cao, Chongxuan Li, Hang Su, Jun Zhu | All are Worth Words: A ViT Backbone for Diffusion Models | Accepted to CVPR 2023 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision transformers (ViT) have shown promise in various vision tasks while
the U-Net based on a convolutional neural network (CNN) remains dominant in
diffusion models. We design a simple and general ViT-based architecture (named
U-ViT) for image generation with diffusion models. U-ViT is characterized by
treating all inputs including the time, condition and noisy image patches as
tokens and employing long skip connections between shallow and deep layers. We
evaluate U-ViT in unconditional and class-conditional image generation, as well
as text-to-image generation tasks, where U-ViT is comparable if not superior to
a CNN-based U-Net of a similar size. In particular, latent diffusion models
with U-ViT achieve record-breaking FID scores of 2.29 in class-conditional
image generation on ImageNet 256x256, and 5.48 in text-to-image generation on
MS-COCO, among methods without accessing large external datasets during the
training of generative models. Our results suggest that, for diffusion-based
image modeling, the long skip connection is crucial while the down-sampling and
up-sampling operators in CNN-based U-Net are not always necessary. We believe
that U-ViT can provide insights for future research on backbones in diffusion
models and benefit generative modeling on large scale cross-modality datasets.
| [
{
"version": "v1",
"created": "Sun, 25 Sep 2022 05:21:59 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Nov 2022 08:44:03 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Mar 2023 18:28:18 GMT"
},
{
"version": "v4",
"created": "Sat, 25 Mar 2023 13:01:42 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Bao",
"Fan",
""
],
[
"Nie",
"Shen",
""
],
[
"Xue",
"Kaiwen",
""
],
[
"Cao",
"Yue",
""
],
[
"Li",
"Chongxuan",
""
],
[
"Su",
"Hang",
""
],
[
"Zhu",
"Jun",
""
]
] | TITLE: All are Worth Words: A ViT Backbone for Diffusion Models
ABSTRACT: Vision transformers (ViT) have shown promise in various vision tasks while
the U-Net based on a convolutional neural network (CNN) remains dominant in
diffusion models. We design a simple and general ViT-based architecture (named
U-ViT) for image generation with diffusion models. U-ViT is characterized by
treating all inputs including the time, condition and noisy image patches as
tokens and employing long skip connections between shallow and deep layers. We
evaluate U-ViT in unconditional and class-conditional image generation, as well
as text-to-image generation tasks, where U-ViT is comparable if not superior to
a CNN-based U-Net of a similar size. In particular, latent diffusion models
with U-ViT achieve record-breaking FID scores of 2.29 in class-conditional
image generation on ImageNet 256x256, and 5.48 in text-to-image generation on
MS-COCO, among methods without accessing large external datasets during the
training of generative models. Our results suggest that, for diffusion-based
image modeling, the long skip connection is crucial while the down-sampling and
up-sampling operators in CNN-based U-Net are not always necessary. We believe
that U-ViT can provide insights for future research on backbones in diffusion
models and benefit generative modeling on large scale cross-modality datasets.
| no_new_dataset | 0.950686 |
2209.14627 | Yuqiao Wen | Yuqiao Wen, Yongchang Hao, Yanshuai Cao, Lili Mou | An Equal-Size Hard EM Algorithm for Diverse Dialogue Generation | Accepted by ICLR 2023 | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Open-domain dialogue systems aim to interact with humans through natural
language texts in an open-ended fashion. Despite the recent success of super
large dialogue systems such as ChatGPT, using medium-to-small-sized dialogue
systems remains the common practice as they are more lightweight and
accessible; however, generating diverse dialogue responses is challenging,
especially with smaller models. In this work, we propose an Equal-size Hard
Expectation--Maximization (EqHard-EM) algorithm to train a multi-decoder model
for diverse dialogue generation. Our algorithm assigns a sample to a decoder in
a hard manner and additionally imposes an equal-assignment constraint to ensure
that all decoders are well-trained. We provide detailed theoretical analysis to
justify our approach. Further, experiments on two large-scale open-domain
dialogue datasets verify that our EqHard-EM algorithm generates high-quality
diverse responses.
| [
{
"version": "v1",
"created": "Thu, 29 Sep 2022 08:41:32 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 20:10:15 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wen",
"Yuqiao",
""
],
[
"Hao",
"Yongchang",
""
],
[
"Cao",
"Yanshuai",
""
],
[
"Mou",
"Lili",
""
]
] | TITLE: An Equal-Size Hard EM Algorithm for Diverse Dialogue Generation
ABSTRACT: Open-domain dialogue systems aim to interact with humans through natural
language texts in an open-ended fashion. Despite the recent success of super
large dialogue systems such as ChatGPT, using medium-to-small-sized dialogue
systems remains the common practice as they are more lightweight and
accessible; however, generating diverse dialogue responses is challenging,
especially with smaller models. In this work, we propose an Equal-size Hard
Expectation--Maximization (EqHard-EM) algorithm to train a multi-decoder model
for diverse dialogue generation. Our algorithm assigns a sample to a decoder in
a hard manner and additionally imposes an equal-assignment constraint to ensure
that all decoders are well-trained. We provide detailed theoretical analysis to
justify our approach. Further, experiments on two large-scale open-domain
dialogue datasets verify that our EqHard-EM algorithm generates high-quality
diverse responses.
| no_new_dataset | 0.951863 |
2210.01781 | Boxiao Pan | Boxiao Pan, Bokui Shen, Davis Rempe, Despoina Paschalidou, Kaichun Mo,
Yanchao Yang, Leonidas J. Guibas | COPILOT: Human-Environment Collision Prediction and Localization from
Egocentric Videos | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ability to forecast human-environment collisions from egocentric
observations is vital to enable collision avoidance in applications such as VR,
AR, and wearable assistive robotics. In this work, we introduce the challenging
problem of predicting collisions in diverse environments from multi-view
egocentric videos captured from body-mounted cameras. Solving this problem
requires a generalizable perception system that can classify which human body
joints will collide and estimate a collision region heatmap to localize
collisions in the environment. To achieve this, we propose a transformer-based
model called COPILOT to perform collision prediction and localization
simultaneously, which accumulates information across multi-view inputs through
a novel 4D space-time-viewpoint attention mechanism. To train our model and
enable future research on this task, we develop a synthetic data generation
framework that produces egocentric videos of virtual humans moving and
colliding within diverse 3D environments. This framework is then used to
establish a large-scale dataset consisting of 8.6M egocentric RGBD frames.
Extensive experiments show that COPILOT generalizes to unseen synthetic as well
as real-world scenes. We further demonstrate COPILOT outputs are useful for
downstream collision avoidance through simple closed-loop control. Please visit
our project webpage at https://sites.google.com/stanford.edu/copilot.
| [
{
"version": "v1",
"created": "Tue, 4 Oct 2022 17:49:23 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 05:27:31 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Pan",
"Boxiao",
""
],
[
"Shen",
"Bokui",
""
],
[
"Rempe",
"Davis",
""
],
[
"Paschalidou",
"Despoina",
""
],
[
"Mo",
"Kaichun",
""
],
[
"Yang",
"Yanchao",
""
],
[
"Guibas",
"Leonidas J.",
""
]
] | TITLE: COPILOT: Human-Environment Collision Prediction and Localization from
Egocentric Videos
ABSTRACT: The ability to forecast human-environment collisions from egocentric
observations is vital to enable collision avoidance in applications such as VR,
AR, and wearable assistive robotics. In this work, we introduce the challenging
problem of predicting collisions in diverse environments from multi-view
egocentric videos captured from body-mounted cameras. Solving this problem
requires a generalizable perception system that can classify which human body
joints will collide and estimate a collision region heatmap to localize
collisions in the environment. To achieve this, we propose a transformer-based
model called COPILOT to perform collision prediction and localization
simultaneously, which accumulates information across multi-view inputs through
a novel 4D space-time-viewpoint attention mechanism. To train our model and
enable future research on this task, we develop a synthetic data generation
framework that produces egocentric videos of virtual humans moving and
colliding within diverse 3D environments. This framework is then used to
establish a large-scale dataset consisting of 8.6M egocentric RGBD frames.
Extensive experiments show that COPILOT generalizes to unseen synthetic as well
as real-world scenes. We further demonstrate COPILOT outputs are useful for
downstream collision avoidance through simple closed-loop control. Please visit
our project webpage at https://sites.google.com/stanford.edu/copilot.
| new_dataset | 0.949435 |
2210.07453 | Jonathan Pilault | Jonathan Pilault, Michael Galkin, Bahare Fatemi, Perouz Taslakian,
David Vasquez, Christopher Pal | Using Graph Algorithms to Pretrain Graph Completion Transformers | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent work on Graph Neural Networks has demonstrated that self-supervised
pretraining can further enhance performance on downstream graph, link, and node
classification tasks. However, the efficacy of pretraining tasks has not been
fully investigated for downstream large knowledge graph completion tasks. Using
a contextualized knowledge graph embedding approach, we investigate five
different pretraining signals, constructed using several graph algorithms and
no external data, as well as their combination. We leverage the versatility of
our Transformer-based model to explore graph structure generation pretraining
tasks (i.e. path and k-hop neighborhood generation), typically inapplicable to
most graph embedding methods. We further propose a new path-finding algorithm
guided by information gain and find that it is the best-performing pretraining
task across three downstream knowledge graph completion datasets. While using
our new path-finding algorithm as a pretraining signal provides 2-3% MRR
improvements, we show that pretraining on all signals together gives the best
knowledge graph completion results. In a multitask setting that combines all
pretraining tasks, our method surpasses the latest and strong performing
knowledge graph embedding methods on all metrics for FB15K-237, on MRR and
Hit@1 for WN18RRand on MRR and hit@10 for JF17K (a knowledge hypergraph
dataset).
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2022 01:41:10 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 15:04:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Pilault",
"Jonathan",
""
],
[
"Galkin",
"Michael",
""
],
[
"Fatemi",
"Bahare",
""
],
[
"Taslakian",
"Perouz",
""
],
[
"Vasquez",
"David",
""
],
[
"Pal",
"Christopher",
""
]
] | TITLE: Using Graph Algorithms to Pretrain Graph Completion Transformers
ABSTRACT: Recent work on Graph Neural Networks has demonstrated that self-supervised
pretraining can further enhance performance on downstream graph, link, and node
classification tasks. However, the efficacy of pretraining tasks has not been
fully investigated for downstream large knowledge graph completion tasks. Using
a contextualized knowledge graph embedding approach, we investigate five
different pretraining signals, constructed using several graph algorithms and
no external data, as well as their combination. We leverage the versatility of
our Transformer-based model to explore graph structure generation pretraining
tasks (i.e. path and k-hop neighborhood generation), typically inapplicable to
most graph embedding methods. We further propose a new path-finding algorithm
guided by information gain and find that it is the best-performing pretraining
task across three downstream knowledge graph completion datasets. While using
our new path-finding algorithm as a pretraining signal provides 2-3% MRR
improvements, we show that pretraining on all signals together gives the best
knowledge graph completion results. In a multitask setting that combines all
pretraining tasks, our method surpasses the latest and strong performing
knowledge graph embedding methods on all metrics for FB15K-237, on MRR and
Hit@1 for WN18RRand on MRR and hit@10 for JF17K (a knowledge hypergraph
dataset).
| no_new_dataset | 0.941708 |
2210.08198 | Boran Hao | Ruidi Chen, Boran Hao, Ioannis Ch. Paschalidis | Distributionally Robust Multiclass Classification and Applications in
Deep Image Classifiers | This work was intended as a replacement of arXiv:2109.12772 and any
subsequent updates will appear there | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We develop a Distributionally Robust Optimization (DRO) formulation for
Multiclass Logistic Regression (MLR), which could tolerate data contaminated by
outliers. The DRO framework uses a probabilistic ambiguity set defined as a
ball of distributions that are close to the empirical distribution of the
training set in the sense of the Wasserstein metric. We relax the DRO
formulation into a regularized learning problem whose regularizer is a norm of
the coefficient matrix. We establish out-of-sample performance guarantees for
the solutions to our model, offering insights on the role of the regularizer in
controlling the prediction error. We apply the proposed method in rendering
deep Vision Transformer (ViT)-based image classifiers robust to random and
adversarial attacks. Specifically, using the MNIST and CIFAR-10 datasets, we
demonstrate reductions in test error rate by up to 83.5% and loss by up to
91.3% compared with baseline methods, by adopting a novel random training
method.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2022 05:09:28 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 18:29:49 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Chen",
"Ruidi",
""
],
[
"Hao",
"Boran",
""
],
[
"Paschalidis",
"Ioannis Ch.",
""
]
] | TITLE: Distributionally Robust Multiclass Classification and Applications in
Deep Image Classifiers
ABSTRACT: We develop a Distributionally Robust Optimization (DRO) formulation for
Multiclass Logistic Regression (MLR), which could tolerate data contaminated by
outliers. The DRO framework uses a probabilistic ambiguity set defined as a
ball of distributions that are close to the empirical distribution of the
training set in the sense of the Wasserstein metric. We relax the DRO
formulation into a regularized learning problem whose regularizer is a norm of
the coefficient matrix. We establish out-of-sample performance guarantees for
the solutions to our model, offering insights on the role of the regularizer in
controlling the prediction error. We apply the proposed method in rendering
deep Vision Transformer (ViT)-based image classifiers robust to random and
adversarial attacks. Specifically, using the MNIST and CIFAR-10 datasets, we
demonstrate reductions in test error rate by up to 83.5% and loss by up to
91.3% compared with baseline methods, by adopting a novel random training
method.
| no_new_dataset | 0.947137 |
2210.08344 | Qi Zhang | Qi Zhang, Yifei Wang, Yisen Wang | How Mask Matters: Towards Theoretical Understandings of Masked
Autoencoders | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Masked Autoencoders (MAE) based on a reconstruction task have risen to be a
promising paradigm for self-supervised learning (SSL) and achieve
state-of-the-art performance across different benchmark datasets. However,
despite its impressive empirical success, there is still limited theoretical
understanding of it. In this paper, we propose a theoretical understanding of
how masking matters for MAE to learn meaningful features. We establish a close
connection between MAE and contrastive learning, which shows that MAE implicit
aligns the mask-induced positive pairs. Built upon this connection, we develop
the first downstream guarantees for MAE methods, and analyze the effect of mask
ratio. Besides, as a result of the implicit alignment, we also point out the
dimensional collapse issue of MAE, and propose a Uniformity-enhanced MAE
(U-MAE) loss that can effectively address this issue and bring significant
improvements on real-world datasets, including CIFAR-10, ImageNet-100, and
ImageNet-1K. Code is available at (https://github.com/zhangq327/U-MAE).
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2022 17:36:03 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 06:47:53 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Qi",
""
],
[
"Wang",
"Yifei",
""
],
[
"Wang",
"Yisen",
""
]
] | TITLE: How Mask Matters: Towards Theoretical Understandings of Masked
Autoencoders
ABSTRACT: Masked Autoencoders (MAE) based on a reconstruction task have risen to be a
promising paradigm for self-supervised learning (SSL) and achieve
state-of-the-art performance across different benchmark datasets. However,
despite its impressive empirical success, there is still limited theoretical
understanding of it. In this paper, we propose a theoretical understanding of
how masking matters for MAE to learn meaningful features. We establish a close
connection between MAE and contrastive learning, which shows that MAE implicit
aligns the mask-induced positive pairs. Built upon this connection, we develop
the first downstream guarantees for MAE methods, and analyze the effect of mask
ratio. Besides, as a result of the implicit alignment, we also point out the
dimensional collapse issue of MAE, and propose a Uniformity-enhanced MAE
(U-MAE) loss that can effectively address this issue and bring significant
improvements on real-world datasets, including CIFAR-10, ImageNet-100, and
ImageNet-1K. Code is available at (https://github.com/zhangq327/U-MAE).
| no_new_dataset | 0.941868 |
2210.09693 | Chaoli Zhang | Chaoli Zhang and Tian Zhou and Qingsong Wen and Liang Sun | TFAD: A Decomposition Time Series Anomaly Detection Architecture with
Time-Frequency Analysis | Accepted by the ACM International Conference on Information and
Knowledge Management (CIKM 2022) | CIKM 2022 | 10.1145/3511808.3557470 | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time series anomaly detection is a challenging problem due to the complex
temporal dependencies and the limited label data. Although some algorithms
including both traditional and deep models have been proposed, most of them
mainly focus on time-domain modeling, and do not fully utilize the information
in the frequency domain of the time series data. In this paper, we propose a
Time-Frequency analysis based time series Anomaly Detection model, or TFAD for
short, to exploit both time and frequency domains for performance improvement.
Besides, we incorporate time series decomposition and data augmentation
mechanisms in the designed time-frequency architecture to further boost the
abilities of performance and interpretability. Empirical studies on widely used
benchmark datasets show that our approach obtains state-of-the-art performance
in univariate and multivariate time series anomaly detection tasks. Code is
provided at https://github.com/DAMO-DI-ML/CIKM22-TFAD.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2022 09:08:57 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 06:26:36 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Chaoli",
""
],
[
"Zhou",
"Tian",
""
],
[
"Wen",
"Qingsong",
""
],
[
"Sun",
"Liang",
""
]
] | TITLE: TFAD: A Decomposition Time Series Anomaly Detection Architecture with
Time-Frequency Analysis
ABSTRACT: Time series anomaly detection is a challenging problem due to the complex
temporal dependencies and the limited label data. Although some algorithms
including both traditional and deep models have been proposed, most of them
mainly focus on time-domain modeling, and do not fully utilize the information
in the frequency domain of the time series data. In this paper, we propose a
Time-Frequency analysis based time series Anomaly Detection model, or TFAD for
short, to exploit both time and frequency domains for performance improvement.
Besides, we incorporate time series decomposition and data augmentation
mechanisms in the designed time-frequency architecture to further boost the
abilities of performance and interpretability. Empirical studies on widely used
benchmark datasets show that our approach obtains state-of-the-art performance
in univariate and multivariate time series anomaly detection tasks. Code is
provided at https://github.com/DAMO-DI-ML/CIKM22-TFAD.
| no_new_dataset | 0.949201 |
2210.15664 | Vladislav Golyanik | Edith Tretschk, Navami Kairanda, Mallikarjun B R, Rishabh Dabral, Adam
Kortylewski, Bernhard Egger, Marc Habermann, Pascal Fua, Christian Theobalt,
Vladislav Golyanik | State of the Art in Dense Monocular Non-Rigid 3D Reconstruction | 36 pages, 18 figures, 3 tables; State-of-the-Art Report at
EUROGRAPHICS 2023 | Computer Graphics Forum, 2023 | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 3D reconstruction of deformable (or non-rigid) scenes from a set of monocular
2D image observations is a long-standing and actively researched area of
computer vision and graphics. It is an ill-posed inverse problem, since --
without additional prior assumptions -- it permits infinitely many solutions
leading to accurate projection to the input 2D images. Non-rigid reconstruction
is a foundational building block for downstream applications like robotics,
AR/VR, or visual content creation. The key advantage of using monocular cameras
is their omnipresence and availability to the end users as well as their ease
of use compared to more sophisticated camera set-ups such as stereo or
multi-view systems. This survey focuses on state-of-the-art methods for dense
non-rigid 3D reconstruction of various deformable objects and composite scenes
from monocular videos or sets of monocular views. It reviews the fundamentals
of 3D reconstruction and deformation modeling from 2D image observations. We
then start from general methods -- that handle arbitrary scenes and make only a
few prior assumptions -- and proceed towards techniques making stronger
assumptions about the observed objects and types of deformations (e.g. human
faces, bodies, hands, and animals). A significant part of this STAR is also
devoted to classification and a high-level comparison of the methods, as well
as an overview of the datasets for training and evaluation of the discussed
techniques. We conclude by discussing open challenges in the field and the
social aspects associated with the usage of the reviewed methods.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2022 17:59:53 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 18:45:56 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Tretschk",
"Edith",
""
],
[
"Kairanda",
"Navami",
""
],
[
"R",
"Mallikarjun B",
""
],
[
"Dabral",
"Rishabh",
""
],
[
"Kortylewski",
"Adam",
""
],
[
"Egger",
"Bernhard",
""
],
[
"Habermann",
"Marc",
""
],
[
"Fua",
"Pascal",
""
],
[
"Theobalt",
"Christian",
""
],
[
"Golyanik",
"Vladislav",
""
]
] | TITLE: State of the Art in Dense Monocular Non-Rigid 3D Reconstruction
ABSTRACT: 3D reconstruction of deformable (or non-rigid) scenes from a set of monocular
2D image observations is a long-standing and actively researched area of
computer vision and graphics. It is an ill-posed inverse problem, since --
without additional prior assumptions -- it permits infinitely many solutions
leading to accurate projection to the input 2D images. Non-rigid reconstruction
is a foundational building block for downstream applications like robotics,
AR/VR, or visual content creation. The key advantage of using monocular cameras
is their omnipresence and availability to the end users as well as their ease
of use compared to more sophisticated camera set-ups such as stereo or
multi-view systems. This survey focuses on state-of-the-art methods for dense
non-rigid 3D reconstruction of various deformable objects and composite scenes
from monocular videos or sets of monocular views. It reviews the fundamentals
of 3D reconstruction and deformation modeling from 2D image observations. We
then start from general methods -- that handle arbitrary scenes and make only a
few prior assumptions -- and proceed towards techniques making stronger
assumptions about the observed objects and types of deformations (e.g. human
faces, bodies, hands, and animals). A significant part of this STAR is also
devoted to classification and a high-level comparison of the methods, as well
as an overview of the datasets for training and evaluation of the discussed
techniques. We conclude by discussing open challenges in the field and the
social aspects associated with the usage of the reviewed methods.
| no_new_dataset | 0.9455 |
2210.16046 | Masakazu Yoshimura | Masakazu Yoshimura, Junji Otsuka, Atsushi Irie, Takeshi Ohashi | Rawgment: Noise-Accounted RAW Augmentation Enables Recognition in a Wide
Variety of Environments | Accepted to CVPR2023 | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Image recognition models that work in challenging environments (e.g.,
extremely dark, blurry, or high dynamic range conditions) must be useful.
However, creating training datasets for such environments is expensive and hard
due to the difficulties of data collection and annotation. It is desirable if
we could get a robust model without the need for hard-to-obtain datasets. One
simple approach is to apply data augmentation such as color jitter and blur to
standard RGB (sRGB) images in simple scenes. Unfortunately, this approach
struggles to yield realistic images in terms of pixel intensity and noise
distribution due to not considering the non-linearity of Image Signal
Processors (ISPs) and noise characteristics of image sensors. Instead, we
propose a noise-accounted RAW image augmentation method. In essence, color
jitter and blur augmentation are applied to a RAW image before applying
non-linear ISP, resulting in realistic intensity. Furthermore, we introduce a
noise amount alignment method that calibrates the domain gap in the noise
property caused by the augmentation. We show that our proposed noise-accounted
RAW augmentation method doubles the image recognition accuracy in challenging
environments only with simple training data.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2022 10:33:45 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 06:17:13 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yoshimura",
"Masakazu",
""
],
[
"Otsuka",
"Junji",
""
],
[
"Irie",
"Atsushi",
""
],
[
"Ohashi",
"Takeshi",
""
]
] | TITLE: Rawgment: Noise-Accounted RAW Augmentation Enables Recognition in a Wide
Variety of Environments
ABSTRACT: Image recognition models that work in challenging environments (e.g.,
extremely dark, blurry, or high dynamic range conditions) must be useful.
However, creating training datasets for such environments is expensive and hard
due to the difficulties of data collection and annotation. It is desirable if
we could get a robust model without the need for hard-to-obtain datasets. One
simple approach is to apply data augmentation such as color jitter and blur to
standard RGB (sRGB) images in simple scenes. Unfortunately, this approach
struggles to yield realistic images in terms of pixel intensity and noise
distribution due to not considering the non-linearity of Image Signal
Processors (ISPs) and noise characteristics of image sensors. Instead, we
propose a noise-accounted RAW image augmentation method. In essence, color
jitter and blur augmentation are applied to a RAW image before applying
non-linear ISP, resulting in realistic intensity. Furthermore, we introduce a
noise amount alignment method that calibrates the domain gap in the noise
property caused by the augmentation. We show that our proposed noise-accounted
RAW augmentation method doubles the image recognition accuracy in challenging
environments only with simple training data.
| no_new_dataset | 0.954351 |
2211.04188 | Francesco Barbato | Francesco Barbato, Giulia Rizzoli, Pietro Zanuttigh | DepthFormer: Multimodal Positional Encodings and Cross-Input Attention
for Transformer-Based Segmentation Networks | Accepted at ICASSP 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most approaches for semantic segmentation use only information from color
cameras to parse the scenes, yet recent advancements show that using depth data
allows to further improve performances. In this work, we focus on
transformer-based deep learning architectures, that have achieved
state-of-the-art performances on the segmentation task, and we propose to
employ depth information by embedding it in the positional encoding.
Effectively, we extend the network to multimodal data without adding any
parameters and in a natural way that makes use of the strength of transformers'
self-attention modules. We also investigate the idea of performing
cross-modality operations inside the attention module, swapping the key inputs
between the depth and color branches. Our approach consistently improves
performances on the Cityscapes benchmark.
| [
{
"version": "v1",
"created": "Tue, 8 Nov 2022 12:01:31 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 12:54:49 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Barbato",
"Francesco",
""
],
[
"Rizzoli",
"Giulia",
""
],
[
"Zanuttigh",
"Pietro",
""
]
] | TITLE: DepthFormer: Multimodal Positional Encodings and Cross-Input Attention
for Transformer-Based Segmentation Networks
ABSTRACT: Most approaches for semantic segmentation use only information from color
cameras to parse the scenes, yet recent advancements show that using depth data
allows to further improve performances. In this work, we focus on
transformer-based deep learning architectures, that have achieved
state-of-the-art performances on the segmentation task, and we propose to
employ depth information by embedding it in the positional encoding.
Effectively, we extend the network to multimodal data without adding any
parameters and in a natural way that makes use of the strength of transformers'
self-attention modules. We also investigate the idea of performing
cross-modality operations inside the attention module, swapping the key inputs
between the depth and color branches. Our approach consistently improves
performances on the Cityscapes benchmark.
| no_new_dataset | 0.947235 |
2211.05272 | Chengyang Zhao | Haoran Geng, Helin Xu, Chengyang Zhao, Chao Xu, Li Yi, Siyuan Huang,
He Wang | GAPartNet: Cross-Category Domain-Generalizable Object Perception and
Manipulation via Generalizable and Actionable Parts | To appear in CVPR 2023 (Highlight) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For years, researchers have been devoted to generalizable object perception
and manipulation, where cross-category generalizability is highly desired yet
underexplored. In this work, we propose to learn such cross-category skills via
Generalizable and Actionable Parts (GAParts). By identifying and defining 9
GAPart classes (lids, handles, etc.) in 27 object categories, we construct a
large-scale part-centric interactive dataset, GAPartNet, where we provide rich,
part-level annotations (semantics, poses) for 8,489 part instances on 1,166
objects. Based on GAPartNet, we investigate three cross-category tasks: part
segmentation, part pose estimation, and part-based object manipulation. Given
the significant domain gaps between seen and unseen object categories, we
propose a robust 3D segmentation method from the perspective of domain
generalization by integrating adversarial learning techniques. Our method
outperforms all existing methods by a large margin, no matter on seen or unseen
categories. Furthermore, with part segmentation and pose estimation results, we
leverage the GAPart pose definition to design part-based manipulation
heuristics that can generalize well to unseen object categories in both the
simulator and the real world. Our dataset, code, and demos are available on our
project page.
| [
{
"version": "v1",
"created": "Thu, 10 Nov 2022 00:30:22 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 23:59:07 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Geng",
"Haoran",
""
],
[
"Xu",
"Helin",
""
],
[
"Zhao",
"Chengyang",
""
],
[
"Xu",
"Chao",
""
],
[
"Yi",
"Li",
""
],
[
"Huang",
"Siyuan",
""
],
[
"Wang",
"He",
""
]
] | TITLE: GAPartNet: Cross-Category Domain-Generalizable Object Perception and
Manipulation via Generalizable and Actionable Parts
ABSTRACT: For years, researchers have been devoted to generalizable object perception
and manipulation, where cross-category generalizability is highly desired yet
underexplored. In this work, we propose to learn such cross-category skills via
Generalizable and Actionable Parts (GAParts). By identifying and defining 9
GAPart classes (lids, handles, etc.) in 27 object categories, we construct a
large-scale part-centric interactive dataset, GAPartNet, where we provide rich,
part-level annotations (semantics, poses) for 8,489 part instances on 1,166
objects. Based on GAPartNet, we investigate three cross-category tasks: part
segmentation, part pose estimation, and part-based object manipulation. Given
the significant domain gaps between seen and unseen object categories, we
propose a robust 3D segmentation method from the perspective of domain
generalization by integrating adversarial learning techniques. Our method
outperforms all existing methods by a large margin, no matter on seen or unseen
categories. Furthermore, with part segmentation and pose estimation results, we
leverage the GAPart pose definition to design part-based manipulation
heuristics that can generalize well to unseen object categories in both the
simulator and the real world. Our dataset, code, and demos are available on our
project page.
| new_dataset | 0.951369 |
2211.06315 | Hanyi Hu | Hanyi Hu, Long Zhang, Shuan Li, Zhi Liu, Yao Yang, Chongning Na | Fraudulent User Detection Via Behavior Information Aggregation Network
(BIAN) On Large-Scale Financial Social Network | 6 pages, 1 figure | null | null | null | cs.SI cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Financial frauds cause billions of losses annually and yet it lacks efficient
approaches in detecting frauds considering user profile and their behaviors
simultaneously in social network . A social network forms a graph structure
whilst Graph neural networks (GNN), a promising research domain in Deep
Learning, can seamlessly process non-Euclidean graph data . In financial fraud
detection, the modus operandi of criminals can be identified by analyzing user
profile and their behaviors such as transaction, loaning etc. as well as their
social connectivity. Currently, most GNNs are incapable of selecting important
neighbors since the neighbors' edge attributes (i.e., behaviors) are ignored.
In this paper, we propose a novel behavior information aggregation network
(BIAN) to combine the user behaviors with other user features. Different from
its close "relatives" such as Graph Attention Networks (GAT) and Graph
Transformer Networks (GTN), it aggregates neighbors based on neighboring edge
attribute distribution, namely, user behaviors in financial social network. The
experimental results on a real-world large-scale financial social network
dataset, DGraph, show that BIAN obtains the 10.2% gain in AUROC comparing with
the State-Of-The-Art models.
| [
{
"version": "v1",
"created": "Fri, 4 Nov 2022 08:33:06 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 10:04:27 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Hu",
"Hanyi",
""
],
[
"Zhang",
"Long",
""
],
[
"Li",
"Shuan",
""
],
[
"Liu",
"Zhi",
""
],
[
"Yang",
"Yao",
""
],
[
"Na",
"Chongning",
""
]
] | TITLE: Fraudulent User Detection Via Behavior Information Aggregation Network
(BIAN) On Large-Scale Financial Social Network
ABSTRACT: Financial frauds cause billions of losses annually and yet it lacks efficient
approaches in detecting frauds considering user profile and their behaviors
simultaneously in social network . A social network forms a graph structure
whilst Graph neural networks (GNN), a promising research domain in Deep
Learning, can seamlessly process non-Euclidean graph data . In financial fraud
detection, the modus operandi of criminals can be identified by analyzing user
profile and their behaviors such as transaction, loaning etc. as well as their
social connectivity. Currently, most GNNs are incapable of selecting important
neighbors since the neighbors' edge attributes (i.e., behaviors) are ignored.
In this paper, we propose a novel behavior information aggregation network
(BIAN) to combine the user behaviors with other user features. Different from
its close "relatives" such as Graph Attention Networks (GAT) and Graph
Transformer Networks (GTN), it aggregates neighbors based on neighboring edge
attribute distribution, namely, user behaviors in financial social network. The
experimental results on a real-world large-scale financial social network
dataset, DGraph, show that BIAN obtains the 10.2% gain in AUROC comparing with
the State-Of-The-Art models.
| no_new_dataset | 0.9463 |
2211.11004 | Jiawei Du | Jiawei Du, Yidi Jiang, Vincent Y. F. Tan, Joey Tianyi Zhou, Haizhou Li | Minimizing the Accumulated Trajectory Error to Improve Dataset
Distillation | null | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Model-based deep learning has achieved astounding successes due in part to
the availability of large-scale real-world data. However, processing such
massive amounts of data comes at a considerable cost in terms of computations,
storage, training and the search for good neural architectures. Dataset
distillation has thus recently come to the fore. This paradigm involves
distilling information from large real-world datasets into tiny and compact
synthetic datasets such that processing the latter ideally yields similar
performances as the former. State-of-the-art methods primarily rely on learning
the synthetic dataset by matching the gradients obtained during training
between the real and synthetic data. However, these gradient-matching methods
suffer from the so-called accumulated trajectory error caused by the
discrepancy between the distillation and subsequent evaluation. To mitigate the
adverse impact of this accumulated trajectory error, we propose a novel
approach that encourages the optimization algorithm to seek a flat trajectory.
We show that the weights trained on synthetic data are robust against the
accumulated errors perturbations with the regularization towards the flat
trajectory. Our method, called Flat Trajectory Distillation (FTD), is shown to
boost the performance of gradient-matching methods by up to 4.7% on a subset of
images of the ImageNet dataset with higher resolution images. We also validate
the effectiveness and generalizability of our method with datasets of different
resolutions and demonstrate its applicability to neural architecture search.
Code is available at https://github.com/AngusDujw/FTD-distillation.
| [
{
"version": "v1",
"created": "Sun, 20 Nov 2022 15:49:11 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Nov 2022 02:26:42 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 03:57:25 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Du",
"Jiawei",
""
],
[
"Jiang",
"Yidi",
""
],
[
"Tan",
"Vincent Y. F.",
""
],
[
"Zhou",
"Joey Tianyi",
""
],
[
"Li",
"Haizhou",
""
]
] | TITLE: Minimizing the Accumulated Trajectory Error to Improve Dataset
Distillation
ABSTRACT: Model-based deep learning has achieved astounding successes due in part to
the availability of large-scale real-world data. However, processing such
massive amounts of data comes at a considerable cost in terms of computations,
storage, training and the search for good neural architectures. Dataset
distillation has thus recently come to the fore. This paradigm involves
distilling information from large real-world datasets into tiny and compact
synthetic datasets such that processing the latter ideally yields similar
performances as the former. State-of-the-art methods primarily rely on learning
the synthetic dataset by matching the gradients obtained during training
between the real and synthetic data. However, these gradient-matching methods
suffer from the so-called accumulated trajectory error caused by the
discrepancy between the distillation and subsequent evaluation. To mitigate the
adverse impact of this accumulated trajectory error, we propose a novel
approach that encourages the optimization algorithm to seek a flat trajectory.
We show that the weights trained on synthetic data are robust against the
accumulated errors perturbations with the regularization towards the flat
trajectory. Our method, called Flat Trajectory Distillation (FTD), is shown to
boost the performance of gradient-matching methods by up to 4.7% on a subset of
images of the ImageNet dataset with higher resolution images. We also validate
the effectiveness and generalizability of our method with datasets of different
resolutions and demonstrate its applicability to neural architecture search.
Code is available at https://github.com/AngusDujw/FTD-distillation.
| no_new_dataset | 0.947478 |
2211.11177 | Shitao Tang | Shitao Tang, Sicong Tang, Andrea Tagliasacchi, Ping Tan and Yasutaka
Furukawa | NeuMap: Neural Coordinate Mapping by Auto-Transdecoder for Camera
Localization | CVPR2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper presents an end-to-end neural mapping method for camera
localization, dubbed NeuMap, encoding a whole scene into a grid of latent
codes, with which a Transformer-based auto-decoder regresses 3D coordinates of
query pixels. State-of-the-art feature matching methods require each scene to
be stored as a 3D point cloud with per-point features, consuming several
gigabytes of storage per scene. While compression is possible, performance
drops significantly at high compression rates. Conversely, coordinate
regression methods achieve high compression by storing scene information in a
neural network but suffer from reduced robustness. NeuMap combines the
advantages of both approaches by utilizing 1) learnable latent codes for
efficient scene representation and 2) a scene-agnostic Transformer-based
auto-decoder to infer coordinates for query pixels. This scene-agnostic network
design learns robust matching priors from large-scale data and enables rapid
optimization of codes for new scenes while keeping the network weights fixed.
Extensive evaluations on five benchmarks show that NeuMap significantly
outperforms other coordinate regression methods and achieves comparable
performance to feature matching methods while requiring a much smaller scene
representation size. For example, NeuMap achieves 39.1% accuracy in the Aachen
night benchmark with only 6MB of data, whereas alternative methods require
100MB or several gigabytes and fail completely under high compression settings.
The codes are available at https://github.com/Tangshitao/NeuMap
| [
{
"version": "v1",
"created": "Mon, 21 Nov 2022 04:46:22 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 06:22:15 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Tang",
"Shitao",
""
],
[
"Tang",
"Sicong",
""
],
[
"Tagliasacchi",
"Andrea",
""
],
[
"Tan",
"Ping",
""
],
[
"Furukawa",
"Yasutaka",
""
]
] | TITLE: NeuMap: Neural Coordinate Mapping by Auto-Transdecoder for Camera
Localization
ABSTRACT: This paper presents an end-to-end neural mapping method for camera
localization, dubbed NeuMap, encoding a whole scene into a grid of latent
codes, with which a Transformer-based auto-decoder regresses 3D coordinates of
query pixels. State-of-the-art feature matching methods require each scene to
be stored as a 3D point cloud with per-point features, consuming several
gigabytes of storage per scene. While compression is possible, performance
drops significantly at high compression rates. Conversely, coordinate
regression methods achieve high compression by storing scene information in a
neural network but suffer from reduced robustness. NeuMap combines the
advantages of both approaches by utilizing 1) learnable latent codes for
efficient scene representation and 2) a scene-agnostic Transformer-based
auto-decoder to infer coordinates for query pixels. This scene-agnostic network
design learns robust matching priors from large-scale data and enables rapid
optimization of codes for new scenes while keeping the network weights fixed.
Extensive evaluations on five benchmarks show that NeuMap significantly
outperforms other coordinate regression methods and achieves comparable
performance to feature matching methods while requiring a much smaller scene
representation size. For example, NeuMap achieves 39.1% accuracy in the Aachen
night benchmark with only 6MB of data, whereas alternative methods require
100MB or several gigabytes and fail completely under high compression settings.
The codes are available at https://github.com/Tangshitao/NeuMap
| no_new_dataset | 0.951233 |
2211.11448 | HongYu Liu | Hongyu Liu and Yibing Song and Qifeng Chen | Delving StyleGAN Inversion for Image Editing: A Foundation Latent Space
Viewpoint | Accepted by CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | GAN inversion and editing via StyleGAN maps an input image into the embedding
spaces ($\mathcal{W}$, $\mathcal{W^+}$, and $\mathcal{F}$) to simultaneously
maintain image fidelity and meaningful manipulation. From latent space
$\mathcal{W}$ to extended latent space $\mathcal{W^+}$ to feature space
$\mathcal{F}$ in StyleGAN, the editability of GAN inversion decreases while its
reconstruction quality increases. Recent GAN inversion methods typically
explore $\mathcal{W^+}$ and $\mathcal{F}$ rather than $\mathcal{W}$ to improve
reconstruction fidelity while maintaining editability. As $\mathcal{W^+}$ and
$\mathcal{F}$ are derived from $\mathcal{W}$ that is essentially the foundation
latent space of StyleGAN, these GAN inversion methods focusing on
$\mathcal{W^+}$ and $\mathcal{F}$ spaces could be improved by stepping back to
$\mathcal{W}$. In this work, we propose to first obtain the precise latent code
in foundation latent space $\mathcal{W}$. We introduce contrastive learning to
align $\mathcal{W}$ and the image space for precise latent code discovery. %The
obtaining process is by using contrastive learning to align $\mathcal{W}$ and
the image space. Then, we leverage a cross-attention encoder to transform the
obtained latent code in $\mathcal{W}$ into $\mathcal{W^+}$ and $\mathcal{F}$,
accordingly. Our experiments show that our exploration of the foundation latent
space $\mathcal{W}$ improves the representation ability of latent codes in
$\mathcal{W^+}$ and features in $\mathcal{F}$, which yields state-of-the-art
reconstruction fidelity and editability results on the standard benchmarks.
Project page: https://kumapowerliu.github.io/CLCAE.
| [
{
"version": "v1",
"created": "Mon, 21 Nov 2022 13:35:32 GMT"
},
{
"version": "v2",
"created": "Fri, 10 Mar 2023 05:45:28 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 18:25:15 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Liu",
"Hongyu",
""
],
[
"Song",
"Yibing",
""
],
[
"Chen",
"Qifeng",
""
]
] | TITLE: Delving StyleGAN Inversion for Image Editing: A Foundation Latent Space
Viewpoint
ABSTRACT: GAN inversion and editing via StyleGAN maps an input image into the embedding
spaces ($\mathcal{W}$, $\mathcal{W^+}$, and $\mathcal{F}$) to simultaneously
maintain image fidelity and meaningful manipulation. From latent space
$\mathcal{W}$ to extended latent space $\mathcal{W^+}$ to feature space
$\mathcal{F}$ in StyleGAN, the editability of GAN inversion decreases while its
reconstruction quality increases. Recent GAN inversion methods typically
explore $\mathcal{W^+}$ and $\mathcal{F}$ rather than $\mathcal{W}$ to improve
reconstruction fidelity while maintaining editability. As $\mathcal{W^+}$ and
$\mathcal{F}$ are derived from $\mathcal{W}$ that is essentially the foundation
latent space of StyleGAN, these GAN inversion methods focusing on
$\mathcal{W^+}$ and $\mathcal{F}$ spaces could be improved by stepping back to
$\mathcal{W}$. In this work, we propose to first obtain the precise latent code
in foundation latent space $\mathcal{W}$. We introduce contrastive learning to
align $\mathcal{W}$ and the image space for precise latent code discovery. %The
obtaining process is by using contrastive learning to align $\mathcal{W}$ and
the image space. Then, we leverage a cross-attention encoder to transform the
obtained latent code in $\mathcal{W}$ into $\mathcal{W^+}$ and $\mathcal{F}$,
accordingly. Our experiments show that our exploration of the foundation latent
space $\mathcal{W}$ improves the representation ability of latent codes in
$\mathcal{W^+}$ and features in $\mathcal{F}$, which yields state-of-the-art
reconstruction fidelity and editability results on the standard benchmarks.
Project page: https://kumapowerliu.github.io/CLCAE.
| no_new_dataset | 0.943243 |
2211.11635 | Aochuan Chen | Aochuan Chen, Yuguang Yao, Pin-Yu Chen, Yihua Zhang, Sijia Liu | Understanding and Improving Visual Prompting: A Label-Mapping
Perspective | CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We revisit and advance visual prompting (VP), an input prompting technique
for vision tasks. VP can reprogram a fixed, pre-trained source model to
accomplish downstream tasks in the target domain by simply incorporating
universal prompts (in terms of input perturbation patterns) into downstream
data points. Yet, it remains elusive why VP stays effective even given a
ruleless label mapping (LM) between the source classes and the target classes.
Inspired by the above, we ask: How is LM interrelated with VP? And how to
exploit such a relationship to improve its accuracy on target tasks? We peer
into the influence of LM on VP and provide an affirmative answer that a better
'quality' of LM (assessed by mapping precision and explanation) can
consistently improve the effectiveness of VP. This is in contrast to the prior
art where the factor of LM was missing. To optimize LM, we propose a new VP
framework, termed ILM-VP (iterative label mapping-based visual prompting),
which automatically re-maps the source labels to the target labels and
progressively improves the target task accuracy of VP. Further, when using a
contrastive language-image pretrained (CLIP) model, we propose to integrate an
LM process to assist the text prompt selection of CLIP and to improve the
target task accuracy. Extensive experiments demonstrate that our proposal
significantly outperforms state-of-the-art VP methods. As highlighted below, we
show that when reprogramming an ImageNet-pretrained ResNet-18 to 13 target
tasks, our method outperforms baselines by a substantial margin, e.g., 7.9% and
6.7% accuracy improvements in transfer learning to the target Flowers102 and
CIFAR100 datasets. Besides, our proposal on CLIP-based VP provides 13.7% and
7.1% accuracy improvements on Flowers102 and DTD respectively. Our code is
available at https://github.com/OPTML-Group/ILM-VP.
| [
{
"version": "v1",
"created": "Mon, 21 Nov 2022 16:49:47 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Mar 2023 14:34:03 GMT"
},
{
"version": "v3",
"created": "Fri, 10 Mar 2023 22:54:45 GMT"
},
{
"version": "v4",
"created": "Thu, 23 Mar 2023 14:02:21 GMT"
},
{
"version": "v5",
"created": "Fri, 24 Mar 2023 18:06:06 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Chen",
"Aochuan",
""
],
[
"Yao",
"Yuguang",
""
],
[
"Chen",
"Pin-Yu",
""
],
[
"Zhang",
"Yihua",
""
],
[
"Liu",
"Sijia",
""
]
] | TITLE: Understanding and Improving Visual Prompting: A Label-Mapping
Perspective
ABSTRACT: We revisit and advance visual prompting (VP), an input prompting technique
for vision tasks. VP can reprogram a fixed, pre-trained source model to
accomplish downstream tasks in the target domain by simply incorporating
universal prompts (in terms of input perturbation patterns) into downstream
data points. Yet, it remains elusive why VP stays effective even given a
ruleless label mapping (LM) between the source classes and the target classes.
Inspired by the above, we ask: How is LM interrelated with VP? And how to
exploit such a relationship to improve its accuracy on target tasks? We peer
into the influence of LM on VP and provide an affirmative answer that a better
'quality' of LM (assessed by mapping precision and explanation) can
consistently improve the effectiveness of VP. This is in contrast to the prior
art where the factor of LM was missing. To optimize LM, we propose a new VP
framework, termed ILM-VP (iterative label mapping-based visual prompting),
which automatically re-maps the source labels to the target labels and
progressively improves the target task accuracy of VP. Further, when using a
contrastive language-image pretrained (CLIP) model, we propose to integrate an
LM process to assist the text prompt selection of CLIP and to improve the
target task accuracy. Extensive experiments demonstrate that our proposal
significantly outperforms state-of-the-art VP methods. As highlighted below, we
show that when reprogramming an ImageNet-pretrained ResNet-18 to 13 target
tasks, our method outperforms baselines by a substantial margin, e.g., 7.9% and
6.7% accuracy improvements in transfer learning to the target Flowers102 and
CIFAR100 datasets. Besides, our proposal on CLIP-based VP provides 13.7% and
7.1% accuracy improvements on Flowers102 and DTD respectively. Our code is
available at https://github.com/OPTML-Group/ILM-VP.
| no_new_dataset | 0.949809 |
2211.11646 | Junkai Huang | Benran Hu, Junkai Huang, Yichen Liu, Yu-Wing Tai, Chi-Keung Tang | NeRF-RPN: A general framework for object detection in NeRFs | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents the first significant object detection framework,
NeRF-RPN, which directly operates on NeRF. Given a pre-trained NeRF model,
NeRF-RPN aims to detect all bounding boxes of objects in a scene. By exploiting
a novel voxel representation that incorporates multi-scale 3D neural volumetric
features, we demonstrate it is possible to regress the 3D bounding boxes of
objects in NeRF directly without rendering the NeRF at any viewpoint. NeRF-RPN
is a general framework and can be applied to detect objects without class
labels. We experimented NeRF-RPN with various backbone architectures, RPN head
designs and loss functions. All of them can be trained in an end-to-end manner
to estimate high quality 3D bounding boxes. To facilitate future research in
object detection for NeRF, we built a new benchmark dataset which consists of
both synthetic and real-world data with careful labeling and clean up. Code and
dataset are available at https://github.com/lyclyc52/NeRF_RPN.
| [
{
"version": "v1",
"created": "Mon, 21 Nov 2022 17:02:01 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Nov 2022 01:55:00 GMT"
},
{
"version": "v3",
"created": "Mon, 27 Mar 2023 16:40:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Hu",
"Benran",
""
],
[
"Huang",
"Junkai",
""
],
[
"Liu",
"Yichen",
""
],
[
"Tai",
"Yu-Wing",
""
],
[
"Tang",
"Chi-Keung",
""
]
] | TITLE: NeRF-RPN: A general framework for object detection in NeRFs
ABSTRACT: This paper presents the first significant object detection framework,
NeRF-RPN, which directly operates on NeRF. Given a pre-trained NeRF model,
NeRF-RPN aims to detect all bounding boxes of objects in a scene. By exploiting
a novel voxel representation that incorporates multi-scale 3D neural volumetric
features, we demonstrate it is possible to regress the 3D bounding boxes of
objects in NeRF directly without rendering the NeRF at any viewpoint. NeRF-RPN
is a general framework and can be applied to detect objects without class
labels. We experimented NeRF-RPN with various backbone architectures, RPN head
designs and loss functions. All of them can be trained in an end-to-end manner
to estimate high quality 3D bounding boxes. To facilitate future research in
object detection for NeRF, we built a new benchmark dataset which consists of
both synthetic and real-world data with careful labeling and clean up. Code and
dataset are available at https://github.com/lyclyc52/NeRF_RPN.
| new_dataset | 0.957358 |
2211.12425 | Bo Dang | Bo Dang, Yansheng Li, Yongjun Zhang, Jiayi Ma | Progressive Learning with Cross-Window Consistency for Semi-Supervised
Semantic Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-supervised semantic segmentation focuses on the exploration of a small
amount of labeled data and a large amount of unlabeled data, which is more in
line with the demands of real-world image understanding applications. However,
it is still hindered by the inability to fully and effectively leverage
unlabeled images. In this paper, we reveal that cross-window consistency (CWC)
is helpful in comprehensively extracting auxiliary supervision from unlabeled
data. Additionally, we propose a novel CWC-driven progressive learning
framework to optimize the deep network by mining weak-to-strong constraints
from massive unlabeled data. More specifically, this paper presents a biased
cross-window consistency (BCC) loss with an importance factor, which helps the
deep network explicitly constrain confidence maps from overlapping regions in
different windows to maintain semantic consistency with larger contexts. In
addition, we propose a dynamic pseudo-label memory bank (DPM) to provide
high-consistency and high-reliability pseudo-labels to further optimize the
network. Extensive experiments on three representative datasets of urban views,
medical scenarios, and satellite scenes demonstrate our framework consistently
outperforms the state-of-the-art methods with a large margin. Code will be
available publicly.
| [
{
"version": "v1",
"created": "Tue, 22 Nov 2022 17:31:43 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 14:26:49 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Dang",
"Bo",
""
],
[
"Li",
"Yansheng",
""
],
[
"Zhang",
"Yongjun",
""
],
[
"Ma",
"Jiayi",
""
]
] | TITLE: Progressive Learning with Cross-Window Consistency for Semi-Supervised
Semantic Segmentation
ABSTRACT: Semi-supervised semantic segmentation focuses on the exploration of a small
amount of labeled data and a large amount of unlabeled data, which is more in
line with the demands of real-world image understanding applications. However,
it is still hindered by the inability to fully and effectively leverage
unlabeled images. In this paper, we reveal that cross-window consistency (CWC)
is helpful in comprehensively extracting auxiliary supervision from unlabeled
data. Additionally, we propose a novel CWC-driven progressive learning
framework to optimize the deep network by mining weak-to-strong constraints
from massive unlabeled data. More specifically, this paper presents a biased
cross-window consistency (BCC) loss with an importance factor, which helps the
deep network explicitly constrain confidence maps from overlapping regions in
different windows to maintain semantic consistency with larger contexts. In
addition, we propose a dynamic pseudo-label memory bank (DPM) to provide
high-consistency and high-reliability pseudo-labels to further optimize the
network. Extensive experiments on three representative datasets of urban views,
medical scenarios, and satellite scenes demonstrate our framework consistently
outperforms the state-of-the-art methods with a large margin. Code will be
available publicly.
| no_new_dataset | 0.943452 |
2211.13190 | Nikolai Kalischek | Nikolai Kalischek, Rodrigo C. Daudt, Torben Peters, Reinhard Furrer,
Jan D. Wegner, Konrad Schindler | BiasBed -- Rigorous Texture Bias Evaluation | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The well-documented presence of texture bias in modern convolutional neural
networks has led to a plethora of algorithms that promote an emphasis on shape
cues, often to support generalization to new domains. Yet, common datasets,
benchmarks and general model selection strategies are missing, and there is no
agreed, rigorous evaluation protocol. In this paper, we investigate
difficulties and limitations when training networks with reduced texture bias.
In particular, we also show that proper evaluation and meaningful comparisons
between methods are not trivial. We introduce BiasBed, a testbed for texture-
and style-biased training, including multiple datasets and a range of existing
algorithms. It comes with an extensive evaluation protocol that includes
rigorous hypothesis testing to gauge the significance of the results, despite
the considerable training instability of some style bias methods. Our extensive
experiments, shed new light on the need for careful, statistically founded
evaluation protocols for style bias (and beyond). E.g., we find that some
algorithms proposed in the literature do not significantly mitigate the impact
of style bias at all. With the release of BiasBed, we hope to foster a common
understanding of consistent and meaningful comparisons, and consequently faster
progress towards learning methods free of texture bias. Code is available at
https://github.com/D1noFuzi/BiasBed
| [
{
"version": "v1",
"created": "Wed, 23 Nov 2022 18:22:59 GMT"
},
{
"version": "v2",
"created": "Thu, 1 Dec 2022 09:08:28 GMT"
},
{
"version": "v3",
"created": "Fri, 24 Mar 2023 20:21:52 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Kalischek",
"Nikolai",
""
],
[
"Daudt",
"Rodrigo C.",
""
],
[
"Peters",
"Torben",
""
],
[
"Furrer",
"Reinhard",
""
],
[
"Wegner",
"Jan D.",
""
],
[
"Schindler",
"Konrad",
""
]
] | TITLE: BiasBed -- Rigorous Texture Bias Evaluation
ABSTRACT: The well-documented presence of texture bias in modern convolutional neural
networks has led to a plethora of algorithms that promote an emphasis on shape
cues, often to support generalization to new domains. Yet, common datasets,
benchmarks and general model selection strategies are missing, and there is no
agreed, rigorous evaluation protocol. In this paper, we investigate
difficulties and limitations when training networks with reduced texture bias.
In particular, we also show that proper evaluation and meaningful comparisons
between methods are not trivial. We introduce BiasBed, a testbed for texture-
and style-biased training, including multiple datasets and a range of existing
algorithms. It comes with an extensive evaluation protocol that includes
rigorous hypothesis testing to gauge the significance of the results, despite
the considerable training instability of some style bias methods. Our extensive
experiments, shed new light on the need for careful, statistically founded
evaluation protocols for style bias (and beyond). E.g., we find that some
algorithms proposed in the literature do not significantly mitigate the impact
of style bias at all. With the release of BiasBed, we hope to foster a common
understanding of consistent and meaningful comparisons, and consequently faster
progress towards learning methods free of texture bias. Code is available at
https://github.com/D1noFuzi/BiasBed
| no_new_dataset | 0.937326 |
2211.13437 | Yatai Ji | Yatai Ji, Rongcheng Tu, Jie Jiang, Weijie Kong, Chengfei Cai, Wenzhe
Zhao, Hongfa Wang, Yujiu Yang, Wei Liu | Seeing What You Miss: Vision-Language Pre-training with Semantic
Completion Learning | CVPR 2023 accept | null | null | null | cs.CV cs.CL cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cross-modal alignment is essential for vision-language pre-training (VLP)
models to learn the correct corresponding information across different
modalities. For this purpose, inspired by the success of masked language
modeling (MLM) tasks in the NLP pre-training area, numerous masked modeling
tasks have been proposed for VLP to further promote cross-modal interactions.
The core idea of previous masked modeling tasks is to focus on reconstructing
the masked tokens based on visible context for learning local-to-local
alignment. However, most of them pay little attention to the global semantic
features generated for the masked data, resulting in a limited cross-modal
alignment ability of global representations. Therefore, in this paper, we
propose a novel Semantic Completion Learning (SCL) task, complementary to
existing masked modeling tasks, to facilitate global-to-local alignment.
Specifically, the SCL task complements the missing semantics of masked data by
capturing the corresponding information from the other modality, promoting
learning more representative global features which have a great impact on the
performance of downstream tasks. Moreover, we present a flexible vision
encoder, which enables our model to perform image-text and video-text
multimodal tasks simultaneously. Experimental results show that our proposed
method obtains state-of-the-art performance on various vision-language
benchmarks, such as visual question answering, image-text retrieval, and
video-text retrieval.
| [
{
"version": "v1",
"created": "Thu, 24 Nov 2022 06:39:16 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 13:59:36 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ji",
"Yatai",
""
],
[
"Tu",
"Rongcheng",
""
],
[
"Jiang",
"Jie",
""
],
[
"Kong",
"Weijie",
""
],
[
"Cai",
"Chengfei",
""
],
[
"Zhao",
"Wenzhe",
""
],
[
"Wang",
"Hongfa",
""
],
[
"Yang",
"Yujiu",
""
],
[
"Liu",
"Wei",
""
]
] | TITLE: Seeing What You Miss: Vision-Language Pre-training with Semantic
Completion Learning
ABSTRACT: Cross-modal alignment is essential for vision-language pre-training (VLP)
models to learn the correct corresponding information across different
modalities. For this purpose, inspired by the success of masked language
modeling (MLM) tasks in the NLP pre-training area, numerous masked modeling
tasks have been proposed for VLP to further promote cross-modal interactions.
The core idea of previous masked modeling tasks is to focus on reconstructing
the masked tokens based on visible context for learning local-to-local
alignment. However, most of them pay little attention to the global semantic
features generated for the masked data, resulting in a limited cross-modal
alignment ability of global representations. Therefore, in this paper, we
propose a novel Semantic Completion Learning (SCL) task, complementary to
existing masked modeling tasks, to facilitate global-to-local alignment.
Specifically, the SCL task complements the missing semantics of masked data by
capturing the corresponding information from the other modality, promoting
learning more representative global features which have a great impact on the
performance of downstream tasks. Moreover, we present a flexible vision
encoder, which enables our model to perform image-text and video-text
multimodal tasks simultaneously. Experimental results show that our proposed
method obtains state-of-the-art performance on various vision-language
benchmarks, such as visual question answering, image-text retrieval, and
video-text retrieval.
| no_new_dataset | 0.942771 |
2211.14905 | Sauradip Nag | Sauradip Nag, Mengmeng Xu, Xiatian Zhu, Juan-Manuel Perez-Rua, Bernard
Ghanem, Yi-Zhe Song and Tao Xiang | Multi-Modal Few-Shot Temporal Action Detection | Technical Report | null | null | null | cs.CV cs.AI cs.CL cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-shot (FS) and zero-shot (ZS) learning are two different approaches for
scaling temporal action detection (TAD) to new classes. The former adapts a
pretrained vision model to a new task represented by as few as a single video
per class, whilst the latter requires no training examples by exploiting a
semantic description of the new class. In this work, we introduce a new
multi-modality few-shot (MMFS) TAD problem, which can be considered as a
marriage of FS-TAD and ZS-TAD by leveraging few-shot support videos and new
class names jointly. To tackle this problem, we further introduce a novel
MUlti-modality PromPt mETa-learning (MUPPET) method. This is enabled by
efficiently bridging pretrained vision and language models whilst maximally
reusing already learned capacity. Concretely, we construct multi-modal prompts
by mapping support videos into the textual token space of a vision-language
model using a meta-learned adapter-equipped visual semantics tokenizer. To
tackle large intra-class variation, we further design a query feature
regulation scheme. Extensive experiments on ActivityNetv1.3 and THUMOS14
demonstrate that our MUPPET outperforms state-of-the-art alternative methods,
often by a large margin. We also show that our MUPPET can be easily extended to
tackle the few-shot object detection problem and again achieves the
state-of-the-art performance on MS-COCO dataset. The code will be available in
https://github.com/sauradip/MUPPET
| [
{
"version": "v1",
"created": "Sun, 27 Nov 2022 18:13:05 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 08:39:13 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Nag",
"Sauradip",
""
],
[
"Xu",
"Mengmeng",
""
],
[
"Zhu",
"Xiatian",
""
],
[
"Perez-Rua",
"Juan-Manuel",
""
],
[
"Ghanem",
"Bernard",
""
],
[
"Song",
"Yi-Zhe",
""
],
[
"Xiang",
"Tao",
""
]
] | TITLE: Multi-Modal Few-Shot Temporal Action Detection
ABSTRACT: Few-shot (FS) and zero-shot (ZS) learning are two different approaches for
scaling temporal action detection (TAD) to new classes. The former adapts a
pretrained vision model to a new task represented by as few as a single video
per class, whilst the latter requires no training examples by exploiting a
semantic description of the new class. In this work, we introduce a new
multi-modality few-shot (MMFS) TAD problem, which can be considered as a
marriage of FS-TAD and ZS-TAD by leveraging few-shot support videos and new
class names jointly. To tackle this problem, we further introduce a novel
MUlti-modality PromPt mETa-learning (MUPPET) method. This is enabled by
efficiently bridging pretrained vision and language models whilst maximally
reusing already learned capacity. Concretely, we construct multi-modal prompts
by mapping support videos into the textual token space of a vision-language
model using a meta-learned adapter-equipped visual semantics tokenizer. To
tackle large intra-class variation, we further design a query feature
regulation scheme. Extensive experiments on ActivityNetv1.3 and THUMOS14
demonstrate that our MUPPET outperforms state-of-the-art alternative methods,
often by a large margin. We also show that our MUPPET can be easily extended to
tackle the few-shot object detection problem and again achieves the
state-of-the-art performance on MS-COCO dataset. The code will be available in
https://github.com/sauradip/MUPPET
| no_new_dataset | 0.947088 |
2211.15388 | Yufan Zhou | Yufan Zhou, Bingchen Liu, Yizhe Zhu, Xiao Yang, Changyou Chen, Jinhui
Xu | Shifted Diffusion for Text-to-image Generation | Accepted by CVPR 2023. Code at
https://github.com/drboog/Shifted_Diffusion | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | We present Corgi, a novel method for text-to-image generation. Corgi is based
on our proposed shifted diffusion model, which achieves better image embedding
generation from input text. Unlike the baseline diffusion model used in DALL-E
2, our method seamlessly encodes prior knowledge of the pre-trained CLIP model
in its diffusion process by designing a new initialization distribution and a
new transition step of the diffusion. Compared to the strong DALL-E 2 baseline,
our method performs better in generating image embedding from the text in terms
of both efficiency and effectiveness, resulting in better text-to-image
generation. Extensive large-scale experiments are conducted and evaluated in
terms of both quantitative measures and human evaluation, indicating a stronger
generation ability of our method compared to existing ones. Furthermore, our
model enables semi-supervised and language-free training for text-to-image
generation, where only part or none of the images in the training dataset have
an associated caption. Trained with only 1.7% of the images being captioned,
our semi-supervised model obtains FID results comparable to DALL-E 2 on
zero-shot text-to-image generation evaluated on MS-COCO. Corgi also achieves
new state-of-the-art results across different datasets on downstream
language-free text-to-image generation tasks, outperforming the previous
method, Lafite, by a large margin.
| [
{
"version": "v1",
"created": "Thu, 24 Nov 2022 03:25:04 GMT"
},
{
"version": "v2",
"created": "Thu, 23 Mar 2023 22:21:20 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhou",
"Yufan",
""
],
[
"Liu",
"Bingchen",
""
],
[
"Zhu",
"Yizhe",
""
],
[
"Yang",
"Xiao",
""
],
[
"Chen",
"Changyou",
""
],
[
"Xu",
"Jinhui",
""
]
] | TITLE: Shifted Diffusion for Text-to-image Generation
ABSTRACT: We present Corgi, a novel method for text-to-image generation. Corgi is based
on our proposed shifted diffusion model, which achieves better image embedding
generation from input text. Unlike the baseline diffusion model used in DALL-E
2, our method seamlessly encodes prior knowledge of the pre-trained CLIP model
in its diffusion process by designing a new initialization distribution and a
new transition step of the diffusion. Compared to the strong DALL-E 2 baseline,
our method performs better in generating image embedding from the text in terms
of both efficiency and effectiveness, resulting in better text-to-image
generation. Extensive large-scale experiments are conducted and evaluated in
terms of both quantitative measures and human evaluation, indicating a stronger
generation ability of our method compared to existing ones. Furthermore, our
model enables semi-supervised and language-free training for text-to-image
generation, where only part or none of the images in the training dataset have
an associated caption. Trained with only 1.7% of the images being captioned,
our semi-supervised model obtains FID results comparable to DALL-E 2 on
zero-shot text-to-image generation evaluated on MS-COCO. Corgi also achieves
new state-of-the-art results across different datasets on downstream
language-free text-to-image generation tasks, outperforming the previous
method, Lafite, by a large margin.
| no_new_dataset | 0.950778 |
2211.17174 | Frederik Pahde | Frederik Pahde, Galip \"Umit Yolcu, Alexander Binder, Wojciech Samek,
Sebastian Lapuschkin | Optimizing Explanations by Network Canonization and Hyperparameter
Search | null | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Explainable AI (XAI) is slowly becoming a key component for many AI
applications. Rule-based and modified backpropagation XAI approaches however
often face challenges when being applied to modern model architectures
including innovative layer building blocks, which is caused by two reasons.
Firstly, the high flexibility of rule-based XAI methods leads to numerous
potential parameterizations. Secondly, many XAI methods break the
implementation-invariance axiom because they struggle with certain model
components, e.g., BatchNorm layers. The latter can be addressed with model
canonization, which is the process of re-structuring the model to disregard
problematic components without changing the underlying function. While model
canonization is straightforward for simple architectures (e.g., VGG, ResNet),
it can be challenging for more complex and highly interconnected models (e.g.,
DenseNet). Moreover, there is only little quantifiable evidence that model
canonization is beneficial for XAI. In this work, we propose canonizations for
currently relevant model blocks applicable to popular deep neural network
architectures,including VGG, ResNet, EfficientNet, DenseNets, as well as
Relation Networks. We further suggest a XAI evaluation framework with which we
quantify and compare the effect sof model canonization for various XAI methods
in image classification tasks on the Pascal-VOC and ILSVRC2017 datasets, as
well as for Visual Question Answering using CLEVR-XAI. Moreover, addressing the
former issue outlined above, we demonstrate how our evaluation framework can be
applied to perform hyperparameter search for XAI methods to optimize the
quality of explanations.
| [
{
"version": "v1",
"created": "Wed, 30 Nov 2022 17:17:55 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 09:42:13 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Pahde",
"Frederik",
""
],
[
"Yolcu",
"Galip Ümit",
""
],
[
"Binder",
"Alexander",
""
],
[
"Samek",
"Wojciech",
""
],
[
"Lapuschkin",
"Sebastian",
""
]
] | TITLE: Optimizing Explanations by Network Canonization and Hyperparameter
Search
ABSTRACT: Explainable AI (XAI) is slowly becoming a key component for many AI
applications. Rule-based and modified backpropagation XAI approaches however
often face challenges when being applied to modern model architectures
including innovative layer building blocks, which is caused by two reasons.
Firstly, the high flexibility of rule-based XAI methods leads to numerous
potential parameterizations. Secondly, many XAI methods break the
implementation-invariance axiom because they struggle with certain model
components, e.g., BatchNorm layers. The latter can be addressed with model
canonization, which is the process of re-structuring the model to disregard
problematic components without changing the underlying function. While model
canonization is straightforward for simple architectures (e.g., VGG, ResNet),
it can be challenging for more complex and highly interconnected models (e.g.,
DenseNet). Moreover, there is only little quantifiable evidence that model
canonization is beneficial for XAI. In this work, we propose canonizations for
currently relevant model blocks applicable to popular deep neural network
architectures,including VGG, ResNet, EfficientNet, DenseNets, as well as
Relation Networks. We further suggest a XAI evaluation framework with which we
quantify and compare the effect sof model canonization for various XAI methods
in image classification tasks on the Pascal-VOC and ILSVRC2017 datasets, as
well as for Visual Question Answering using CLEVR-XAI. Moreover, addressing the
former issue outlined above, we demonstrate how our evaluation framework can be
applied to perform hyperparameter search for XAI methods to optimize the
quality of explanations.
| no_new_dataset | 0.942612 |
2212.00785 | Junbum Cha | Junbum Cha, Jonghwan Mun, Byungseok Roh | Learning to Generate Text-grounded Mask for Open-world Semantic
Segmentation from Only Image-Text Pairs | CVPR 2023 camera-ready | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We tackle open-world semantic segmentation, which aims at learning to segment
arbitrary visual concepts in images, by using only image-text pairs without
dense annotations. Existing open-world segmentation methods have shown
impressive advances by employing contrastive learning (CL) to learn diverse
visual concepts and transferring the learned image-level understanding to the
segmentation task. However, these CL-based methods suffer from a train-test
discrepancy, since it only considers image-text alignment during training,
whereas segmentation requires region-text alignment during testing. In this
paper, we proposed a novel Text-grounded Contrastive Learning (TCL) framework
that enables a model to directly learn region-text alignment. Our method
generates a segmentation mask for a given text, extracts text-grounded image
embedding from the masked region, and aligns it with text embedding via TCL. By
learning region-text alignment directly, our framework encourages a model to
directly improve the quality of generated segmentation masks. In addition, for
a rigorous and fair comparison, we present a unified evaluation protocol with
widely used 8 semantic segmentation datasets. TCL achieves state-of-the-art
zero-shot segmentation performances with large margins in all datasets. Code is
available at https://github.com/kakaobrain/tcl.
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2022 18:59:03 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 11:16:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Cha",
"Junbum",
""
],
[
"Mun",
"Jonghwan",
""
],
[
"Roh",
"Byungseok",
""
]
] | TITLE: Learning to Generate Text-grounded Mask for Open-world Semantic
Segmentation from Only Image-Text Pairs
ABSTRACT: We tackle open-world semantic segmentation, which aims at learning to segment
arbitrary visual concepts in images, by using only image-text pairs without
dense annotations. Existing open-world segmentation methods have shown
impressive advances by employing contrastive learning (CL) to learn diverse
visual concepts and transferring the learned image-level understanding to the
segmentation task. However, these CL-based methods suffer from a train-test
discrepancy, since it only considers image-text alignment during training,
whereas segmentation requires region-text alignment during testing. In this
paper, we proposed a novel Text-grounded Contrastive Learning (TCL) framework
that enables a model to directly learn region-text alignment. Our method
generates a segmentation mask for a given text, extracts text-grounded image
embedding from the masked region, and aligns it with text embedding via TCL. By
learning region-text alignment directly, our framework encourages a model to
directly improve the quality of generated segmentation masks. In addition, for
a rigorous and fair comparison, we present a unified evaluation protocol with
widely used 8 semantic segmentation datasets. TCL achieves state-of-the-art
zero-shot segmentation performances with large margins in all datasets. Code is
available at https://github.com/kakaobrain/tcl.
| no_new_dataset | 0.949012 |
2212.03640 | Muhammad Maaz Mr | Hanoona Rasheed, Muhammad Uzair Khattak, Muhammad Maaz, Salman Khan,
Fahad Shahbaz Khan | Fine-tuned CLIP Models are Efficient Video Learners | Accepted at CVPR 2023 | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large-scale multi-modal training with image-text pairs imparts strong
generalization to CLIP model. Since training on a similar scale for videos is
infeasible, recent approaches focus on the effective transfer of image-based
CLIP to the video domain. In this pursuit, new parametric modules are added to
learn temporal information and inter-frame relationships which require
meticulous design efforts. Furthermore, when the resulting models are learned
on videos, they tend to overfit on the given task distribution and lack in
generalization aspect. This begs the following question: How to effectively
transfer image-level CLIP representations to videos? In this work, we show that
a simple Video Fine-tuned CLIP (ViFi-CLIP) baseline is generally sufficient to
bridge the domain gap from images to videos. Our qualitative analysis
illustrates that the frame-level processing from CLIP image-encoder followed by
feature pooling and similarity matching with corresponding text embeddings
helps in implicitly modeling the temporal cues within ViFi-CLIP. Such
fine-tuning helps the model to focus on scene dynamics, moving objects and
inter-object relationships. For low-data regimes where full fine-tuning is not
viable, we propose a `bridge and prompt' approach that first uses fine-tuning
to bridge the domain gap and then learns prompts on language and vision side to
adapt CLIP representations. We extensively evaluate this simple yet strong
baseline on zero-shot, base-to-novel generalization, few-shot and fully
supervised settings across five video benchmarks. Our code is available at
https://github.com/muzairkhattak/ViFi-CLIP.
| [
{
"version": "v1",
"created": "Tue, 6 Dec 2022 18:59:58 GMT"
},
{
"version": "v2",
"created": "Fri, 10 Mar 2023 15:18:18 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 11:40:16 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Rasheed",
"Hanoona",
""
],
[
"Khattak",
"Muhammad Uzair",
""
],
[
"Maaz",
"Muhammad",
""
],
[
"Khan",
"Salman",
""
],
[
"Khan",
"Fahad Shahbaz",
""
]
] | TITLE: Fine-tuned CLIP Models are Efficient Video Learners
ABSTRACT: Large-scale multi-modal training with image-text pairs imparts strong
generalization to CLIP model. Since training on a similar scale for videos is
infeasible, recent approaches focus on the effective transfer of image-based
CLIP to the video domain. In this pursuit, new parametric modules are added to
learn temporal information and inter-frame relationships which require
meticulous design efforts. Furthermore, when the resulting models are learned
on videos, they tend to overfit on the given task distribution and lack in
generalization aspect. This begs the following question: How to effectively
transfer image-level CLIP representations to videos? In this work, we show that
a simple Video Fine-tuned CLIP (ViFi-CLIP) baseline is generally sufficient to
bridge the domain gap from images to videos. Our qualitative analysis
illustrates that the frame-level processing from CLIP image-encoder followed by
feature pooling and similarity matching with corresponding text embeddings
helps in implicitly modeling the temporal cues within ViFi-CLIP. Such
fine-tuning helps the model to focus on scene dynamics, moving objects and
inter-object relationships. For low-data regimes where full fine-tuning is not
viable, we propose a `bridge and prompt' approach that first uses fine-tuning
to bridge the domain gap and then learns prompts on language and vision side to
adapt CLIP representations. We extensively evaluate this simple yet strong
baseline on zero-shot, base-to-novel generalization, few-shot and fully
supervised settings across five video benchmarks. Our code is available at
https://github.com/muzairkhattak/ViFi-CLIP.
| no_new_dataset | 0.948728 |
2212.05590 | Sheng Zhang | Sheng Zhang, Salman Khan, Zhiqiang Shen, Muzammal Naseer, Guangyi
Chen, Fahad Khan | PromptCAL: Contrastive Affinity Learning via Auxiliary Prompts for
Generalized Novel Category Discovery | Accepted by CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although existing semi-supervised learning models achieve remarkable success
in learning with unannotated in-distribution data, they mostly fail to learn on
unlabeled data sampled from novel semantic classes due to their closed-set
assumption. In this work, we target a pragmatic but under-explored Generalized
Novel Category Discovery (GNCD) setting. The GNCD setting aims to categorize
unlabeled training data coming from known and novel classes by leveraging the
information of partially labeled known classes. We propose a two-stage
Contrastive Affinity Learning method with auxiliary visual Prompts, dubbed
PromptCAL, to address this challenging problem. Our approach discovers reliable
pairwise sample affinities to learn better semantic clustering of both known
and novel classes for the class token and visual prompts. First, we propose a
discriminative prompt regularization loss to reinforce semantic
discriminativeness of prompt-adapted pre-trained vision transformer for refined
affinity relationships.Besides, we propose contrastive affinity learning to
calibrate semantic representations based on our iterative semi-supervised
affinity graph generation method for semantically-enhanced supervision.
Extensive experimental evaluation demonstrates that our PromptCAL method is
more effective in discovering novel classes even with limited annotations and
surpasses the current state-of-the-art on generic and fine-grained benchmarks
(e.g., with nearly 11% gain on CUB-200, and 9% on ImageNet-100) on overall
accuracy. Our code is available at https://github.com/sheng-eatamath/PromptCAL.
| [
{
"version": "v1",
"created": "Sun, 11 Dec 2022 20:06:14 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 10:30:22 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Sheng",
""
],
[
"Khan",
"Salman",
""
],
[
"Shen",
"Zhiqiang",
""
],
[
"Naseer",
"Muzammal",
""
],
[
"Chen",
"Guangyi",
""
],
[
"Khan",
"Fahad",
""
]
] | TITLE: PromptCAL: Contrastive Affinity Learning via Auxiliary Prompts for
Generalized Novel Category Discovery
ABSTRACT: Although existing semi-supervised learning models achieve remarkable success
in learning with unannotated in-distribution data, they mostly fail to learn on
unlabeled data sampled from novel semantic classes due to their closed-set
assumption. In this work, we target a pragmatic but under-explored Generalized
Novel Category Discovery (GNCD) setting. The GNCD setting aims to categorize
unlabeled training data coming from known and novel classes by leveraging the
information of partially labeled known classes. We propose a two-stage
Contrastive Affinity Learning method with auxiliary visual Prompts, dubbed
PromptCAL, to address this challenging problem. Our approach discovers reliable
pairwise sample affinities to learn better semantic clustering of both known
and novel classes for the class token and visual prompts. First, we propose a
discriminative prompt regularization loss to reinforce semantic
discriminativeness of prompt-adapted pre-trained vision transformer for refined
affinity relationships.Besides, we propose contrastive affinity learning to
calibrate semantic representations based on our iterative semi-supervised
affinity graph generation method for semantically-enhanced supervision.
Extensive experimental evaluation demonstrates that our PromptCAL method is
more effective in discovering novel classes even with limited annotations and
surpasses the current state-of-the-art on generic and fine-grained benchmarks
(e.g., with nearly 11% gain on CUB-200, and 9% on ImageNet-100) on overall
accuracy. Our code is available at https://github.com/sheng-eatamath/PromptCAL.
| no_new_dataset | 0.951142 |
2212.05973 | Yunsung Lee | Hyojun Go, Yunsung Lee, Jin-Young Kim, Seunghyun Lee, Myeongho Jeong,
Hyun Seung Lee, and Seungtaek Choi | Towards Practical Plug-and-Play Diffusion Models | CVPR 2023 camera-ready | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion-based generative models have achieved remarkable success in image
generation. Their guidance formulation allows an external model to
plug-and-play control the generation process for various tasks without
finetuning the diffusion model. However, the direct use of publicly available
off-the-shelf models for guidance fails due to their poor performance on noisy
inputs. For that, the existing practice is to fine-tune the guidance models
with labeled data corrupted with noises. In this paper, we argue that this
practice has limitations in two aspects: (1) performing on inputs with
extremely various noises is too hard for a single guidance model; (2)
collecting labeled datasets hinders scaling up for various tasks. To tackle the
limitations, we propose a novel strategy that leverages multiple experts where
each expert is specialized in a particular noise range and guides the reverse
process of the diffusion at its corresponding timesteps. However, as it is
infeasible to manage multiple networks and utilize labeled data, we present a
practical guidance framework termed Practical Plug-And-Play (PPAP), which
leverages parameter-efficient fine-tuning and data-free knowledge transfer. We
exhaustively conduct ImageNet class conditional generation experiments to show
that our method can successfully guide diffusion with small trainable
parameters and no labeled data. Finally, we show that image classifiers, depth
estimators, and semantic segmentation models can guide publicly available GLIDE
through our framework in a plug-and-play manner. Our code is available at
https://github.com/riiid/PPAP.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 15:29:46 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 13:17:48 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Go",
"Hyojun",
""
],
[
"Lee",
"Yunsung",
""
],
[
"Kim",
"Jin-Young",
""
],
[
"Lee",
"Seunghyun",
""
],
[
"Jeong",
"Myeongho",
""
],
[
"Lee",
"Hyun Seung",
""
],
[
"Choi",
"Seungtaek",
""
]
] | TITLE: Towards Practical Plug-and-Play Diffusion Models
ABSTRACT: Diffusion-based generative models have achieved remarkable success in image
generation. Their guidance formulation allows an external model to
plug-and-play control the generation process for various tasks without
finetuning the diffusion model. However, the direct use of publicly available
off-the-shelf models for guidance fails due to their poor performance on noisy
inputs. For that, the existing practice is to fine-tune the guidance models
with labeled data corrupted with noises. In this paper, we argue that this
practice has limitations in two aspects: (1) performing on inputs with
extremely various noises is too hard for a single guidance model; (2)
collecting labeled datasets hinders scaling up for various tasks. To tackle the
limitations, we propose a novel strategy that leverages multiple experts where
each expert is specialized in a particular noise range and guides the reverse
process of the diffusion at its corresponding timesteps. However, as it is
infeasible to manage multiple networks and utilize labeled data, we present a
practical guidance framework termed Practical Plug-And-Play (PPAP), which
leverages parameter-efficient fine-tuning and data-free knowledge transfer. We
exhaustively conduct ImageNet class conditional generation experiments to show
that our method can successfully guide diffusion with small trainable
parameters and no labeled data. Finally, we show that image classifiers, depth
estimators, and semantic segmentation models can guide publicly available GLIDE
through our framework in a plug-and-play manner. Our code is available at
https://github.com/riiid/PPAP.
| no_new_dataset | 0.947914 |
2212.07692 | Fran\c{c}ois Lecomte | Fran\c{c}ois Lecomte, Jean-Louis Dillenseger, St\'ephane Cotin | CNN-based real-time 2D-3D deformable registration from a single X-ray
projection | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Purpose: The purpose of this paper is to present a method for real-time 2D-3D
non-rigid registration using a single fluoroscopic image. Such a method can
find applications in surgery, interventional radiology and radiotherapy. By
estimating a three-dimensional displacement field from a 2D X-ray image,
anatomical structures segmented in the preoperative scan can be projected onto
the 2D image, thus providing a mixed reality view. Methods: A dataset composed
of displacement fields and 2D projections of the anatomy is generated from the
preoperative scan. From this dataset, a neural network is trained to recover
the unknown 3D displacement field from a single projection image. Results: Our
method is validated on lung 4D CT data at different stages of the lung
deformation. The training is performed on a 3D CT using random (non
domain-specific) diffeomorphic deformations, to which perturbations mimicking
the pose uncertainty are added. The model achieves a mean TRE over a series of
landmarks ranging from 2.3 to 5.5 mm depending on the amplitude of deformation.
Conclusion: In this paper, a CNN-based method for real-time 2D-3D non-rigid
registration is presented. This method is able to cope with pose estimation
uncertainties, making it applicable to actual clinical scenarios, such as lung
surgery, where the C-arm pose is planned before the intervention.
| [
{
"version": "v1",
"created": "Thu, 15 Dec 2022 09:57:19 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 14:22:54 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Lecomte",
"François",
""
],
[
"Dillenseger",
"Jean-Louis",
""
],
[
"Cotin",
"Stéphane",
""
]
] | TITLE: CNN-based real-time 2D-3D deformable registration from a single X-ray
projection
ABSTRACT: Purpose: The purpose of this paper is to present a method for real-time 2D-3D
non-rigid registration using a single fluoroscopic image. Such a method can
find applications in surgery, interventional radiology and radiotherapy. By
estimating a three-dimensional displacement field from a 2D X-ray image,
anatomical structures segmented in the preoperative scan can be projected onto
the 2D image, thus providing a mixed reality view. Methods: A dataset composed
of displacement fields and 2D projections of the anatomy is generated from the
preoperative scan. From this dataset, a neural network is trained to recover
the unknown 3D displacement field from a single projection image. Results: Our
method is validated on lung 4D CT data at different stages of the lung
deformation. The training is performed on a 3D CT using random (non
domain-specific) diffeomorphic deformations, to which perturbations mimicking
the pose uncertainty are added. The model achieves a mean TRE over a series of
landmarks ranging from 2.3 to 5.5 mm depending on the amplitude of deformation.
Conclusion: In this paper, a CNN-based method for real-time 2D-3D non-rigid
registration is presented. This method is able to cope with pose estimation
uncertainties, making it applicable to actual clinical scenarios, such as lung
surgery, where the C-arm pose is planned before the intervention.
| new_dataset | 0.974845 |
2212.08062 | Bowen Zhang | Bowen Zhang, Chenyang Qi, Pan Zhang, Bo Zhang, HsiangTao Wu, Dong
Chen, Qifeng Chen, Yong Wang, Fang Wen | MetaPortrait: Identity-Preserving Talking Head Generation with Fast
Personalized Adaptation | CVPR 2023, project page: https://meta-portrait.github.io | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we propose an ID-preserving talking head generation framework,
which advances previous methods in two aspects. First, as opposed to
interpolating from sparse flow, we claim that dense landmarks are crucial to
achieving accurate geometry-aware flow fields. Second, inspired by
face-swapping methods, we adaptively fuse the source identity during synthesis,
so that the network better preserves the key characteristics of the image
portrait. Although the proposed model surpasses prior generation fidelity on
established benchmarks, to further make the talking head generation qualified
for real usage, personalized fine-tuning is usually needed. However, this
process is rather computationally demanding that is unaffordable to standard
users. To solve this, we propose a fast adaptation model using a meta-learning
approach. The learned model can be adapted to a high-quality personalized model
as fast as 30 seconds. Last but not the least, a spatial-temporal enhancement
module is proposed to improve the fine details while ensuring temporal
coherency. Extensive experiments prove the significant superiority of our
approach over the state of the arts in both one-shot and personalized settings.
| [
{
"version": "v1",
"created": "Thu, 15 Dec 2022 18:59:33 GMT"
},
{
"version": "v2",
"created": "Sat, 17 Dec 2022 02:38:01 GMT"
},
{
"version": "v3",
"created": "Mon, 27 Mar 2023 02:16:13 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Bowen",
""
],
[
"Qi",
"Chenyang",
""
],
[
"Zhang",
"Pan",
""
],
[
"Zhang",
"Bo",
""
],
[
"Wu",
"HsiangTao",
""
],
[
"Chen",
"Dong",
""
],
[
"Chen",
"Qifeng",
""
],
[
"Wang",
"Yong",
""
],
[
"Wen",
"Fang",
""
]
] | TITLE: MetaPortrait: Identity-Preserving Talking Head Generation with Fast
Personalized Adaptation
ABSTRACT: In this work, we propose an ID-preserving talking head generation framework,
which advances previous methods in two aspects. First, as opposed to
interpolating from sparse flow, we claim that dense landmarks are crucial to
achieving accurate geometry-aware flow fields. Second, inspired by
face-swapping methods, we adaptively fuse the source identity during synthesis,
so that the network better preserves the key characteristics of the image
portrait. Although the proposed model surpasses prior generation fidelity on
established benchmarks, to further make the talking head generation qualified
for real usage, personalized fine-tuning is usually needed. However, this
process is rather computationally demanding that is unaffordable to standard
users. To solve this, we propose a fast adaptation model using a meta-learning
approach. The learned model can be adapted to a high-quality personalized model
as fast as 30 seconds. Last but not the least, a spatial-temporal enhancement
module is proposed to improve the fine details while ensuring temporal
coherency. Extensive experiments prove the significant superiority of our
approach over the state of the arts in both one-shot and personalized settings.
| no_new_dataset | 0.946051 |
2212.11042 | Chun-Han Yao | Chun-Han Yao, Wei-Chih Hung, Yuanzhen Li, Michael Rubinstein,
Ming-Hsuan Yang, Varun Jampani | Hi-LASSIE: High-Fidelity Articulated Shape and Skeleton Discovery from
Sparse Image Ensemble | Project page: https://chhankyao.github.io/hi-lassie/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Automatically estimating 3D skeleton, shape, camera viewpoints, and part
articulation from sparse in-the-wild image ensembles is a severely
under-constrained and challenging problem. Most prior methods rely on
large-scale image datasets, dense temporal correspondence, or human annotations
like camera pose, 2D keypoints, and shape templates. We propose Hi-LASSIE,
which performs 3D articulated reconstruction from only 20-30 online images in
the wild without any user-defined shape or skeleton templates. We follow the
recent work of LASSIE that tackles a similar problem setting and make two
significant advances. First, instead of relying on a manually annotated 3D
skeleton, we automatically estimate a class-specific skeleton from the selected
reference image. Second, we improve the shape reconstructions with novel
instance-specific optimization strategies that allow reconstructions to
faithful fit on each instance while preserving the class-specific priors
learned across all images. Experiments on in-the-wild image ensembles show that
Hi-LASSIE obtains higher fidelity state-of-the-art 3D reconstructions despite
requiring minimum user input.
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2022 14:31:33 GMT"
},
{
"version": "v2",
"created": "Wed, 28 Dec 2022 12:53:23 GMT"
},
{
"version": "v3",
"created": "Thu, 19 Jan 2023 00:26:26 GMT"
},
{
"version": "v4",
"created": "Sat, 25 Mar 2023 17:59:59 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yao",
"Chun-Han",
""
],
[
"Hung",
"Wei-Chih",
""
],
[
"Li",
"Yuanzhen",
""
],
[
"Rubinstein",
"Michael",
""
],
[
"Yang",
"Ming-Hsuan",
""
],
[
"Jampani",
"Varun",
""
]
] | TITLE: Hi-LASSIE: High-Fidelity Articulated Shape and Skeleton Discovery from
Sparse Image Ensemble
ABSTRACT: Automatically estimating 3D skeleton, shape, camera viewpoints, and part
articulation from sparse in-the-wild image ensembles is a severely
under-constrained and challenging problem. Most prior methods rely on
large-scale image datasets, dense temporal correspondence, or human annotations
like camera pose, 2D keypoints, and shape templates. We propose Hi-LASSIE,
which performs 3D articulated reconstruction from only 20-30 online images in
the wild without any user-defined shape or skeleton templates. We follow the
recent work of LASSIE that tackles a similar problem setting and make two
significant advances. First, instead of relying on a manually annotated 3D
skeleton, we automatically estimate a class-specific skeleton from the selected
reference image. Second, we improve the shape reconstructions with novel
instance-specific optimization strategies that allow reconstructions to
faithful fit on each instance while preserving the class-specific priors
learned across all images. Experiments on in-the-wild image ensembles show that
Hi-LASSIE obtains higher fidelity state-of-the-art 3D reconstructions despite
requiring minimum user input.
| no_new_dataset | 0.949995 |
2212.12053 | Dongdong Wang | Dongdong Wang and Boqing Gong and Liqiang Wang | On Calibrating Semantic Segmentation Models: Analyses and An Algorithm | Accepted to CVPR-2023 (8 pages, 4 figures) | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of semantic segmentation calibration. Lots of solutions
have been proposed to approach model miscalibration of confidence in image
classification. However, to date, confidence calibration research on semantic
segmentation is still limited. We provide a systematic study on the calibration
of semantic segmentation models and propose a simple yet effective approach.
First, we find that model capacity, crop size, multi-scale testing, and
prediction correctness have impact on calibration. Among them, prediction
correctness, especially misprediction, is more important to miscalibration due
to over-confidence. Next, we propose a simple, unifying, and effective
approach, namely selective scaling, by separating correct/incorrect prediction
for scaling and more focusing on misprediction logit smoothing. Then, we study
popular existing calibration methods and compare them with selective scaling on
semantic segmentation calibration. We conduct extensive experiments with a
variety of benchmarks on both in-domain and domain-shift calibration and show
that selective scaling consistently outperforms other methods.
| [
{
"version": "v1",
"created": "Thu, 22 Dec 2022 22:05:16 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Mar 2023 16:41:02 GMT"
},
{
"version": "v3",
"created": "Mon, 20 Mar 2023 15:22:59 GMT"
},
{
"version": "v4",
"created": "Sat, 25 Mar 2023 06:15:41 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Dongdong",
""
],
[
"Gong",
"Boqing",
""
],
[
"Wang",
"Liqiang",
""
]
] | TITLE: On Calibrating Semantic Segmentation Models: Analyses and An Algorithm
ABSTRACT: We study the problem of semantic segmentation calibration. Lots of solutions
have been proposed to approach model miscalibration of confidence in image
classification. However, to date, confidence calibration research on semantic
segmentation is still limited. We provide a systematic study on the calibration
of semantic segmentation models and propose a simple yet effective approach.
First, we find that model capacity, crop size, multi-scale testing, and
prediction correctness have impact on calibration. Among them, prediction
correctness, especially misprediction, is more important to miscalibration due
to over-confidence. Next, we propose a simple, unifying, and effective
approach, namely selective scaling, by separating correct/incorrect prediction
for scaling and more focusing on misprediction logit smoothing. Then, we study
popular existing calibration methods and compare them with selective scaling on
semantic segmentation calibration. We conduct extensive experiments with a
variety of benchmarks on both in-domain and domain-shift calibration and show
that selective scaling consistently outperforms other methods.
| no_new_dataset | 0.945147 |
2212.14694 | Divyanshi Dwivedi | Siddharth Patwardhan, Utso Majumder, Aditya Das Sarma, Mayukha Pal,
Divyanshi Dwivedi and Prasanta K. Panigrahi | Machine Learning as an Accurate Predictor for Percolation Threshold of
Diverse Networks | null | null | null | null | physics.soc-ph cs.LG cs.SI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The percolation threshold is an important measure to determine the inherent
rigidity of large networks. Predictors of the percolation threshold for large
networks are computationally intense to run, hence it is a necessity to develop
predictors of the percolation threshold of networks, that do not rely on
numerical simulations. We demonstrate the efficacy of five machine
learning-based regression techniques for the accurate prediction of the
percolation threshold. The dataset generated to train the machine learning
models contains a total of 777 real and synthetic networks. It consists of 5
statistical and structural properties of networks as features and the
numerically computed percolation threshold as the output attribute. We
establish that the machine learning models outperform three existing empirical
estimators of bond percolation threshold, and extend this experiment to predict
site and explosive percolation. Further, we compared the performance of our
models in predicting the percolation threshold using RMSE values. The gradient
boosting regressor, multilayer perceptron and random forests regression models
achieve the least RMSE values among considered models.
| [
{
"version": "v1",
"created": "Fri, 16 Dec 2022 20:01:46 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 18:15:55 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Patwardhan",
"Siddharth",
""
],
[
"Majumder",
"Utso",
""
],
[
"Sarma",
"Aditya Das",
""
],
[
"Pal",
"Mayukha",
""
],
[
"Dwivedi",
"Divyanshi",
""
],
[
"Panigrahi",
"Prasanta K.",
""
]
] | TITLE: Machine Learning as an Accurate Predictor for Percolation Threshold of
Diverse Networks
ABSTRACT: The percolation threshold is an important measure to determine the inherent
rigidity of large networks. Predictors of the percolation threshold for large
networks are computationally intense to run, hence it is a necessity to develop
predictors of the percolation threshold of networks, that do not rely on
numerical simulations. We demonstrate the efficacy of five machine
learning-based regression techniques for the accurate prediction of the
percolation threshold. The dataset generated to train the machine learning
models contains a total of 777 real and synthetic networks. It consists of 5
statistical and structural properties of networks as features and the
numerically computed percolation threshold as the output attribute. We
establish that the machine learning models outperform three existing empirical
estimators of bond percolation threshold, and extend this experiment to predict
site and explosive percolation. Further, we compared the performance of our
models in predicting the percolation threshold using RMSE values. The gradient
boosting regressor, multilayer perceptron and random forests regression models
achieve the least RMSE values among considered models.
| no_new_dataset | 0.894605 |
2301.00182 | Wenhao Wu | Wenhao Wu, Xiaohan Wang, Haipeng Luo, Jingdong Wang, Yi Yang, Wanli
Ouyang | Bidirectional Cross-Modal Knowledge Exploration for Video Recognition
with Pre-trained Vision-Language Models | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-language models (VLMs) pre-trained on large-scale image-text pairs
have demonstrated impressive transferability on various visual tasks.
Transferring knowledge from such powerful VLMs is a promising direction for
building effective video recognition models. However, current exploration in
this field is still limited. We believe that the greatest value of pre-trained
VLMs lies in building a bridge between visual and textual domains. In this
paper, we propose a novel framework called BIKE, which utilizes the cross-modal
bridge to explore bidirectional knowledge: i) We introduce the Video Attribute
Association mechanism, which leverages the Video-to-Text knowledge to generate
textual auxiliary attributes for complementing video recognition. ii) We also
present a Temporal Concept Spotting mechanism that uses the Text-to-Video
expertise to capture temporal saliency in a parameter-free manner, leading to
enhanced video representation. Extensive studies on six popular video datasets,
including Kinetics-400 & 600, UCF-101, HMDB-51, ActivityNet and Charades, show
that our method achieves state-of-the-art performance in various recognition
scenarios, such as general, zero-shot, and few-shot video recognition. Our best
model achieves a state-of-the-art accuracy of 88.6% on the challenging
Kinetics-400 using the released CLIP model. The code is available at
https://github.com/whwu95/BIKE .
| [
{
"version": "v1",
"created": "Sat, 31 Dec 2022 11:36:53 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 12:12:30 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wu",
"Wenhao",
""
],
[
"Wang",
"Xiaohan",
""
],
[
"Luo",
"Haipeng",
""
],
[
"Wang",
"Jingdong",
""
],
[
"Yang",
"Yi",
""
],
[
"Ouyang",
"Wanli",
""
]
] | TITLE: Bidirectional Cross-Modal Knowledge Exploration for Video Recognition
with Pre-trained Vision-Language Models
ABSTRACT: Vision-language models (VLMs) pre-trained on large-scale image-text pairs
have demonstrated impressive transferability on various visual tasks.
Transferring knowledge from such powerful VLMs is a promising direction for
building effective video recognition models. However, current exploration in
this field is still limited. We believe that the greatest value of pre-trained
VLMs lies in building a bridge between visual and textual domains. In this
paper, we propose a novel framework called BIKE, which utilizes the cross-modal
bridge to explore bidirectional knowledge: i) We introduce the Video Attribute
Association mechanism, which leverages the Video-to-Text knowledge to generate
textual auxiliary attributes for complementing video recognition. ii) We also
present a Temporal Concept Spotting mechanism that uses the Text-to-Video
expertise to capture temporal saliency in a parameter-free manner, leading to
enhanced video representation. Extensive studies on six popular video datasets,
including Kinetics-400 & 600, UCF-101, HMDB-51, ActivityNet and Charades, show
that our method achieves state-of-the-art performance in various recognition
scenarios, such as general, zero-shot, and few-shot video recognition. Our best
model achieves a state-of-the-art accuracy of 88.6% on the challenging
Kinetics-400 using the released CLIP model. The code is available at
https://github.com/whwu95/BIKE .
| no_new_dataset | 0.945701 |
2301.00746 | Santhosh Kumar Ramakrishnan | Santhosh Kumar Ramakrishnan, Ziad Al-Halah, Kristen Grauman | NaQ: Leveraging Narrations as Queries to Supervise Episodic Memory | 13 pages, 7 figures, appearing in CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Searching long egocentric videos with natural language queries (NLQ) has
compelling applications in augmented reality and robotics, where a fluid index
into everything that a person (agent) has seen before could augment human
memory and surface relevant information on demand. However, the structured
nature of the learning problem (free-form text query inputs, localized video
temporal window outputs) and its needle-in-a-haystack nature makes it both
technically challenging and expensive to supervise. We introduce
Narrations-as-Queries (NaQ), a data augmentation strategy that transforms
standard video-text narrations into training data for a video query
localization model. Validating our idea on the Ego4D benchmark, we find it has
tremendous impact in practice. NaQ improves multiple top models by substantial
margins (even doubling their accuracy), and yields the very best results to
date on the Ego4D NLQ challenge, soundly outperforming all challenge winners in
the CVPR and ECCV 2022 competitions and topping the current public leaderboard.
Beyond achieving the state-of-the-art for NLQ, we also demonstrate unique
properties of our approach such as the ability to perform zero-shot and
few-shot NLQ, and improved performance on queries about long-tail object
categories. Code and models:
{\small\url{http://vision.cs.utexas.edu/projects/naq}}.
| [
{
"version": "v1",
"created": "Mon, 2 Jan 2023 16:40:15 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 04:46:18 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ramakrishnan",
"Santhosh Kumar",
""
],
[
"Al-Halah",
"Ziad",
""
],
[
"Grauman",
"Kristen",
""
]
] | TITLE: NaQ: Leveraging Narrations as Queries to Supervise Episodic Memory
ABSTRACT: Searching long egocentric videos with natural language queries (NLQ) has
compelling applications in augmented reality and robotics, where a fluid index
into everything that a person (agent) has seen before could augment human
memory and surface relevant information on demand. However, the structured
nature of the learning problem (free-form text query inputs, localized video
temporal window outputs) and its needle-in-a-haystack nature makes it both
technically challenging and expensive to supervise. We introduce
Narrations-as-Queries (NaQ), a data augmentation strategy that transforms
standard video-text narrations into training data for a video query
localization model. Validating our idea on the Ego4D benchmark, we find it has
tremendous impact in practice. NaQ improves multiple top models by substantial
margins (even doubling their accuracy), and yields the very best results to
date on the Ego4D NLQ challenge, soundly outperforming all challenge winners in
the CVPR and ECCV 2022 competitions and topping the current public leaderboard.
Beyond achieving the state-of-the-art for NLQ, we also demonstrate unique
properties of our approach such as the ability to perform zero-shot and
few-shot NLQ, and improved performance on queries about long-tail object
categories. Code and models:
{\small\url{http://vision.cs.utexas.edu/projects/naq}}.
| no_new_dataset | 0.949669 |
2301.00896 | Xingxing Wei | Wei Xingxing and Wang Songping and Yan Huanqian | Efficient Robustness Assessment via Adversarial Spatial-Temporal Focus
on Videos | accepted by TPAMI2023 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Adversarial robustness assessment for video recognition models has raised
concerns owing to their wide applications on safety-critical tasks. Compared
with images, videos have much high dimension, which brings huge computational
costs when generating adversarial videos. This is especially serious for the
query-based black-box attacks where gradient estimation for the threat models
is usually utilized, and high dimensions will lead to a large number of
queries. To mitigate this issue, we propose to simultaneously eliminate the
temporal and spatial redundancy within the video to achieve an effective and
efficient gradient estimation on the reduced searching space, and thus query
number could decrease. To implement this idea, we design the novel Adversarial
spatial-temporal Focus (AstFocus) attack on videos, which performs attacks on
the simultaneously focused key frames and key regions from the inter-frames and
intra-frames in the video. AstFocus attack is based on the cooperative
Multi-Agent Reinforcement Learning (MARL) framework. One agent is responsible
for selecting key frames, and another agent is responsible for selecting key
regions. These two agents are jointly trained by the common rewards received
from the black-box threat models to perform a cooperative prediction. By
continuously querying, the reduced searching space composed of key frames and
key regions is becoming precise, and the whole query number becomes less than
that on the original video. Extensive experiments on four mainstream video
recognition models and three widely used action recognition datasets
demonstrate that the proposed AstFocus attack outperforms the SOTA methods,
which is prevenient in fooling rate, query number, time, and perturbation
magnitude at the same.
| [
{
"version": "v1",
"created": "Tue, 3 Jan 2023 00:28:57 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 01:57:56 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Xingxing",
"Wei",
""
],
[
"Songping",
"Wang",
""
],
[
"Huanqian",
"Yan",
""
]
] | TITLE: Efficient Robustness Assessment via Adversarial Spatial-Temporal Focus
on Videos
ABSTRACT: Adversarial robustness assessment for video recognition models has raised
concerns owing to their wide applications on safety-critical tasks. Compared
with images, videos have much high dimension, which brings huge computational
costs when generating adversarial videos. This is especially serious for the
query-based black-box attacks where gradient estimation for the threat models
is usually utilized, and high dimensions will lead to a large number of
queries. To mitigate this issue, we propose to simultaneously eliminate the
temporal and spatial redundancy within the video to achieve an effective and
efficient gradient estimation on the reduced searching space, and thus query
number could decrease. To implement this idea, we design the novel Adversarial
spatial-temporal Focus (AstFocus) attack on videos, which performs attacks on
the simultaneously focused key frames and key regions from the inter-frames and
intra-frames in the video. AstFocus attack is based on the cooperative
Multi-Agent Reinforcement Learning (MARL) framework. One agent is responsible
for selecting key frames, and another agent is responsible for selecting key
regions. These two agents are jointly trained by the common rewards received
from the black-box threat models to perform a cooperative prediction. By
continuously querying, the reduced searching space composed of key frames and
key regions is becoming precise, and the whole query number becomes less than
that on the original video. Extensive experiments on four mainstream video
recognition models and three widely used action recognition datasets
demonstrate that the proposed AstFocus attack outperforms the SOTA methods,
which is prevenient in fooling rate, query number, time, and perturbation
magnitude at the same.
| no_new_dataset | 0.951997 |
2301.01970 | Ying Wei | Shuailei Ma, Yuefeng Wang, Jiaqi Fan, Ying Wei, Thomas H. Li, Hongli
Liu and Fanbing Lv | CAT: LoCalization and IdentificAtion Cascade Detection Transformer for
Open-World Object Detection | CVPR 2023 camera-ready version | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Open-world object detection (OWOD), as a more general and challenging goal,
requires the model trained from data on known objects to detect both known and
unknown objects and incrementally learn to identify these unknown objects. The
existing works which employ standard detection framework and fixed
pseudo-labelling mechanism (PLM) have the following problems: (i) The inclusion
of detecting unknown objects substantially reduces the model's ability to
detect known ones. (ii) The PLM does not adequately utilize the priori
knowledge of inputs. (iii) The fixed selection manner of PLM cannot guarantee
that the model is trained in the right direction. We observe that humans
subconsciously prefer to focus on all foreground objects and then identify each
one in detail, rather than localize and identify a single object
simultaneously, for alleviating the confusion. This motivates us to propose a
novel solution called CAT: LoCalization and IdentificAtion Cascade Detection
Transformer which decouples the detection process via the shared decoder in the
cascade decoding way. In the meanwhile, we propose the self-adaptive
pseudo-labelling mechanism which combines the model-driven with input-driven
PLM and self-adaptively generates robust pseudo-labels for unknown objects,
significantly improving the ability of CAT to retrieve unknown objects.
Comprehensive experiments on two benchmark datasets, i.e., MS-COCO and PASCAL
VOC, show that our model outperforms the state-of-the-art in terms of all
metrics in the task of OWOD, incremental object detection (IOD) and open-set
detection.
| [
{
"version": "v1",
"created": "Thu, 5 Jan 2023 09:11:16 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 12:42:10 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Mar 2023 02:55:57 GMT"
},
{
"version": "v4",
"created": "Tue, 14 Mar 2023 11:40:34 GMT"
},
{
"version": "v5",
"created": "Wed, 15 Mar 2023 03:02:19 GMT"
},
{
"version": "v6",
"created": "Mon, 27 Mar 2023 11:37:27 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ma",
"Shuailei",
""
],
[
"Wang",
"Yuefeng",
""
],
[
"Fan",
"Jiaqi",
""
],
[
"Wei",
"Ying",
""
],
[
"Li",
"Thomas H.",
""
],
[
"Liu",
"Hongli",
""
],
[
"Lv",
"Fanbing",
""
]
] | TITLE: CAT: LoCalization and IdentificAtion Cascade Detection Transformer for
Open-World Object Detection
ABSTRACT: Open-world object detection (OWOD), as a more general and challenging goal,
requires the model trained from data on known objects to detect both known and
unknown objects and incrementally learn to identify these unknown objects. The
existing works which employ standard detection framework and fixed
pseudo-labelling mechanism (PLM) have the following problems: (i) The inclusion
of detecting unknown objects substantially reduces the model's ability to
detect known ones. (ii) The PLM does not adequately utilize the priori
knowledge of inputs. (iii) The fixed selection manner of PLM cannot guarantee
that the model is trained in the right direction. We observe that humans
subconsciously prefer to focus on all foreground objects and then identify each
one in detail, rather than localize and identify a single object
simultaneously, for alleviating the confusion. This motivates us to propose a
novel solution called CAT: LoCalization and IdentificAtion Cascade Detection
Transformer which decouples the detection process via the shared decoder in the
cascade decoding way. In the meanwhile, we propose the self-adaptive
pseudo-labelling mechanism which combines the model-driven with input-driven
PLM and self-adaptively generates robust pseudo-labels for unknown objects,
significantly improving the ability of CAT to retrieve unknown objects.
Comprehensive experiments on two benchmark datasets, i.e., MS-COCO and PASCAL
VOC, show that our model outperforms the state-of-the-art in terms of all
metrics in the task of OWOD, incremental object detection (IOD) and open-set
detection.
| no_new_dataset | 0.951142 |
2301.02074 | Piyush Bagad | Piyush Bagad and Makarand Tapaswi and Cees G. M. Snoek | Test of Time: Instilling Video-Language Models with a Sense of Time | Accepted for publication at CVPR 2023. Project page:
https://bpiyush.github.io/testoftime-website/index.html | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modelling and understanding time remains a challenge in contemporary video
understanding models. With language emerging as a key driver towards powerful
generalization, it is imperative for foundational video-language models to have
a sense of time. In this paper, we consider a specific aspect of temporal
understanding: consistency of time order as elicited by before/after relations.
We establish that seven existing video-language models struggle to understand
even such simple temporal relations. We then question whether it is feasible to
equip these foundational models with temporal awareness without re-training
them from scratch. Towards this, we propose a temporal adaptation recipe on top
of one such model, VideoCLIP, based on post-pretraining on a small amount of
video-text data. We conduct a zero-shot evaluation of the adapted models on six
datasets for three downstream tasks which require varying degrees of time
awareness. We observe encouraging performance gains especially when the task
needs higher time awareness. Our work serves as a first step towards probing
and instilling a sense of time in existing video-language models without the
need for data and compute-intense training from scratch.
| [
{
"version": "v1",
"created": "Thu, 5 Jan 2023 14:14:36 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 12:44:50 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Bagad",
"Piyush",
""
],
[
"Tapaswi",
"Makarand",
""
],
[
"Snoek",
"Cees G. M.",
""
]
] | TITLE: Test of Time: Instilling Video-Language Models with a Sense of Time
ABSTRACT: Modelling and understanding time remains a challenge in contemporary video
understanding models. With language emerging as a key driver towards powerful
generalization, it is imperative for foundational video-language models to have
a sense of time. In this paper, we consider a specific aspect of temporal
understanding: consistency of time order as elicited by before/after relations.
We establish that seven existing video-language models struggle to understand
even such simple temporal relations. We then question whether it is feasible to
equip these foundational models with temporal awareness without re-training
them from scratch. Towards this, we propose a temporal adaptation recipe on top
of one such model, VideoCLIP, based on post-pretraining on a small amount of
video-text data. We conduct a zero-shot evaluation of the adapted models on six
datasets for three downstream tasks which require varying degrees of time
awareness. We observe encouraging performance gains especially when the task
needs higher time awareness. Our work serves as a first step towards probing
and instilling a sense of time in existing video-language models without the
need for data and compute-intense training from scratch.
| no_new_dataset | 0.949995 |
2301.02700 | Rameen Abdal | Rameen Abdal, Hsin-Ying Lee, Peihao Zhu, Menglei Chai, Aliaksandr
Siarohin, Peter Wonka, Sergey Tulyakov | 3DAvatarGAN: Bridging Domains for Personalized Editable Avatars | Project Page: https://rameenabdal.github.io/3DAvatarGAN/ | null | null | null | cs.CV cs.GR | http://creativecommons.org/licenses/by/4.0/ | Modern 3D-GANs synthesize geometry and texture by training on large-scale
datasets with a consistent structure. Training such models on stylized,
artistic data, with often unknown, highly variable geometry, and camera
information has not yet been shown possible. Can we train a 3D GAN on such
artistic data, while maintaining multi-view consistency and texture quality? To
this end, we propose an adaptation framework, where the source domain is a
pre-trained 3D-GAN, while the target domain is a 2D-GAN trained on artistic
datasets. We then distill the knowledge from a 2D generator to the source 3D
generator. To do that, we first propose an optimization-based method to align
the distributions of camera parameters across domains. Second, we propose
regularizations necessary to learn high-quality texture, while avoiding
degenerate geometric solutions, such as flat shapes. Third, we show a
deformation-based technique for modeling exaggerated geometry of artistic
domains, enabling -- as a byproduct -- personalized geometric editing. Finally,
we propose a novel inversion method for 3D-GANs linking the latent spaces of
the source and the target domains. Our contributions -- for the first time --
allow for the generation, editing, and animation of personalized artistic 3D
avatars on artistic datasets.
| [
{
"version": "v1",
"created": "Fri, 6 Jan 2023 19:58:47 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 11:26:54 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Abdal",
"Rameen",
""
],
[
"Lee",
"Hsin-Ying",
""
],
[
"Zhu",
"Peihao",
""
],
[
"Chai",
"Menglei",
""
],
[
"Siarohin",
"Aliaksandr",
""
],
[
"Wonka",
"Peter",
""
],
[
"Tulyakov",
"Sergey",
""
]
] | TITLE: 3DAvatarGAN: Bridging Domains for Personalized Editable Avatars
ABSTRACT: Modern 3D-GANs synthesize geometry and texture by training on large-scale
datasets with a consistent structure. Training such models on stylized,
artistic data, with often unknown, highly variable geometry, and camera
information has not yet been shown possible. Can we train a 3D GAN on such
artistic data, while maintaining multi-view consistency and texture quality? To
this end, we propose an adaptation framework, where the source domain is a
pre-trained 3D-GAN, while the target domain is a 2D-GAN trained on artistic
datasets. We then distill the knowledge from a 2D generator to the source 3D
generator. To do that, we first propose an optimization-based method to align
the distributions of camera parameters across domains. Second, we propose
regularizations necessary to learn high-quality texture, while avoiding
degenerate geometric solutions, such as flat shapes. Third, we show a
deformation-based technique for modeling exaggerated geometry of artistic
domains, enabling -- as a byproduct -- personalized geometric editing. Finally,
we propose a novel inversion method for 3D-GANs linking the latent spaces of
the source and the target domains. Our contributions -- for the first time --
allow for the generation, editing, and animation of personalized artistic 3D
avatars on artistic datasets.
| no_new_dataset | 0.949012 |
2301.05175 | Diogo Luvizon | Diogo Luvizon, Marc Habermann, Vladislav Golyanik, Adam Kortylewski,
Christian Theobalt | Scene-Aware 3D Multi-Human Motion Capture from a Single Camera | Accepted to Eurographics 2023. See also github:
https://github.com/dluvizon/scene-aware-3d-multi-human project page:
https://vcai.mpi-inf.mpg.de/projects/scene-aware-3d-multi-human/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we consider the problem of estimating the 3D position of
multiple humans in a scene as well as their body shape and articulation from a
single RGB video recorded with a static camera. In contrast to expensive
marker-based or multi-view systems, our lightweight setup is ideal for private
users as it enables an affordable 3D motion capture that is easy to install and
does not require expert knowledge. To deal with this challenging setting, we
leverage recent advances in computer vision using large-scale pre-trained
models for a variety of modalities, including 2D body joints, joint angles,
normalized disparity maps, and human segmentation masks. Thus, we introduce the
first non-linear optimization-based approach that jointly solves for the
absolute 3D position of each human, their articulated pose, their individual
shapes as well as the scale of the scene. In particular, we estimate the scene
depth and person unique scale from normalized disparity predictions using the
2D body joints and joint angles. Given the per-frame scene depth, we
reconstruct a point-cloud of the static scene in 3D space. Finally, given the
per-frame 3D estimates of the humans and scene point-cloud, we perform a
space-time coherent optimization over the video to ensure temporal, spatial and
physical plausibility. We evaluate our method on established multi-person 3D
human pose benchmarks where we consistently outperform previous methods and we
qualitatively demonstrate that our method is robust to in-the-wild conditions
including challenging scenes with people of different sizes.
| [
{
"version": "v1",
"created": "Thu, 12 Jan 2023 18:01:28 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 12:21:23 GMT"
},
{
"version": "v3",
"created": "Mon, 27 Mar 2023 06:59:55 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Luvizon",
"Diogo",
""
],
[
"Habermann",
"Marc",
""
],
[
"Golyanik",
"Vladislav",
""
],
[
"Kortylewski",
"Adam",
""
],
[
"Theobalt",
"Christian",
""
]
] | TITLE: Scene-Aware 3D Multi-Human Motion Capture from a Single Camera
ABSTRACT: In this work, we consider the problem of estimating the 3D position of
multiple humans in a scene as well as their body shape and articulation from a
single RGB video recorded with a static camera. In contrast to expensive
marker-based or multi-view systems, our lightweight setup is ideal for private
users as it enables an affordable 3D motion capture that is easy to install and
does not require expert knowledge. To deal with this challenging setting, we
leverage recent advances in computer vision using large-scale pre-trained
models for a variety of modalities, including 2D body joints, joint angles,
normalized disparity maps, and human segmentation masks. Thus, we introduce the
first non-linear optimization-based approach that jointly solves for the
absolute 3D position of each human, their articulated pose, their individual
shapes as well as the scale of the scene. In particular, we estimate the scene
depth and person unique scale from normalized disparity predictions using the
2D body joints and joint angles. Given the per-frame scene depth, we
reconstruct a point-cloud of the static scene in 3D space. Finally, given the
per-frame 3D estimates of the humans and scene point-cloud, we perform a
space-time coherent optimization over the video to ensure temporal, spatial and
physical plausibility. We evaluate our method on established multi-person 3D
human pose benchmarks where we consistently outperform previous methods and we
qualitatively demonstrate that our method is robust to in-the-wild conditions
including challenging scenes with people of different sizes.
| no_new_dataset | 0.94699 |
2301.06866 | Aniket Agarwal | Aniket Agarwal, Alex Zhang, Karthik Narasimhan, Igor Gilitschenski,
Vishvak Murahari, Yash Kant | Building Scalable Video Understanding Benchmarks through Sports | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Existing benchmarks for evaluating long video understanding falls short on
two critical aspects, either lacking in scale or quality of annotations. These
limitations arise from the difficulty in collecting dense annotations for long
videos, which often require manually labeling each frame. In this work, we
introduce an automated Annotation and Video Stream Alignment Pipeline
(abbreviated ASAP). We demonstrate the generality of ASAP by aligning unlabeled
videos of four different sports with corresponding freely available dense web
annotations (i.e. commentary). We then leverage ASAP scalability to create
LCric, a large-scale long video understanding benchmark, with over 1000 hours
of densely annotated long Cricket videos (with an average sample length of ~50
mins) collected at virtually zero annotation cost. We benchmark and analyze
state-of-the-art video understanding models on LCric through a large set of
compositional multi-choice and regression queries. We establish a human
baseline that indicates significant room for new research to explore. Our human
studies indicate that ASAP can align videos and annotations with high fidelity,
precision, and speed. The dataset along with the code for ASAP and baselines
can be accessed here: https://asap-benchmark.github.io/.
| [
{
"version": "v1",
"created": "Tue, 17 Jan 2023 13:20:21 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Jan 2023 15:33:02 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 08:34:56 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Agarwal",
"Aniket",
""
],
[
"Zhang",
"Alex",
""
],
[
"Narasimhan",
"Karthik",
""
],
[
"Gilitschenski",
"Igor",
""
],
[
"Murahari",
"Vishvak",
""
],
[
"Kant",
"Yash",
""
]
] | TITLE: Building Scalable Video Understanding Benchmarks through Sports
ABSTRACT: Existing benchmarks for evaluating long video understanding falls short on
two critical aspects, either lacking in scale or quality of annotations. These
limitations arise from the difficulty in collecting dense annotations for long
videos, which often require manually labeling each frame. In this work, we
introduce an automated Annotation and Video Stream Alignment Pipeline
(abbreviated ASAP). We demonstrate the generality of ASAP by aligning unlabeled
videos of four different sports with corresponding freely available dense web
annotations (i.e. commentary). We then leverage ASAP scalability to create
LCric, a large-scale long video understanding benchmark, with over 1000 hours
of densely annotated long Cricket videos (with an average sample length of ~50
mins) collected at virtually zero annotation cost. We benchmark and analyze
state-of-the-art video understanding models on LCric through a large set of
compositional multi-choice and regression queries. We establish a human
baseline that indicates significant room for new research to explore. Our human
studies indicate that ASAP can align videos and annotations with high fidelity,
precision, and speed. The dataset along with the code for ASAP and baselines
can be accessed here: https://asap-benchmark.github.io/.
| new_dataset | 0.957675 |
2301.07302 | Ram Ramrakhya | Ram Ramrakhya, Dhruv Batra, Erik Wijmans, Abhishek Das | PIRLNav: Pretraining with Imitation and RL Finetuning for ObjectNav | 8 pages + supplement | null | null | null | cs.LG cs.AI cs.RO | http://creativecommons.org/licenses/by/4.0/ | We study ObjectGoal Navigation -- where a virtual robot situated in a new
environment is asked to navigate to an object. Prior work has shown that
imitation learning (IL) using behavior cloning (BC) on a dataset of human
demonstrations achieves promising results. However, this has limitations -- 1)
BC policies generalize poorly to new states, since the training mimics actions
not their consequences, and 2) collecting demonstrations is expensive. On the
other hand, reinforcement learning (RL) is trivially scalable, but requires
careful reward engineering to achieve desirable behavior. We present PIRLNav, a
two-stage learning scheme for BC pretraining on human demonstrations followed
by RL-finetuning. This leads to a policy that achieves a success rate of
$65.0\%$ on ObjectNav ($+5.0\%$ absolute over previous state-of-the-art). Using
this BC$\rightarrow$RL training recipe, we present a rigorous empirical
analysis of design choices. First, we investigate whether human demonstrations
can be replaced with `free' (automatically generated) sources of
demonstrations, e.g. shortest paths (SP) or task-agnostic frontier exploration
(FE) trajectories. We find that BC$\rightarrow$RL on human demonstrations
outperforms BC$\rightarrow$RL on SP and FE trajectories, even when controlled
for same BC-pretraining success on train, and even on a subset of val episodes
where BC-pretraining success favors the SP or FE policies. Next, we study how
RL-finetuning performance scales with the size of the BC pretraining dataset.
We find that as we increase the size of BC-pretraining dataset and get to high
BC accuracies, improvements from RL-finetuning are smaller, and that $90\%$ of
the performance of our best BC$\rightarrow$RL policy can be achieved with less
than half the number of BC demonstrations. Finally, we analyze failure modes of
our ObjectNav policies, and present guidelines for further improving them.
| [
{
"version": "v1",
"created": "Wed, 18 Jan 2023 04:40:50 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 14:49:25 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ramrakhya",
"Ram",
""
],
[
"Batra",
"Dhruv",
""
],
[
"Wijmans",
"Erik",
""
],
[
"Das",
"Abhishek",
""
]
] | TITLE: PIRLNav: Pretraining with Imitation and RL Finetuning for ObjectNav
ABSTRACT: We study ObjectGoal Navigation -- where a virtual robot situated in a new
environment is asked to navigate to an object. Prior work has shown that
imitation learning (IL) using behavior cloning (BC) on a dataset of human
demonstrations achieves promising results. However, this has limitations -- 1)
BC policies generalize poorly to new states, since the training mimics actions
not their consequences, and 2) collecting demonstrations is expensive. On the
other hand, reinforcement learning (RL) is trivially scalable, but requires
careful reward engineering to achieve desirable behavior. We present PIRLNav, a
two-stage learning scheme for BC pretraining on human demonstrations followed
by RL-finetuning. This leads to a policy that achieves a success rate of
$65.0\%$ on ObjectNav ($+5.0\%$ absolute over previous state-of-the-art). Using
this BC$\rightarrow$RL training recipe, we present a rigorous empirical
analysis of design choices. First, we investigate whether human demonstrations
can be replaced with `free' (automatically generated) sources of
demonstrations, e.g. shortest paths (SP) or task-agnostic frontier exploration
(FE) trajectories. We find that BC$\rightarrow$RL on human demonstrations
outperforms BC$\rightarrow$RL on SP and FE trajectories, even when controlled
for same BC-pretraining success on train, and even on a subset of val episodes
where BC-pretraining success favors the SP or FE policies. Next, we study how
RL-finetuning performance scales with the size of the BC pretraining dataset.
We find that as we increase the size of BC-pretraining dataset and get to high
BC accuracies, improvements from RL-finetuning are smaller, and that $90\%$ of
the performance of our best BC$\rightarrow$RL policy can be achieved with less
than half the number of BC demonstrations. Finally, we analyze failure modes of
our ObjectNav policies, and present guidelines for further improving them.
| no_new_dataset | 0.945096 |
2301.09629 | Qiuhong Anna Wei | Qiuhong Anna Wei, Sijie Ding, Jeong Joon Park, Rahul Sajnani, Adrien
Poulenard, Srinath Sridhar, Leonidas Guibas | LEGO-Net: Learning Regular Rearrangements of Objects in Rooms | Project page: https://ivl.cs.brown.edu/projects/lego-net | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Humans universally dislike the task of cleaning up a messy room. If machines
were to help us with this task, they must understand human criteria for regular
arrangements, such as several types of symmetry, co-linearity or
co-circularity, spacing uniformity in linear or circular patterns, and further
inter-object relationships that relate to style and functionality. Previous
approaches for this task relied on human input to explicitly specify goal
state, or synthesized scenes from scratch -- but such methods do not address
the rearrangement of existing messy scenes without providing a goal state. In
this paper, we present LEGO-Net, a data-driven transformer-based iterative
method for LEarning reGular rearrangement of Objects in messy rooms. LEGO-Net
is partly inspired by diffusion models -- it starts with an initial messy state
and iteratively ''de-noises'' the position and orientation of objects to a
regular state while reducing distance traveled. Given randomly perturbed object
positions and orientations in an existing dataset of professionally-arranged
scenes, our method is trained to recover a regular re-arrangement. Results
demonstrate that our method is able to reliably rearrange room scenes and
outperform other methods. We additionally propose a metric for evaluating
regularity in room arrangements using number-theoretic machinery.
| [
{
"version": "v1",
"created": "Mon, 23 Jan 2023 18:58:02 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 22:49:15 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wei",
"Qiuhong Anna",
""
],
[
"Ding",
"Sijie",
""
],
[
"Park",
"Jeong Joon",
""
],
[
"Sajnani",
"Rahul",
""
],
[
"Poulenard",
"Adrien",
""
],
[
"Sridhar",
"Srinath",
""
],
[
"Guibas",
"Leonidas",
""
]
] | TITLE: LEGO-Net: Learning Regular Rearrangements of Objects in Rooms
ABSTRACT: Humans universally dislike the task of cleaning up a messy room. If machines
were to help us with this task, they must understand human criteria for regular
arrangements, such as several types of symmetry, co-linearity or
co-circularity, spacing uniformity in linear or circular patterns, and further
inter-object relationships that relate to style and functionality. Previous
approaches for this task relied on human input to explicitly specify goal
state, or synthesized scenes from scratch -- but such methods do not address
the rearrangement of existing messy scenes without providing a goal state. In
this paper, we present LEGO-Net, a data-driven transformer-based iterative
method for LEarning reGular rearrangement of Objects in messy rooms. LEGO-Net
is partly inspired by diffusion models -- it starts with an initial messy state
and iteratively ''de-noises'' the position and orientation of objects to a
regular state while reducing distance traveled. Given randomly perturbed object
positions and orientations in an existing dataset of professionally-arranged
scenes, our method is trained to recover a regular re-arrangement. Results
demonstrate that our method is able to reliably rearrange room scenes and
outperform other methods. We additionally propose a metric for evaluating
regularity in room arrangements using number-theoretic machinery.
| no_new_dataset | 0.948058 |
2301.11517 | Xiangyu Zhao | Xiangyu Zhao, Hannes St\"ark, Dominique Beaini, Yiren Zhao, Pietro
Li\`o | Task-Agnostic Graph Neural Network Evaluation via Adversarial
Collaboration | 11th International Conference on Learning Representations (ICLR 2023)
Machine Learning for Drug Discovery (MLDD) Workshop. 17 pages, 6 figures, 4
tables | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | It has been increasingly demanding to develop reliable methods to evaluate
the progress of Graph Neural Network (GNN) research for molecular
representation learning. Existing GNN benchmarking methods for molecular
representation learning focus on comparing the GNNs' performances on some
node/graph classification/regression tasks on certain datasets. However, there
lacks a principled, task-agnostic method to directly compare two GNNs.
Additionally, most of the existing self-supervised learning works incorporate
handcrafted augmentations to the data, which has several severe difficulties to
be applied on graphs due to their unique characteristics. To address the
aforementioned issues, we propose GraphAC (Graph Adversarial Collaboration) --
a conceptually novel, principled, task-agnostic, and stable framework for
evaluating GNNs through contrastive self-supervision. We introduce a novel
objective function: the Competitive Barlow Twins, that allow two GNNs to
jointly update themselves from direct competitions against each other. GraphAC
succeeds in distinguishing GNNs of different expressiveness across various
aspects, and has demonstrated to be a principled and reliable GNN evaluation
method, without necessitating any augmentations.
| [
{
"version": "v1",
"created": "Fri, 27 Jan 2023 03:33:11 GMT"
},
{
"version": "v2",
"created": "Sat, 11 Feb 2023 00:30:47 GMT"
},
{
"version": "v3",
"created": "Mon, 27 Mar 2023 02:59:10 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhao",
"Xiangyu",
""
],
[
"Stärk",
"Hannes",
""
],
[
"Beaini",
"Dominique",
""
],
[
"Zhao",
"Yiren",
""
],
[
"Liò",
"Pietro",
""
]
] | TITLE: Task-Agnostic Graph Neural Network Evaluation via Adversarial
Collaboration
ABSTRACT: It has been increasingly demanding to develop reliable methods to evaluate
the progress of Graph Neural Network (GNN) research for molecular
representation learning. Existing GNN benchmarking methods for molecular
representation learning focus on comparing the GNNs' performances on some
node/graph classification/regression tasks on certain datasets. However, there
lacks a principled, task-agnostic method to directly compare two GNNs.
Additionally, most of the existing self-supervised learning works incorporate
handcrafted augmentations to the data, which has several severe difficulties to
be applied on graphs due to their unique characteristics. To address the
aforementioned issues, we propose GraphAC (Graph Adversarial Collaboration) --
a conceptually novel, principled, task-agnostic, and stable framework for
evaluating GNNs through contrastive self-supervision. We introduce a novel
objective function: the Competitive Barlow Twins, that allow two GNNs to
jointly update themselves from direct competitions against each other. GraphAC
succeeds in distinguishing GNNs of different expressiveness across various
aspects, and has demonstrated to be a principled and reliable GNN evaluation
method, without necessitating any augmentations.
| no_new_dataset | 0.949435 |
2302.00463 | Manon Flageat | Manon Flageat and Antoine Cully | Uncertain Quality-Diversity: Evaluation methodology and new methods for
Quality-Diversity in Uncertain Domains | Submitted to Transactions on Evolutionary Computation | null | null | null | cs.NE | http://creativecommons.org/licenses/by/4.0/ | Quality-Diversity optimisation (QD) has proven to yield promising results
across a broad set of applications. However, QD approaches struggle in the
presence of uncertainty in the environment, as it impacts their ability to
quantify the true performance and novelty of solutions. This problem has been
highlighted multiple times independently in previous literature. In this work,
we propose to uniformise the view on this problem through four main
contributions. First, we formalise a common framework for uncertain domains:
the Uncertain QD setting, a special case of QD in which fitness and descriptors
for each solution are no longer fixed values but distribution over possible
values. Second, we propose a new methodology to evaluate Uncertain QD
approaches, relying on a new per-generation sampling budget and a set of
existing and new metrics specifically designed for Uncertain QD. Third, we
propose three new Uncertain QD algorithms: Archive-sampling,
Parallel-Adaptive-sampling and Deep-Grid-sampling. We propose these approaches
taking into account recent advances in the QD community toward the use of
hardware acceleration that enable large numbers of parallel evaluations and
make sampling an affordable approach to uncertainty. Our final and fourth
contribution is to use this new framework and the associated comparison methods
to benchmark existing and novel approaches. We demonstrate once again the
limitation of MAP-Elites in uncertain domains and highlight the performance of
the existing Deep-Grid approach, and of our new algorithms. The goal of this
framework and methods is to become an instrumental benchmark for future works
considering Uncertain QD.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2023 14:12:04 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 17:44:40 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Flageat",
"Manon",
""
],
[
"Cully",
"Antoine",
""
]
] | TITLE: Uncertain Quality-Diversity: Evaluation methodology and new methods for
Quality-Diversity in Uncertain Domains
ABSTRACT: Quality-Diversity optimisation (QD) has proven to yield promising results
across a broad set of applications. However, QD approaches struggle in the
presence of uncertainty in the environment, as it impacts their ability to
quantify the true performance and novelty of solutions. This problem has been
highlighted multiple times independently in previous literature. In this work,
we propose to uniformise the view on this problem through four main
contributions. First, we formalise a common framework for uncertain domains:
the Uncertain QD setting, a special case of QD in which fitness and descriptors
for each solution are no longer fixed values but distribution over possible
values. Second, we propose a new methodology to evaluate Uncertain QD
approaches, relying on a new per-generation sampling budget and a set of
existing and new metrics specifically designed for Uncertain QD. Third, we
propose three new Uncertain QD algorithms: Archive-sampling,
Parallel-Adaptive-sampling and Deep-Grid-sampling. We propose these approaches
taking into account recent advances in the QD community toward the use of
hardware acceleration that enable large numbers of parallel evaluations and
make sampling an affordable approach to uncertainty. Our final and fourth
contribution is to use this new framework and the associated comparison methods
to benchmark existing and novel approaches. We demonstrate once again the
limitation of MAP-Elites in uncertain domains and highlight the performance of
the existing Deep-Grid approach, and of our new algorithms. The goal of this
framework and methods is to become an instrumental benchmark for future works
considering Uncertain QD.
| no_new_dataset | 0.941868 |
2302.02335 | Yu-Chu Yu | Yu-Chu Yu and Hsuan-Tien Lin | Semi-Supervised Domain Adaptation with Source Label Adaptation | Accepted by CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Semi-Supervised Domain Adaptation (SSDA) involves learning to classify unseen
target data with a few labeled and lots of unlabeled target data, along with
many labeled source data from a related domain. Current SSDA approaches usually
aim at aligning the target data to the labeled source data with feature space
mapping and pseudo-label assignments. Nevertheless, such a source-oriented
model can sometimes align the target data to source data of the wrong classes,
degrading the classification performance. This paper presents a novel
source-adaptive paradigm that adapts the source data to match the target data.
Our key idea is to view the source data as a noisily-labeled version of the
ideal target data. Then, we propose an SSDA model that cleans up the label
noise dynamically with the help of a robust cleaner component designed from the
target perspective. Since the paradigm is very different from the core ideas
behind existing SSDA approaches, our proposed model can be easily coupled with
them to improve their performance. Empirical results on two state-of-the-art
SSDA approaches demonstrate that the proposed model effectively cleans up the
noise within the source labels and exhibits superior performance over those
approaches across benchmark datasets. Our code is available at
https://github.com/chu0802/SLA .
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2023 08:34:04 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 20:54:44 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yu",
"Yu-Chu",
""
],
[
"Lin",
"Hsuan-Tien",
""
]
] | TITLE: Semi-Supervised Domain Adaptation with Source Label Adaptation
ABSTRACT: Semi-Supervised Domain Adaptation (SSDA) involves learning to classify unseen
target data with a few labeled and lots of unlabeled target data, along with
many labeled source data from a related domain. Current SSDA approaches usually
aim at aligning the target data to the labeled source data with feature space
mapping and pseudo-label assignments. Nevertheless, such a source-oriented
model can sometimes align the target data to source data of the wrong classes,
degrading the classification performance. This paper presents a novel
source-adaptive paradigm that adapts the source data to match the target data.
Our key idea is to view the source data as a noisily-labeled version of the
ideal target data. Then, we propose an SSDA model that cleans up the label
noise dynamically with the help of a robust cleaner component designed from the
target perspective. Since the paradigm is very different from the core ideas
behind existing SSDA approaches, our proposed model can be easily coupled with
them to improve their performance. Empirical results on two state-of-the-art
SSDA approaches demonstrate that the proposed model effectively cleans up the
noise within the source labels and exhibits superior performance over those
approaches across benchmark datasets. Our code is available at
https://github.com/chu0802/SLA .
| no_new_dataset | 0.947381 |
2302.04843 | Stevo Rackovic | Stevo Rackovi\'c, Cl\'audia Soares, Du\v{s}an Jakoveti\'c, Zoranka
Desnica | Accurate and Interpretable Solution of the Inverse Rig for Realistic
Blendshape Models with Quadratic Corrective Terms | null | null | null | null | cs.GR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We propose a new model-based algorithm solving the inverse rig problem in
facial animation retargeting, exhibiting higher accuracy of the fit and
sparser, more interpretable weight vector compared to SOTA. The proposed method
targets a specific subdomain of human face animation - highly-realistic
blendshape models used in the production of movies and video games. In this
paper, we formulate an optimization problem that takes into account all the
requirements of targeted models. Our objective goes beyond a linear blendshape
model and employs the quadratic corrective terms necessary for correctly
fitting fine details of the mesh. We show that the solution to the proposed
problem yields highly accurate mesh reconstruction even when general-purpose
solvers, like SQP, are used. The results obtained using SQP are highly accurate
in the mesh space but do not exhibit favorable qualities in terms of weight
sparsity and smoothness, and for this reason, we further propose a novel
algorithm relying on a MM technique. The algorithm is specifically suited for
solving the proposed objective, yielding a high-accuracy mesh fit while
respecting the constraints and producing a sparse and smooth set of weights
easy to manipulate and interpret by artists. Our algorithm is benchmarked with
SOTA approaches, and shows an overall superiority of the results, yielding a
smooth animation reconstruction with a relative improvement up to 45 percent in
root mean squared mesh error while keeping the cardinality comparable with
benchmark methods. This paper gives a comprehensive set of evaluation metrics
that cover different aspects of the solution, including mesh accuracy, sparsity
of the weights, and smoothness of the animation curves, as well as the
appearance of the produced animation, which human experts evaluated.
| [
{
"version": "v1",
"created": "Thu, 9 Feb 2023 18:51:04 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 09:11:04 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Racković",
"Stevo",
""
],
[
"Soares",
"Cláudia",
""
],
[
"Jakovetić",
"Dušan",
""
],
[
"Desnica",
"Zoranka",
""
]
] | TITLE: Accurate and Interpretable Solution of the Inverse Rig for Realistic
Blendshape Models with Quadratic Corrective Terms
ABSTRACT: We propose a new model-based algorithm solving the inverse rig problem in
facial animation retargeting, exhibiting higher accuracy of the fit and
sparser, more interpretable weight vector compared to SOTA. The proposed method
targets a specific subdomain of human face animation - highly-realistic
blendshape models used in the production of movies and video games. In this
paper, we formulate an optimization problem that takes into account all the
requirements of targeted models. Our objective goes beyond a linear blendshape
model and employs the quadratic corrective terms necessary for correctly
fitting fine details of the mesh. We show that the solution to the proposed
problem yields highly accurate mesh reconstruction even when general-purpose
solvers, like SQP, are used. The results obtained using SQP are highly accurate
in the mesh space but do not exhibit favorable qualities in terms of weight
sparsity and smoothness, and for this reason, we further propose a novel
algorithm relying on a MM technique. The algorithm is specifically suited for
solving the proposed objective, yielding a high-accuracy mesh fit while
respecting the constraints and producing a sparse and smooth set of weights
easy to manipulate and interpret by artists. Our algorithm is benchmarked with
SOTA approaches, and shows an overall superiority of the results, yielding a
smooth animation reconstruction with a relative improvement up to 45 percent in
root mean squared mesh error while keeping the cardinality comparable with
benchmark methods. This paper gives a comprehensive set of evaluation metrics
that cover different aspects of the solution, including mesh accuracy, sparsity
of the weights, and smoothness of the animation curves, as well as the
appearance of the produced animation, which human experts evaluated.
| no_new_dataset | 0.950686 |
2302.07061 | Gengmo Zhou | Gengmo Zhou, Zhifeng Gao, Zhewei Wei, Hang Zheng, Guolin Ke | Do Deep Learning Methods Really Perform Better in Molecular Conformation
Generation? | null | null | null | null | cs.CE cs.LG q-bio.BM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Molecular conformation generation (MCG) is a fundamental and important
problem in drug discovery. Many traditional methods have been developed to
solve the MCG problem, such as systematic searching, model-building, random
searching, distance geometry, molecular dynamics, Monte Carlo methods, etc.
However, they have some limitations depending on the molecular structures.
Recently, there are plenty of deep learning based MCG methods, which claim they
largely outperform the traditional methods. However, to our surprise, we design
a simple and cheap algorithm (parameter-free) based on the traditional methods
and find it is comparable to or even outperforms deep learning based MCG
methods in the widely used GEOM-QM9 and GEOM-Drugs benchmarks. In particular,
our design algorithm is simply the clustering of the RDKIT-generated
conformations. We hope our findings can help the community to revise the deep
learning methods for MCG. The code of the proposed algorithm could be found at
https://gist.github.com/ZhouGengmo/5b565f51adafcd911c0bc115b2ef027c.
| [
{
"version": "v1",
"created": "Tue, 14 Feb 2023 14:21:26 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 14:15:37 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhou",
"Gengmo",
""
],
[
"Gao",
"Zhifeng",
""
],
[
"Wei",
"Zhewei",
""
],
[
"Zheng",
"Hang",
""
],
[
"Ke",
"Guolin",
""
]
] | TITLE: Do Deep Learning Methods Really Perform Better in Molecular Conformation
Generation?
ABSTRACT: Molecular conformation generation (MCG) is a fundamental and important
problem in drug discovery. Many traditional methods have been developed to
solve the MCG problem, such as systematic searching, model-building, random
searching, distance geometry, molecular dynamics, Monte Carlo methods, etc.
However, they have some limitations depending on the molecular structures.
Recently, there are plenty of deep learning based MCG methods, which claim they
largely outperform the traditional methods. However, to our surprise, we design
a simple and cheap algorithm (parameter-free) based on the traditional methods
and find it is comparable to or even outperforms deep learning based MCG
methods in the widely used GEOM-QM9 and GEOM-Drugs benchmarks. In particular,
our design algorithm is simply the clustering of the RDKIT-generated
conformations. We hope our findings can help the community to revise the deep
learning methods for MCG. The code of the proposed algorithm could be found at
https://gist.github.com/ZhouGengmo/5b565f51adafcd911c0bc115b2ef027c.
| no_new_dataset | 0.943191 |
2302.09685 | Ankan Mullick | Ankan Mullick, Ishani Mondal, Sourjyadip Ray, R Raghav, G Sai
Chaitanya, Pawan Goyal | Intent Identification and Entity Extraction for Healthcare Queries in
Indic Languages | null | EACL 2023 Findings Full Paper | null | null | cs.IR cs.CL | http://creativecommons.org/licenses/by/4.0/ | Scarcity of data and technological limitations for resource-poor languages in
developing countries like India poses a threat to the development of
sophisticated NLU systems for healthcare. To assess the current status of
various state-of-the-art language models in healthcare, this paper studies the
problem by initially proposing two different Healthcare datasets, Indian
Healthcare Query Intent-WebMD and 1mg (IHQID-WebMD and IHQID-1mg) and one real
world Indian hospital query data in English and multiple Indic languages
(Hindi, Bengali, Tamil, Telugu, Marathi and Gujarati) which are annotated with
the query intents as well as entities. Our aim is to detect query intents and
extract corresponding entities. We perform extensive experiments on a set of
models in various realistic settings and explore two scenarios based on the
access to English data only (less costly) and access to target language data
(more expensive). We analyze context specific practical relevancy through
empirical analysis. The results, expressed in terms of overall F1 score show
that our approach is practically useful to identify intents and entities.
| [
{
"version": "v1",
"created": "Sun, 19 Feb 2023 22:53:03 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Mullick",
"Ankan",
""
],
[
"Mondal",
"Ishani",
""
],
[
"Ray",
"Sourjyadip",
""
],
[
"Raghav",
"R",
""
],
[
"Chaitanya",
"G Sai",
""
],
[
"Goyal",
"Pawan",
""
]
] | TITLE: Intent Identification and Entity Extraction for Healthcare Queries in
Indic Languages
ABSTRACT: Scarcity of data and technological limitations for resource-poor languages in
developing countries like India poses a threat to the development of
sophisticated NLU systems for healthcare. To assess the current status of
various state-of-the-art language models in healthcare, this paper studies the
problem by initially proposing two different Healthcare datasets, Indian
Healthcare Query Intent-WebMD and 1mg (IHQID-WebMD and IHQID-1mg) and one real
world Indian hospital query data in English and multiple Indic languages
(Hindi, Bengali, Tamil, Telugu, Marathi and Gujarati) which are annotated with
the query intents as well as entities. Our aim is to detect query intents and
extract corresponding entities. We perform extensive experiments on a set of
models in various realistic settings and explore two scenarios based on the
access to English data only (less costly) and access to target language data
(more expensive). We analyze context specific practical relevancy through
empirical analysis. The results, expressed in terms of overall F1 score show
that our approach is practically useful to identify intents and entities.
| no_new_dataset | 0.901575 |
2302.14338 | Wenwen Yu | Wenwen Yu, Yuliang Liu, Wei Hua, Deqiang Jiang, Bo Ren, Xiang Bai | Turning a CLIP Model into a Scene Text Detector | CVPR2023 | null | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | The recent large-scale Contrastive Language-Image Pretraining (CLIP) model
has shown great potential in various downstream tasks via leveraging the
pretrained vision and language knowledge. Scene text, which contains rich
textual and visual information, has an inherent connection with a model like
CLIP. Recently, pretraining approaches based on vision language models have
made effective progresses in the field of text detection. In contrast to these
works, this paper proposes a new method, termed TCM, focusing on Turning the
CLIP Model directly for text detection without pretraining process. We
demonstrate the advantages of the proposed TCM as follows: (1) The underlying
principle of our framework can be applied to improve existing scene text
detector. (2) It facilitates the few-shot training capability of existing
methods, e.g., by using 10% of labeled data, we significantly improve the
performance of the baseline method with an average of 22% in terms of the
F-measure on 4 benchmarks. (3) By turning the CLIP model into existing scene
text detection methods, we further achieve promising domain adaptation ability.
The code will be publicly released at https://github.com/wenwenyu/TCM.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 06:06:12 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 04:36:26 GMT"
},
{
"version": "v3",
"created": "Sun, 26 Mar 2023 12:58:19 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yu",
"Wenwen",
""
],
[
"Liu",
"Yuliang",
""
],
[
"Hua",
"Wei",
""
],
[
"Jiang",
"Deqiang",
""
],
[
"Ren",
"Bo",
""
],
[
"Bai",
"Xiang",
""
]
] | TITLE: Turning a CLIP Model into a Scene Text Detector
ABSTRACT: The recent large-scale Contrastive Language-Image Pretraining (CLIP) model
has shown great potential in various downstream tasks via leveraging the
pretrained vision and language knowledge. Scene text, which contains rich
textual and visual information, has an inherent connection with a model like
CLIP. Recently, pretraining approaches based on vision language models have
made effective progresses in the field of text detection. In contrast to these
works, this paper proposes a new method, termed TCM, focusing on Turning the
CLIP Model directly for text detection without pretraining process. We
demonstrate the advantages of the proposed TCM as follows: (1) The underlying
principle of our framework can be applied to improve existing scene text
detector. (2) It facilitates the few-shot training capability of existing
methods, e.g., by using 10% of labeled data, we significantly improve the
performance of the baseline method with an average of 22% in terms of the
F-measure on 4 benchmarks. (3) By turning the CLIP model into existing scene
text detection methods, we further achieve promising domain adaptation ability.
The code will be publicly released at https://github.com/wenwenyu/TCM.
| no_new_dataset | 0.949012 |
2302.14411 | Bart Bogaerts | Bart Bogaerts, Bas Ketsman, Younes Zeboudj, Heba Aamer, Ruben Taelman,
Ruben Verborgh | Distributed Subweb Specifications for Traversing the Web | Under consideration in Theory and Practice of Logic Programming
(TPLP) | null | null | null | cs.DB cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Link Traversal-based Query Processing (ltqp), in which a sparql query is
evaluated over a web of documents rather than a single dataset, is often seen
as a theoretically interesting yet impractical technique. However, in a time
where the hypercentralization of data has increasingly come under scrutiny, a
decentralized Web of Data with a simple document-based interface is appealing,
as it enables data publishers to control their data and access rights. While
ltqp allows evaluating complex queries over such webs, it suffers from
performance issues (due to the high number of documents containing data) as
well as information quality concerns (due to the many sources providing such
documents). In existing ltqp approaches, the burden of finding sources to query
is entirely in the hands of the data consumer. In this paper, we argue that to
solve these issues, data publishers should also be able to suggest sources of
interest and guide the data consumer towards relevant and trustworthy data. We
introduce a theoretical framework that enables such guided link traversal and
study its properties. We illustrate with a theoretic example that this can
improve query results and reduce the number of network requests. We evaluate
our proposal experimentally on a virtual linked web with specifications and
indeed observe that not just the data quality but also the efficiency of
querying improves.
Under consideration in Theory and Practice of Logic Programming (TPLP).
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 08:46:04 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 07:58:23 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Bogaerts",
"Bart",
""
],
[
"Ketsman",
"Bas",
""
],
[
"Zeboudj",
"Younes",
""
],
[
"Aamer",
"Heba",
""
],
[
"Taelman",
"Ruben",
""
],
[
"Verborgh",
"Ruben",
""
]
] | TITLE: Distributed Subweb Specifications for Traversing the Web
ABSTRACT: Link Traversal-based Query Processing (ltqp), in which a sparql query is
evaluated over a web of documents rather than a single dataset, is often seen
as a theoretically interesting yet impractical technique. However, in a time
where the hypercentralization of data has increasingly come under scrutiny, a
decentralized Web of Data with a simple document-based interface is appealing,
as it enables data publishers to control their data and access rights. While
ltqp allows evaluating complex queries over such webs, it suffers from
performance issues (due to the high number of documents containing data) as
well as information quality concerns (due to the many sources providing such
documents). In existing ltqp approaches, the burden of finding sources to query
is entirely in the hands of the data consumer. In this paper, we argue that to
solve these issues, data publishers should also be able to suggest sources of
interest and guide the data consumer towards relevant and trustworthy data. We
introduce a theoretical framework that enables such guided link traversal and
study its properties. We illustrate with a theoretic example that this can
improve query results and reduce the number of network requests. We evaluate
our proposal experimentally on a virtual linked web with specifications and
indeed observe that not just the data quality but also the efficiency of
querying improves.
Under consideration in Theory and Practice of Logic Programming (TPLP).
| no_new_dataset | 0.946101 |
2302.14673 | Chengzhi Wu | Chengzhi Wu, Junwei Zheng, Julius Pfrommer, J\"urgen Beyerer | Attention-based Point Cloud Edge Sampling | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Point cloud sampling is a less explored research topic for this data
representation. The most commonly used sampling methods are still classical
random sampling and farthest point sampling. With the development of neural
networks, various methods have been proposed to sample point clouds in a
task-based learning manner. However, these methods are mostly generative-based,
rather than selecting points directly using mathematical statistics. Inspired
by the Canny edge detection algorithm for images and with the help of the
attention mechanism, this paper proposes a non-generative Attention-based Point
cloud Edge Sampling method (APES), which captures salient points in the point
cloud outline. Both qualitative and quantitative experimental results show the
superior performance of our sampling method on common benchmark tasks.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 15:36:17 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 14:32:23 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wu",
"Chengzhi",
""
],
[
"Zheng",
"Junwei",
""
],
[
"Pfrommer",
"Julius",
""
],
[
"Beyerer",
"Jürgen",
""
]
] | TITLE: Attention-based Point Cloud Edge Sampling
ABSTRACT: Point cloud sampling is a less explored research topic for this data
representation. The most commonly used sampling methods are still classical
random sampling and farthest point sampling. With the development of neural
networks, various methods have been proposed to sample point clouds in a
task-based learning manner. However, these methods are mostly generative-based,
rather than selecting points directly using mathematical statistics. Inspired
by the Canny edge detection algorithm for images and with the help of the
attention mechanism, this paper proposes a non-generative Attention-based Point
cloud Edge Sampling method (APES), which captures salient points in the point
cloud outline. Both qualitative and quantitative experimental results show the
superior performance of our sampling method on common benchmark tasks.
| no_new_dataset | 0.952264 |
2303.00917 | Chenqi Kong | Chenqi Kong, Haoliang Li, Shiqi Wang | Enhancing General Face Forgery Detection via Vision Transformer with
Low-Rank Adaptation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nowadays, forgery faces pose pressing security concerns over fake news,
fraud, impersonation, etc. Despite the demonstrated success in intra-domain
face forgery detection, existing detection methods lack generalization
capability and tend to suffer from dramatic performance drops when deployed to
unforeseen domains. To mitigate this issue, this paper designs a more general
fake face detection model based on the vision transformer(ViT) architecture. In
the training phase, the pretrained ViT weights are freezed, and only the
Low-Rank Adaptation(LoRA) modules are updated. Additionally, the Single Center
Loss(SCL) is applied to supervise the training process, further improving the
generalization capability of the model. The proposed method achieves
state-of-the-arts detection performances in both cross-manipulation and
cross-dataset evaluations.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 02:26:04 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 07:42:24 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Kong",
"Chenqi",
""
],
[
"Li",
"Haoliang",
""
],
[
"Wang",
"Shiqi",
""
]
] | TITLE: Enhancing General Face Forgery Detection via Vision Transformer with
Low-Rank Adaptation
ABSTRACT: Nowadays, forgery faces pose pressing security concerns over fake news,
fraud, impersonation, etc. Despite the demonstrated success in intra-domain
face forgery detection, existing detection methods lack generalization
capability and tend to suffer from dramatic performance drops when deployed to
unforeseen domains. To mitigate this issue, this paper designs a more general
fake face detection model based on the vision transformer(ViT) architecture. In
the training phase, the pretrained ViT weights are freezed, and only the
Low-Rank Adaptation(LoRA) modules are updated. Additionally, the Single Center
Loss(SCL) is applied to supervise the training process, further improving the
generalization capability of the model. The proposed method achieves
state-of-the-arts detection performances in both cross-manipulation and
cross-dataset evaluations.
| no_new_dataset | 0.950365 |
2303.00938 | Jialiang Zhang | Yinzhen Xu, Weikang Wan, Jialiang Zhang, Haoran Liu, Zikang Shan, Hao
Shen, Ruicheng Wang, Haoran Geng, Yijia Weng, Jiayi Chen, Tengyu Liu, Li Yi,
He Wang | UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse
Proposal Generation and Goal-Conditioned Policy | Accepted to CVPR 2023 | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we tackle the problem of learning universal robotic dexterous
grasping from a point cloud observation under a table-top setting. The goal is
to grasp and lift up objects in high-quality and diverse ways and generalize
across hundreds of categories and even the unseen. Inspired by successful
pipelines used in parallel gripper grasping, we split the task into two stages:
1) grasp proposal (pose) generation and 2) goal-conditioned grasp execution.
For the first stage, we propose a novel probabilistic model of grasp pose
conditioned on the point cloud observation that factorizes rotation from
translation and articulation. Trained on our synthesized large-scale dexterous
grasp dataset, this model enables us to sample diverse and high-quality
dexterous grasp poses for the object point cloud.For the second stage, we
propose to replace the motion planning used in parallel gripper grasping with a
goal-conditioned grasp policy, due to the complexity involved in dexterous
grasping execution. Note that it is very challenging to learn this highly
generalizable grasp policy that only takes realistic inputs without oracle
states. We thus propose several important innovations, including state
canonicalization, object curriculum, and teacher-student distillation.
Integrating the two stages, our final pipeline becomes the first to achieve
universal generalization for dexterous grasping, demonstrating an average
success rate of more than 60\% on thousands of object instances, which
significantly outperforms all baselines, meanwhile showing only a minimal
generalization gap.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 03:23:18 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 07:35:32 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Xu",
"Yinzhen",
""
],
[
"Wan",
"Weikang",
""
],
[
"Zhang",
"Jialiang",
""
],
[
"Liu",
"Haoran",
""
],
[
"Shan",
"Zikang",
""
],
[
"Shen",
"Hao",
""
],
[
"Wang",
"Ruicheng",
""
],
[
"Geng",
"Haoran",
""
],
[
"Weng",
"Yijia",
""
],
[
"Chen",
"Jiayi",
""
],
[
"Liu",
"Tengyu",
""
],
[
"Yi",
"Li",
""
],
[
"Wang",
"He",
""
]
] | TITLE: UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse
Proposal Generation and Goal-Conditioned Policy
ABSTRACT: In this work, we tackle the problem of learning universal robotic dexterous
grasping from a point cloud observation under a table-top setting. The goal is
to grasp and lift up objects in high-quality and diverse ways and generalize
across hundreds of categories and even the unseen. Inspired by successful
pipelines used in parallel gripper grasping, we split the task into two stages:
1) grasp proposal (pose) generation and 2) goal-conditioned grasp execution.
For the first stage, we propose a novel probabilistic model of grasp pose
conditioned on the point cloud observation that factorizes rotation from
translation and articulation. Trained on our synthesized large-scale dexterous
grasp dataset, this model enables us to sample diverse and high-quality
dexterous grasp poses for the object point cloud.For the second stage, we
propose to replace the motion planning used in parallel gripper grasping with a
goal-conditioned grasp policy, due to the complexity involved in dexterous
grasping execution. Note that it is very challenging to learn this highly
generalizable grasp policy that only takes realistic inputs without oracle
states. We thus propose several important innovations, including state
canonicalization, object curriculum, and teacher-student distillation.
Integrating the two stages, our final pipeline becomes the first to achieve
universal generalization for dexterous grasping, demonstrating an average
success rate of more than 60\% on thousands of object instances, which
significantly outperforms all baselines, meanwhile showing only a minimal
generalization gap.
| new_dataset | 0.841044 |
2303.01276 | Zicheng Wang | Zicheng Wang, Zhen Zhao, Xiaoxia Xing, Dong Xu, Xiangyu Kong, Luping
Zhou | Conflict-Based Cross-View Consistency for Semi-Supervised Semantic
Segmentation | accepted by CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-supervised semantic segmentation (SSS) has recently gained increasing
research interest as it can reduce the requirement for large-scale
fully-annotated training data. The current methods often suffer from the
confirmation bias from the pseudo-labelling process, which can be alleviated by
the co-training framework. The current co-training-based SSS methods rely on
hand-crafted perturbations to prevent the different sub-nets from collapsing
into each other, but these artificial perturbations cannot lead to the optimal
solution. In this work, we propose a new conflict-based cross-view consistency
(CCVC) method based on a two-branch co-training framework which aims at
enforcing the two sub-nets to learn informative features from irrelevant views.
In particular, we first propose a new cross-view consistency (CVC) strategy
that encourages the two sub-nets to learn distinct features from the same input
by introducing a feature discrepancy loss, while these distinct features are
expected to generate consistent prediction scores of the input. The CVC
strategy helps to prevent the two sub-nets from stepping into the collapse. In
addition, we further propose a conflict-based pseudo-labelling (CPL) method to
guarantee the model will learn more useful information from conflicting
predictions, which will lead to a stable training process. We validate our new
CCVC approach on the SSS benchmark datasets where our method achieves new
state-of-the-art performance. Our code is available at
https://github.com/xiaoyao3302/CCVC.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 14:02:16 GMT"
},
{
"version": "v2",
"created": "Fri, 3 Mar 2023 10:47:41 GMT"
},
{
"version": "v3",
"created": "Sat, 25 Mar 2023 06:33:15 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Zicheng",
""
],
[
"Zhao",
"Zhen",
""
],
[
"Xing",
"Xiaoxia",
""
],
[
"Xu",
"Dong",
""
],
[
"Kong",
"Xiangyu",
""
],
[
"Zhou",
"Luping",
""
]
] | TITLE: Conflict-Based Cross-View Consistency for Semi-Supervised Semantic
Segmentation
ABSTRACT: Semi-supervised semantic segmentation (SSS) has recently gained increasing
research interest as it can reduce the requirement for large-scale
fully-annotated training data. The current methods often suffer from the
confirmation bias from the pseudo-labelling process, which can be alleviated by
the co-training framework. The current co-training-based SSS methods rely on
hand-crafted perturbations to prevent the different sub-nets from collapsing
into each other, but these artificial perturbations cannot lead to the optimal
solution. In this work, we propose a new conflict-based cross-view consistency
(CCVC) method based on a two-branch co-training framework which aims at
enforcing the two sub-nets to learn informative features from irrelevant views.
In particular, we first propose a new cross-view consistency (CVC) strategy
that encourages the two sub-nets to learn distinct features from the same input
by introducing a feature discrepancy loss, while these distinct features are
expected to generate consistent prediction scores of the input. The CVC
strategy helps to prevent the two sub-nets from stepping into the collapse. In
addition, we further propose a conflict-based pseudo-labelling (CPL) method to
guarantee the model will learn more useful information from conflicting
predictions, which will lead to a stable training process. We validate our new
CCVC approach on the SSS benchmark datasets where our method achieves new
state-of-the-art performance. Our code is available at
https://github.com/xiaoyao3302/CCVC.
| no_new_dataset | 0.948917 |
2303.02375 | Bowen Cai | Bowen Cai, Jinchi Huang, Rongfei Jia, Chengfei Lv, Huan Fu | NeuDA: Neural Deformable Anchor for High-Fidelity Implicit Surface
Reconstruction | Accepted to CVPR 2023, project page:
https://3d-front-future.github.io/neuda | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies implicit surface reconstruction leveraging differentiable
ray casting. Previous works such as IDR and NeuS overlook the spatial context
in 3D space when predicting and rendering the surface, thereby may fail to
capture sharp local topologies such as small holes and structures. To mitigate
the limitation, we propose a flexible neural implicit representation leveraging
hierarchical voxel grids, namely Neural Deformable Anchor (NeuDA), for
high-fidelity surface reconstruction. NeuDA maintains the hierarchical anchor
grids where each vertex stores a 3D position (or anchor) instead of the direct
embedding (or feature). We optimize the anchor grids such that different local
geometry structures can be adaptively encoded. Besides, we dig into the
frequency encoding strategies and introduce a simple hierarchical positional
encoding method for the hierarchical anchor structure to flexibly exploit the
properties of high-frequency and low-frequency geometry and appearance.
Experiments on both the DTU and BlendedMVS datasets demonstrate that NeuDA can
produce promising mesh surfaces.
| [
{
"version": "v1",
"created": "Sat, 4 Mar 2023 10:38:41 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 10:38:58 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Cai",
"Bowen",
""
],
[
"Huang",
"Jinchi",
""
],
[
"Jia",
"Rongfei",
""
],
[
"Lv",
"Chengfei",
""
],
[
"Fu",
"Huan",
""
]
] | TITLE: NeuDA: Neural Deformable Anchor for High-Fidelity Implicit Surface
Reconstruction
ABSTRACT: This paper studies implicit surface reconstruction leveraging differentiable
ray casting. Previous works such as IDR and NeuS overlook the spatial context
in 3D space when predicting and rendering the surface, thereby may fail to
capture sharp local topologies such as small holes and structures. To mitigate
the limitation, we propose a flexible neural implicit representation leveraging
hierarchical voxel grids, namely Neural Deformable Anchor (NeuDA), for
high-fidelity surface reconstruction. NeuDA maintains the hierarchical anchor
grids where each vertex stores a 3D position (or anchor) instead of the direct
embedding (or feature). We optimize the anchor grids such that different local
geometry structures can be adaptively encoded. Besides, we dig into the
frequency encoding strategies and introduce a simple hierarchical positional
encoding method for the hierarchical anchor structure to flexibly exploit the
properties of high-frequency and low-frequency geometry and appearance.
Experiments on both the DTU and BlendedMVS datasets demonstrate that NeuDA can
produce promising mesh surfaces.
| no_new_dataset | 0.953188 |
2303.02841 | Bixing Yan | Bixing Yan, Shaoling Chen, Yuxuan He, Zhihan Li | Model-Agnostic Meta-Learning for Natural Language Understanding Tasks in
Finance | 13 pages, 6 figures, 8 tables | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Natural language understanding(NLU) is challenging for finance due to the
lack of annotated data and the specialized language in that domain. As a
result, researchers have proposed to use pre-trained language model and
multi-task learning to learn robust representations. However, aggressive
fine-tuning often causes over-fitting and multi-task learning may favor tasks
with significantly larger amounts data, etc. To address these problems, in this
paper, we investigate model-agnostic meta-learning algorithm(MAML) in
low-resource financial NLU tasks. Our contribution includes: 1. we explore the
performance of MAML method with multiple types of tasks: GLUE datasets, SNLI,
Sci-Tail and Financial PhraseBank; 2. we study the performance of MAML method
with multiple single-type tasks: a real scenario stock price prediction problem
with twitter text data. Our models achieve the state-of-the-art performance
according to the experimental results, which demonstrate that our method can
adapt fast and well to low-resource situations.
| [
{
"version": "v1",
"created": "Mon, 6 Mar 2023 02:24:48 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 19:45:25 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yan",
"Bixing",
""
],
[
"Chen",
"Shaoling",
""
],
[
"He",
"Yuxuan",
""
],
[
"Li",
"Zhihan",
""
]
] | TITLE: Model-Agnostic Meta-Learning for Natural Language Understanding Tasks in
Finance
ABSTRACT: Natural language understanding(NLU) is challenging for finance due to the
lack of annotated data and the specialized language in that domain. As a
result, researchers have proposed to use pre-trained language model and
multi-task learning to learn robust representations. However, aggressive
fine-tuning often causes over-fitting and multi-task learning may favor tasks
with significantly larger amounts data, etc. To address these problems, in this
paper, we investigate model-agnostic meta-learning algorithm(MAML) in
low-resource financial NLU tasks. Our contribution includes: 1. we explore the
performance of MAML method with multiple types of tasks: GLUE datasets, SNLI,
Sci-Tail and Financial PhraseBank; 2. we study the performance of MAML method
with multiple single-type tasks: a real scenario stock price prediction problem
with twitter text data. Our models achieve the state-of-the-art performance
according to the experimental results, which demonstrate that our method can
adapt fast and well to low-resource situations.
| no_new_dataset | 0.945751 |
2303.06237 | Xiaopeng Jiang | Xiaopeng Jiang, Cristian Borcea | Complement Sparsification: Low-Overhead Model Pruning for Federated
Learning | null | null | null | null | cs.LG cs.AI cs.DC | http://creativecommons.org/licenses/by/4.0/ | Federated Learning (FL) is a privacy-preserving distributed deep learning
paradigm that involves substantial communication and computation effort, which
is a problem for resource-constrained mobile and IoT devices. Model
pruning/sparsification develops sparse models that could solve this problem,
but existing sparsification solutions cannot satisfy at the same time the
requirements for low bidirectional communication overhead between the server
and the clients, low computation overhead at the clients, and good model
accuracy, under the FL assumption that the server does not have access to raw
data to fine-tune the pruned models. We propose Complement Sparsification (CS),
a pruning mechanism that satisfies all these requirements through a
complementary and collaborative pruning done at the server and the clients. At
each round, CS creates a global sparse model that contains the weights that
capture the general data distribution of all clients, while the clients create
local sparse models with the weights pruned from the global model to capture
the local trends. For improved model performance, these two types of
complementary sparse models are aggregated into a dense model in each round,
which is subsequently pruned in an iterative process. CS requires little
computation overhead on the top of vanilla FL for both the server and the
clients. We demonstrate that CS is an approximation of vanilla FL and, thus,
its models perform well. We evaluate CS experimentally with two popular FL
benchmark datasets. CS achieves substantial reduction in bidirectional
communication, while achieving performance comparable with vanilla FL. In
addition, CS outperforms baseline pruning mechanisms for FL.
| [
{
"version": "v1",
"created": "Fri, 10 Mar 2023 23:07:02 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Jiang",
"Xiaopeng",
""
],
[
"Borcea",
"Cristian",
""
]
] | TITLE: Complement Sparsification: Low-Overhead Model Pruning for Federated
Learning
ABSTRACT: Federated Learning (FL) is a privacy-preserving distributed deep learning
paradigm that involves substantial communication and computation effort, which
is a problem for resource-constrained mobile and IoT devices. Model
pruning/sparsification develops sparse models that could solve this problem,
but existing sparsification solutions cannot satisfy at the same time the
requirements for low bidirectional communication overhead between the server
and the clients, low computation overhead at the clients, and good model
accuracy, under the FL assumption that the server does not have access to raw
data to fine-tune the pruned models. We propose Complement Sparsification (CS),
a pruning mechanism that satisfies all these requirements through a
complementary and collaborative pruning done at the server and the clients. At
each round, CS creates a global sparse model that contains the weights that
capture the general data distribution of all clients, while the clients create
local sparse models with the weights pruned from the global model to capture
the local trends. For improved model performance, these two types of
complementary sparse models are aggregated into a dense model in each round,
which is subsequently pruned in an iterative process. CS requires little
computation overhead on the top of vanilla FL for both the server and the
clients. We demonstrate that CS is an approximation of vanilla FL and, thus,
its models perform well. We evaluate CS experimentally with two popular FL
benchmark datasets. CS achieves substantial reduction in bidirectional
communication, while achieving performance comparable with vanilla FL. In
addition, CS outperforms baseline pruning mechanisms for FL.
| no_new_dataset | 0.946101 |
2303.08233 | Timothy Yu | Rindranirina Ramamonjison, Timothy T. Yu, Raymond Li, Haley Li,
Giuseppe Carenini, Bissan Ghaddar, Shiqi He, Mahdi Mostajabdaveh, Amin
Banitalebi-Dehkordi, Zirui Zhou, Yong Zhang | NL4Opt Competition: Formulating Optimization Problems Based on Their
Natural Language Descriptions | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Natural Language for Optimization (NL4Opt) Competition was created to
investigate methods of extracting the meaning and formulation of an
optimization problem based on its text description. Specifically, the goal of
the competition is to increase the accessibility and usability of optimization
solvers by allowing non-experts to interface with them using natural language.
We separate this challenging goal into two sub-tasks: (1) recognize and label
the semantic entities that correspond to the components of the optimization
problem; (2) generate a meaning representation (i.e., a logical form) of the
problem from its detected problem entities. The first task aims to reduce
ambiguity by detecting and tagging the entities of the optimization problems.
The second task creates an intermediate representation of the linear
programming (LP) problem that is converted into a format that can be used by
commercial solvers. In this report, we present the LP word problem dataset and
shared tasks for the NeurIPS 2022 competition. Furthermore, we investigate and
compare the performance of the ChatGPT large language model against the winning
solutions. Through this competition, we hope to bring interest towards the
development of novel machine learning applications and datasets for
optimization modeling.
| [
{
"version": "v1",
"created": "Tue, 14 Mar 2023 20:59:04 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 01:10:12 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Ramamonjison",
"Rindranirina",
""
],
[
"Yu",
"Timothy T.",
""
],
[
"Li",
"Raymond",
""
],
[
"Li",
"Haley",
""
],
[
"Carenini",
"Giuseppe",
""
],
[
"Ghaddar",
"Bissan",
""
],
[
"He",
"Shiqi",
""
],
[
"Mostajabdaveh",
"Mahdi",
""
],
[
"Banitalebi-Dehkordi",
"Amin",
""
],
[
"Zhou",
"Zirui",
""
],
[
"Zhang",
"Yong",
""
]
] | TITLE: NL4Opt Competition: Formulating Optimization Problems Based on Their
Natural Language Descriptions
ABSTRACT: The Natural Language for Optimization (NL4Opt) Competition was created to
investigate methods of extracting the meaning and formulation of an
optimization problem based on its text description. Specifically, the goal of
the competition is to increase the accessibility and usability of optimization
solvers by allowing non-experts to interface with them using natural language.
We separate this challenging goal into two sub-tasks: (1) recognize and label
the semantic entities that correspond to the components of the optimization
problem; (2) generate a meaning representation (i.e., a logical form) of the
problem from its detected problem entities. The first task aims to reduce
ambiguity by detecting and tagging the entities of the optimization problems.
The second task creates an intermediate representation of the linear
programming (LP) problem that is converted into a format that can be used by
commercial solvers. In this report, we present the LP word problem dataset and
shared tasks for the NeurIPS 2022 competition. Furthermore, we investigate and
compare the performance of the ChatGPT large language model against the winning
solutions. Through this competition, we hope to bring interest towards the
development of novel machine learning applications and datasets for
optimization modeling.
| new_dataset | 0.960212 |
2303.10752 | Haozhe Si | Haozhe Si, Bin Zhao, Dong Wang, Yunpeng Gao, Mulin Chen, Zhigang Wang,
Xuelong Li | Fully Self-Supervised Depth Estimation from Defocus Clue | CVPR 2023 camera-ready version. The code is released at
https://github.com/Ehzoahis/DEReD | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Depth-from-defocus (DFD), modeling the relationship between depth and defocus
pattern in images, has demonstrated promising performance in depth estimation.
Recently, several self-supervised works try to overcome the difficulties in
acquiring accurate depth ground-truth. However, they depend on the all-in-focus
(AIF) images, which cannot be captured in real-world scenarios. Such limitation
discourages the applications of DFD methods. To tackle this issue, we propose a
completely self-supervised framework that estimates depth purely from a sparse
focal stack. We show that our framework circumvents the needs for the depth and
AIF image ground-truth, and receives superior predictions, thus closing the gap
between the theoretical success of DFD works and their applications in the real
world. In particular, we propose (i) a more realistic setting for DFD tasks,
where no depth or AIF image ground-truth is available; (ii) a novel
self-supervision framework that provides reliable predictions of depth and AIF
image under the challenging setting. The proposed framework uses a neural model
to predict the depth and AIF image, and utilizes an optical model to validate
and refine the prediction. We verify our framework on three benchmark datasets
with rendered focal stacks and real focal stacks. Qualitative and quantitative
evaluations show that our method provides a strong baseline for self-supervised
DFD tasks.
| [
{
"version": "v1",
"created": "Sun, 19 Mar 2023 19:59:48 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Mar 2023 06:02:06 GMT"
},
{
"version": "v3",
"created": "Thu, 23 Mar 2023 00:47:10 GMT"
},
{
"version": "v4",
"created": "Mon, 27 Mar 2023 04:26:50 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Si",
"Haozhe",
""
],
[
"Zhao",
"Bin",
""
],
[
"Wang",
"Dong",
""
],
[
"Gao",
"Yunpeng",
""
],
[
"Chen",
"Mulin",
""
],
[
"Wang",
"Zhigang",
""
],
[
"Li",
"Xuelong",
""
]
] | TITLE: Fully Self-Supervised Depth Estimation from Defocus Clue
ABSTRACT: Depth-from-defocus (DFD), modeling the relationship between depth and defocus
pattern in images, has demonstrated promising performance in depth estimation.
Recently, several self-supervised works try to overcome the difficulties in
acquiring accurate depth ground-truth. However, they depend on the all-in-focus
(AIF) images, which cannot be captured in real-world scenarios. Such limitation
discourages the applications of DFD methods. To tackle this issue, we propose a
completely self-supervised framework that estimates depth purely from a sparse
focal stack. We show that our framework circumvents the needs for the depth and
AIF image ground-truth, and receives superior predictions, thus closing the gap
between the theoretical success of DFD works and their applications in the real
world. In particular, we propose (i) a more realistic setting for DFD tasks,
where no depth or AIF image ground-truth is available; (ii) a novel
self-supervision framework that provides reliable predictions of depth and AIF
image under the challenging setting. The proposed framework uses a neural model
to predict the depth and AIF image, and utilizes an optical model to validate
and refine the prediction. We verify our framework on three benchmark datasets
with rendered focal stacks and real focal stacks. Qualitative and quantitative
evaluations show that our method provides a strong baseline for self-supervised
DFD tasks.
| no_new_dataset | 0.946892 |
2303.10967 | Xiaokang Chen | Xiaokang Chen, Yajie Xing and Gang Zeng | Real-time 3D Semantic Scene Completion Via Feature Aggregation and
Conditioned Prediction | Accepted by ICIP | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Semantic Scene Completion (SSC) aims to simultaneously predict the volumetric
occupancy and semantic category of a 3D scene. In this paper, we propose a
real-time semantic scene completion method with a feature aggregation strategy
and conditioned prediction module. Feature aggregation fuses feature with
different receptive fields and gathers context to improve scene completion
performance. And the conditioned prediction module adopts a two-step prediction
scheme that takes volumetric occupancy as a condition to enhance semantic
completion prediction. We conduct experiments on three recognized benchmarks
NYU, NYUCAD, and SUNCG. Our method achieves competitive performance at a speed
of 110 FPS on one GTX 1080 Ti GPU.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:41:50 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 07:36:34 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Chen",
"Xiaokang",
""
],
[
"Xing",
"Yajie",
""
],
[
"Zeng",
"Gang",
""
]
] | TITLE: Real-time 3D Semantic Scene Completion Via Feature Aggregation and
Conditioned Prediction
ABSTRACT: Semantic Scene Completion (SSC) aims to simultaneously predict the volumetric
occupancy and semantic category of a 3D scene. In this paper, we propose a
real-time semantic scene completion method with a feature aggregation strategy
and conditioned prediction module. Feature aggregation fuses feature with
different receptive fields and gathers context to improve scene completion
performance. And the conditioned prediction module adopts a two-step prediction
scheme that takes volumetric occupancy as a condition to enhance semantic
completion prediction. We conduct experiments on three recognized benchmarks
NYU, NYUCAD, and SUNCG. Our method achieves competitive performance at a speed
of 110 FPS on one GTX 1080 Ti GPU.
| no_new_dataset | 0.956227 |
2303.11548 | Sarthak Bhagat | Sahil Goyal, Shagun Uppal, Sarthak Bhagat, Yi Yu, Yifang Yin, Rajiv
Ratn Shah | Emotionally Enhanced Talking Face Generation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several works have developed end-to-end pipelines for generating lip-synced
talking faces with various real-world applications, such as teaching and
language translation in videos. However, these prior works fail to create
realistic-looking videos since they focus little on people's expressions and
emotions. Moreover, these methods' effectiveness largely depends on the faces
in the training dataset, which means they may not perform well on unseen faces.
To mitigate this, we build a talking face generation framework conditioned on a
categorical emotion to generate videos with appropriate expressions, making
them more realistic and convincing. With a broad range of six emotions, i.e.,
\emph{happiness}, \emph{sadness}, \emph{fear}, \emph{anger}, \emph{disgust},
and \emph{neutral}, we show that our model can adapt to arbitrary identities,
emotions, and languages. Our proposed framework is equipped with a
user-friendly web interface with a real-time experience for talking face
generation with emotions. We also conduct a user study for subjective
evaluation of our interface's usability, design, and functionality. Project
page: https://midas.iiitd.edu.in/emo/
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 02:33:27 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 04:41:50 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Goyal",
"Sahil",
""
],
[
"Uppal",
"Shagun",
""
],
[
"Bhagat",
"Sarthak",
""
],
[
"Yu",
"Yi",
""
],
[
"Yin",
"Yifang",
""
],
[
"Shah",
"Rajiv Ratn",
""
]
] | TITLE: Emotionally Enhanced Talking Face Generation
ABSTRACT: Several works have developed end-to-end pipelines for generating lip-synced
talking faces with various real-world applications, such as teaching and
language translation in videos. However, these prior works fail to create
realistic-looking videos since they focus little on people's expressions and
emotions. Moreover, these methods' effectiveness largely depends on the faces
in the training dataset, which means they may not perform well on unseen faces.
To mitigate this, we build a talking face generation framework conditioned on a
categorical emotion to generate videos with appropriate expressions, making
them more realistic and convincing. With a broad range of six emotions, i.e.,
\emph{happiness}, \emph{sadness}, \emph{fear}, \emph{anger}, \emph{disgust},
and \emph{neutral}, we show that our model can adapt to arbitrary identities,
emotions, and languages. Our proposed framework is equipped with a
user-friendly web interface with a real-time experience for talking face
generation with emotions. We also conduct a user study for subjective
evaluation of our interface's usability, design, and functionality. Project
page: https://midas.iiitd.edu.in/emo/
| no_new_dataset | 0.94625 |
2303.11694 | Shi Linfeng | Linfeng Shi, Yan Li, Xi Zhu | Anchor Free remote sensing detector based on solving discrete polar
coordinate equation | 20 pages,15 figures | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As the rapid development of depth learning, object detection in aviatic
remote sensing images has become increasingly popular in recent years. Most of
the current Anchor Free detectors based on key point detection sampling
directly regression and classification features, with the design of object loss
function based on the horizontal bounding box. It is more challenging for
complex and diverse aviatic remote sensing object. In this paper, we propose an
Anchor Free aviatic remote sensing object detector (BWP-Det) to detect rotating
and multi-scale object. Specifically, we design a interactive
double-branch(IDB) up-sampling network, in which one branch gradually
up-sampling is used for the prediction of Heatmap, and the other branch is used
for the regression of boundary box parameters. We improve a weighted
multi-scale convolution (WmConv) in order to highlight the difference between
foreground and background. We extracted Pixel level attention features from the
middle layer to guide the two branches to pay attention to effective object
information in the sampling process. Finally, referring to the calculation idea
of horizontal IoU, we design a rotating IoU based on the split polar coordinate
plane, namely JIoU, which is expressed as the intersection ratio following
discretization of the inner ellipse of the rotating bounding box, to solve the
correlation between angle and side length in the regression process of the
rotating bounding box. Ultimately, BWP-Det, our experiments on DOTA, UCAS-AOD
and NWPU VHR-10 datasets show, achieves advanced performance with simpler
models and fewer regression parameters.
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 09:28:47 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 06:43:43 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Shi",
"Linfeng",
""
],
[
"Li",
"Yan",
""
],
[
"Zhu",
"Xi",
""
]
] | TITLE: Anchor Free remote sensing detector based on solving discrete polar
coordinate equation
ABSTRACT: As the rapid development of depth learning, object detection in aviatic
remote sensing images has become increasingly popular in recent years. Most of
the current Anchor Free detectors based on key point detection sampling
directly regression and classification features, with the design of object loss
function based on the horizontal bounding box. It is more challenging for
complex and diverse aviatic remote sensing object. In this paper, we propose an
Anchor Free aviatic remote sensing object detector (BWP-Det) to detect rotating
and multi-scale object. Specifically, we design a interactive
double-branch(IDB) up-sampling network, in which one branch gradually
up-sampling is used for the prediction of Heatmap, and the other branch is used
for the regression of boundary box parameters. We improve a weighted
multi-scale convolution (WmConv) in order to highlight the difference between
foreground and background. We extracted Pixel level attention features from the
middle layer to guide the two branches to pay attention to effective object
information in the sampling process. Finally, referring to the calculation idea
of horizontal IoU, we design a rotating IoU based on the split polar coordinate
plane, namely JIoU, which is expressed as the intersection ratio following
discretization of the inner ellipse of the rotating bounding box, to solve the
correlation between angle and side length in the regression process of the
rotating bounding box. Ultimately, BWP-Det, our experiments on DOTA, UCAS-AOD
and NWPU VHR-10 datasets show, achieves advanced performance with simpler
models and fewer regression parameters.
| no_new_dataset | 0.951323 |
2303.11739 | Nicola Strisciuglio | Maria Leyva-Vallina, Nicola Strisciuglio, Nicolai Petkov | Data-efficient Large Scale Place Recognition with Graded Similarity
Supervision | Accepted at CVPR 2023 | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Visual place recognition (VPR) is a fundamental task of computer vision for
visual localization. Existing methods are trained using image pairs that either
depict the same place or not. Such a binary indication does not consider
continuous relations of similarity between images of the same place taken from
different positions, determined by the continuous nature of camera pose. The
binary similarity induces a noisy supervision signal into the training of VPR
methods, which stall in local minima and require expensive hard mining
algorithms to guarantee convergence. Motivated by the fact that two images of
the same place only partially share visual cues due to camera pose differences,
we deploy an automatic re-annotation strategy to re-label VPR datasets. We
compute graded similarity labels for image pairs based on available
localization metadata. Furthermore, we propose a new Generalized Contrastive
Loss (GCL) that uses graded similarity labels for training contrastive
networks. We demonstrate that the use of the new labels and GCL allow to
dispense from hard-pair mining, and to train image descriptors that perform
better in VPR by nearest neighbor search, obtaining superior or comparable
results than methods that require expensive hard-pair mining and re-ranking
techniques. Code and models available at:
https://github.com/marialeyvallina/generalized_contrastive_loss
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 10:56:57 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 18:54:41 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Leyva-Vallina",
"Maria",
""
],
[
"Strisciuglio",
"Nicola",
""
],
[
"Petkov",
"Nicolai",
""
]
] | TITLE: Data-efficient Large Scale Place Recognition with Graded Similarity
Supervision
ABSTRACT: Visual place recognition (VPR) is a fundamental task of computer vision for
visual localization. Existing methods are trained using image pairs that either
depict the same place or not. Such a binary indication does not consider
continuous relations of similarity between images of the same place taken from
different positions, determined by the continuous nature of camera pose. The
binary similarity induces a noisy supervision signal into the training of VPR
methods, which stall in local minima and require expensive hard mining
algorithms to guarantee convergence. Motivated by the fact that two images of
the same place only partially share visual cues due to camera pose differences,
we deploy an automatic re-annotation strategy to re-label VPR datasets. We
compute graded similarity labels for image pairs based on available
localization metadata. Furthermore, we propose a new Generalized Contrastive
Loss (GCL) that uses graded similarity labels for training contrastive
networks. We demonstrate that the use of the new labels and GCL allow to
dispense from hard-pair mining, and to train image descriptors that perform
better in VPR by nearest neighbor search, obtaining superior or comparable
results than methods that require expensive hard-pair mining and re-ranking
techniques. Code and models available at:
https://github.com/marialeyvallina/generalized_contrastive_loss
| no_new_dataset | 0.95096 |
2303.11749 | Zhenyu Wang | Zhenyu Wang, Yali Li, Xi Chen, Ser-Nam Lim, Antonio Torralba,
Hengshuang Zhao, Shengjin Wang | Detecting Everything in the Open World: Towards Universal Object
Detection | Accepted by CVPR2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In this paper, we formally address universal object detection, which aims to
detect every scene and predict every category. The dependence on human
annotations, the limited visual information, and the novel categories in the
open world severely restrict the universality of traditional detectors. We
propose UniDetector, a universal object detector that has the ability to
recognize enormous categories in the open world. The critical points for the
universality of UniDetector are: 1) it leverages images of multiple sources and
heterogeneous label spaces for training through the alignment of image and text
spaces, which guarantees sufficient information for universal representations.
2) it generalizes to the open world easily while keeping the balance between
seen and unseen classes, thanks to abundant information from both vision and
language modalities. 3) it further promotes the generalization ability to novel
categories through our proposed decoupling training manner and probability
calibration. These contributions allow UniDetector to detect over 7k
categories, the largest measurable category size so far, with only about 500
classes participating in training. Our UniDetector behaves the strong zero-shot
generalization ability on large-vocabulary datasets like LVIS, ImageNetBoxes,
and VisualGenome - it surpasses the traditional supervised baselines by more
than 4\% on average without seeing any corresponding images. On 13 public
detection datasets with various scenes, UniDetector also achieves
state-of-the-art performance with only a 3\% amount of training data.
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 11:15:03 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 03:56:47 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Zhenyu",
""
],
[
"Li",
"Yali",
""
],
[
"Chen",
"Xi",
""
],
[
"Lim",
"Ser-Nam",
""
],
[
"Torralba",
"Antonio",
""
],
[
"Zhao",
"Hengshuang",
""
],
[
"Wang",
"Shengjin",
""
]
] | TITLE: Detecting Everything in the Open World: Towards Universal Object
Detection
ABSTRACT: In this paper, we formally address universal object detection, which aims to
detect every scene and predict every category. The dependence on human
annotations, the limited visual information, and the novel categories in the
open world severely restrict the universality of traditional detectors. We
propose UniDetector, a universal object detector that has the ability to
recognize enormous categories in the open world. The critical points for the
universality of UniDetector are: 1) it leverages images of multiple sources and
heterogeneous label spaces for training through the alignment of image and text
spaces, which guarantees sufficient information for universal representations.
2) it generalizes to the open world easily while keeping the balance between
seen and unseen classes, thanks to abundant information from both vision and
language modalities. 3) it further promotes the generalization ability to novel
categories through our proposed decoupling training manner and probability
calibration. These contributions allow UniDetector to detect over 7k
categories, the largest measurable category size so far, with only about 500
classes participating in training. Our UniDetector behaves the strong zero-shot
generalization ability on large-vocabulary datasets like LVIS, ImageNetBoxes,
and VisualGenome - it surpasses the traditional supervised baselines by more
than 4\% on average without seeing any corresponding images. On 13 public
detection datasets with various scenes, UniDetector also achieves
state-of-the-art performance with only a 3\% amount of training data.
| no_new_dataset | 0.942929 |
2303.11791 | Yihao Wang | Yihao Wang, Zhigang Wang, Bin Zhao, Dong Wang, Mulin Chen, Xuelong Li | Propagate And Calibrate: Real-time Passive Non-line-of-sight Tracking | CVPR 2023 camera-ready version. Codes and dataset are available at
https://againstentropy.github.io/NLOS-Track/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Non-line-of-sight (NLOS) tracking has drawn increasing attention in recent
years, due to its ability to detect object motion out of sight. Most previous
works on NLOS tracking rely on active illumination, e.g., laser, and suffer
from high cost and elaborate experimental conditions. Besides, these techniques
are still far from practical application due to oversimplified settings. In
contrast, we propose a purely passive method to track a person walking in an
invisible room by only observing a relay wall, which is more in line with real
application scenarios, e.g., security. To excavate imperceptible changes in
videos of the relay wall, we introduce difference frames as an essential
carrier of temporal-local motion messages. In addition, we propose PAC-Net,
which consists of alternating propagation and calibration, making it capable of
leveraging both dynamic and static messages on a frame-level granularity. To
evaluate the proposed method, we build and publish the first dynamic passive
NLOS tracking dataset, NLOS-Track, which fills the vacuum of realistic NLOS
datasets. NLOS-Track contains thousands of NLOS video clips and corresponding
trajectories. Both real-shot and synthetic data are included. Our codes and
dataset are available at https://againstentropy.github.io/NLOS-Track/.
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 12:18:57 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 10:11:31 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Yihao",
""
],
[
"Wang",
"Zhigang",
""
],
[
"Zhao",
"Bin",
""
],
[
"Wang",
"Dong",
""
],
[
"Chen",
"Mulin",
""
],
[
"Li",
"Xuelong",
""
]
] | TITLE: Propagate And Calibrate: Real-time Passive Non-line-of-sight Tracking
ABSTRACT: Non-line-of-sight (NLOS) tracking has drawn increasing attention in recent
years, due to its ability to detect object motion out of sight. Most previous
works on NLOS tracking rely on active illumination, e.g., laser, and suffer
from high cost and elaborate experimental conditions. Besides, these techniques
are still far from practical application due to oversimplified settings. In
contrast, we propose a purely passive method to track a person walking in an
invisible room by only observing a relay wall, which is more in line with real
application scenarios, e.g., security. To excavate imperceptible changes in
videos of the relay wall, we introduce difference frames as an essential
carrier of temporal-local motion messages. In addition, we propose PAC-Net,
which consists of alternating propagation and calibration, making it capable of
leveraging both dynamic and static messages on a frame-level granularity. To
evaluate the proposed method, we build and publish the first dynamic passive
NLOS tracking dataset, NLOS-Track, which fills the vacuum of realistic NLOS
datasets. NLOS-Track contains thousands of NLOS video clips and corresponding
trajectories. Both real-shot and synthetic data are included. Our codes and
dataset are available at https://againstentropy.github.io/NLOS-Track/.
| new_dataset | 0.956796 |
2303.11921 | Dingkang Yang | Dingkang Yang, Zhaoyu Chen, Yuzheng Wang, Shunli Wang, Mingcheng Li,
Siao Liu, Xiao Zhao, Shuai Huang, Zhiyan Dong, Peng Zhai, Lihua Zhang | Context De-confounded Emotion Recognition | Accepted by CVPR 2023. CCIM is available at
https://github.com/ydk122024/CCIM | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Context-Aware Emotion Recognition (CAER) is a crucial and challenging task
that aims to perceive the emotional states of the target person with contextual
information. Recent approaches invariably focus on designing sophisticated
architectures or mechanisms to extract seemingly meaningful representations
from subjects and contexts. However, a long-overlooked issue is that a context
bias in existing datasets leads to a significantly unbalanced distribution of
emotional states among different context scenarios. Concretely, the harmful
bias is a confounder that misleads existing models to learn spurious
correlations based on conventional likelihood estimation, significantly
limiting the models' performance. To tackle the issue, this paper provides a
causality-based perspective to disentangle the models from the impact of such
bias, and formulate the causalities among variables in the CAER task via a
tailored causal graph. Then, we propose a Contextual Causal Intervention Module
(CCIM) based on the backdoor adjustment to de-confound the confounder and
exploit the true causal effect for model training. CCIM is plug-in and
model-agnostic, which improves diverse state-of-the-art approaches by
considerable margins. Extensive experiments on three benchmark datasets
demonstrate the effectiveness of our CCIM and the significance of causal
insight.
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 15:12:20 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 07:18:05 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yang",
"Dingkang",
""
],
[
"Chen",
"Zhaoyu",
""
],
[
"Wang",
"Yuzheng",
""
],
[
"Wang",
"Shunli",
""
],
[
"Li",
"Mingcheng",
""
],
[
"Liu",
"Siao",
""
],
[
"Zhao",
"Xiao",
""
],
[
"Huang",
"Shuai",
""
],
[
"Dong",
"Zhiyan",
""
],
[
"Zhai",
"Peng",
""
],
[
"Zhang",
"Lihua",
""
]
] | TITLE: Context De-confounded Emotion Recognition
ABSTRACT: Context-Aware Emotion Recognition (CAER) is a crucial and challenging task
that aims to perceive the emotional states of the target person with contextual
information. Recent approaches invariably focus on designing sophisticated
architectures or mechanisms to extract seemingly meaningful representations
from subjects and contexts. However, a long-overlooked issue is that a context
bias in existing datasets leads to a significantly unbalanced distribution of
emotional states among different context scenarios. Concretely, the harmful
bias is a confounder that misleads existing models to learn spurious
correlations based on conventional likelihood estimation, significantly
limiting the models' performance. To tackle the issue, this paper provides a
causality-based perspective to disentangle the models from the impact of such
bias, and formulate the causalities among variables in the CAER task via a
tailored causal graph. Then, we propose a Contextual Causal Intervention Module
(CCIM) based on the backdoor adjustment to de-confound the confounder and
exploit the true causal effect for model training. CCIM is plug-in and
model-agnostic, which improves diverse state-of-the-art approaches by
considerable margins. Extensive experiments on three benchmark datasets
demonstrate the effectiveness of our CCIM and the significance of causal
insight.
| no_new_dataset | 0.944125 |
2303.12337 | Tuong Do Khanh Long | Nhat Le, Thang Pham, Tuong Do, Erman Tjiputra, Quang D. Tran, Anh
Nguyen | Music-Driven Group Choreography | accepted in CVPR 2023 | null | null | null | cs.MM cs.CV cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | Music-driven choreography is a challenging problem with a wide variety of
industrial applications. Recently, many methods have been proposed to
synthesize dance motions from music for a single dancer. However, generating
dance motion for a group remains an open problem. In this paper, we present
$\rm AIOZ-GDANCE$, a new large-scale dataset for music-driven group dance
generation. Unlike existing datasets that only support single dance, our new
dataset contains group dance videos, hence supporting the study of group
choreography. We propose a semi-autonomous labeling method with humans in the
loop to obtain the 3D ground truth for our dataset. The proposed dataset
consists of 16.7 hours of paired music and 3D motion from in-the-wild videos,
covering 7 dance styles and 16 music genres. We show that naively applying
single dance generation technique to creating group dance motion may lead to
unsatisfactory results, such as inconsistent movements and collisions between
dancers. Based on our new dataset, we propose a new method that takes an input
music sequence and a set of 3D positions of dancers to efficiently produce
multiple group-coherent choreographies. We propose new evaluation metrics for
measuring group dance quality and perform intensive experiments to demonstrate
the effectiveness of our method. Our project facilitates future research on
group dance generation and is available at:
https://aioz-ai.github.io/AIOZ-GDANCE/
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 06:26:56 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 01:59:41 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Le",
"Nhat",
""
],
[
"Pham",
"Thang",
""
],
[
"Do",
"Tuong",
""
],
[
"Tjiputra",
"Erman",
""
],
[
"Tran",
"Quang D.",
""
],
[
"Nguyen",
"Anh",
""
]
] | TITLE: Music-Driven Group Choreography
ABSTRACT: Music-driven choreography is a challenging problem with a wide variety of
industrial applications. Recently, many methods have been proposed to
synthesize dance motions from music for a single dancer. However, generating
dance motion for a group remains an open problem. In this paper, we present
$\rm AIOZ-GDANCE$, a new large-scale dataset for music-driven group dance
generation. Unlike existing datasets that only support single dance, our new
dataset contains group dance videos, hence supporting the study of group
choreography. We propose a semi-autonomous labeling method with humans in the
loop to obtain the 3D ground truth for our dataset. The proposed dataset
consists of 16.7 hours of paired music and 3D motion from in-the-wild videos,
covering 7 dance styles and 16 music genres. We show that naively applying
single dance generation technique to creating group dance motion may lead to
unsatisfactory results, such as inconsistent movements and collisions between
dancers. Based on our new dataset, we propose a new method that takes an input
music sequence and a set of 3D positions of dancers to efficiently produce
multiple group-coherent choreographies. We propose new evaluation metrics for
measuring group dance quality and perform intensive experiments to demonstrate
the effectiveness of our method. Our project facilitates future research on
group dance generation and is available at:
https://aioz-ai.github.io/AIOZ-GDANCE/
| new_dataset | 0.961929 |
2303.12368 | JunYong Choi | JunYong Choi and SeokYeong Lee and Haesol Park and Seung-Won Jung and
Ig-Jae Kim and Junghyun Cho | MAIR: Multi-view Attention Inverse Rendering with 3D Spatially-Varying
Lighting Estimation | Accepted by CVPR 2023; Project Page is
https://bring728.github.io/mair.project/ | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a scene-level inverse rendering framework that uses multi-view
images to decompose the scene into geometry, a SVBRDF, and 3D spatially-varying
lighting. Because multi-view images provide a variety of information about the
scene, multi-view images in object-level inverse rendering have been taken for
granted. However, owing to the absence of multi-view HDR synthetic dataset,
scene-level inverse rendering has mainly been studied using single-view image.
We were able to successfully perform scene-level inverse rendering using
multi-view images by expanding OpenRooms dataset and designing efficient
pipelines to handle multi-view images, and splitting spatially-varying
lighting. Our experiments show that the proposed method not only achieves
better performance than single-view-based methods, but also achieves robust
performance on unseen real-world scene. Also, our sophisticated 3D
spatially-varying lighting volume allows for photorealistic object insertion in
any 3D location.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 08:07:28 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 04:32:11 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Choi",
"JunYong",
""
],
[
"Lee",
"SeokYeong",
""
],
[
"Park",
"Haesol",
""
],
[
"Jung",
"Seung-Won",
""
],
[
"Kim",
"Ig-Jae",
""
],
[
"Cho",
"Junghyun",
""
]
] | TITLE: MAIR: Multi-view Attention Inverse Rendering with 3D Spatially-Varying
Lighting Estimation
ABSTRACT: We propose a scene-level inverse rendering framework that uses multi-view
images to decompose the scene into geometry, a SVBRDF, and 3D spatially-varying
lighting. Because multi-view images provide a variety of information about the
scene, multi-view images in object-level inverse rendering have been taken for
granted. However, owing to the absence of multi-view HDR synthetic dataset,
scene-level inverse rendering has mainly been studied using single-view image.
We were able to successfully perform scene-level inverse rendering using
multi-view images by expanding OpenRooms dataset and designing efficient
pipelines to handle multi-view images, and splitting spatially-varying
lighting. Our experiments show that the proposed method not only achieves
better performance than single-view-based methods, but also achieves robust
performance on unseen real-world scene. Also, our sophisticated 3D
spatially-varying lighting volume allows for photorealistic object insertion in
any 3D location.
| no_new_dataset | 0.944944 |
2303.12423 | Xin Gu | Xin Gu, Guang Chen, Yufei Wang, Libo Zhang, Tiejian Luo, Longyin Wen | Text with Knowledge Graph Augmented Transformer for Video Captioning | Accepted by CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video captioning aims to describe the content of videos using natural
language. Although significant progress has been made, there is still much room
to improve the performance for real-world applications, mainly due to the
long-tail words challenge. In this paper, we propose a text with knowledge
graph augmented transformer (TextKG) for video captioning. Notably, TextKG is a
two-stream transformer, formed by the external stream and internal stream. The
external stream is designed to absorb additional knowledge, which models the
interactions between the additional knowledge, e.g., pre-built knowledge graph,
and the built-in information of videos, e.g., the salient object regions,
speech transcripts, and video captions, to mitigate the long-tail words
challenge. Meanwhile, the internal stream is designed to exploit the
multi-modality information in videos (e.g., the appearance of video frames,
speech transcripts, and video captions) to ensure the quality of caption
results. In addition, the cross attention mechanism is also used in between the
two streams for sharing information. In this way, the two streams can help each
other for more accurate results. Extensive experiments conducted on four
challenging video captioning datasets, i.e., YouCookII, ActivityNet Captions,
MSRVTT, and MSVD, demonstrate that the proposed method performs favorably
against the state-of-the-art methods. Specifically, the proposed TextKG method
outperforms the best published results by improving 18.7% absolute CIDEr scores
on the YouCookII dataset.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 09:49:53 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 10:40:05 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Gu",
"Xin",
""
],
[
"Chen",
"Guang",
""
],
[
"Wang",
"Yufei",
""
],
[
"Zhang",
"Libo",
""
],
[
"Luo",
"Tiejian",
""
],
[
"Wen",
"Longyin",
""
]
] | TITLE: Text with Knowledge Graph Augmented Transformer for Video Captioning
ABSTRACT: Video captioning aims to describe the content of videos using natural
language. Although significant progress has been made, there is still much room
to improve the performance for real-world applications, mainly due to the
long-tail words challenge. In this paper, we propose a text with knowledge
graph augmented transformer (TextKG) for video captioning. Notably, TextKG is a
two-stream transformer, formed by the external stream and internal stream. The
external stream is designed to absorb additional knowledge, which models the
interactions between the additional knowledge, e.g., pre-built knowledge graph,
and the built-in information of videos, e.g., the salient object regions,
speech transcripts, and video captions, to mitigate the long-tail words
challenge. Meanwhile, the internal stream is designed to exploit the
multi-modality information in videos (e.g., the appearance of video frames,
speech transcripts, and video captions) to ensure the quality of caption
results. In addition, the cross attention mechanism is also used in between the
two streams for sharing information. In this way, the two streams can help each
other for more accurate results. Extensive experiments conducted on four
challenging video captioning datasets, i.e., YouCookII, ActivityNet Captions,
MSRVTT, and MSVD, demonstrate that the proposed method performs favorably
against the state-of-the-art methods. Specifically, the proposed TextKG method
outperforms the best published results by improving 18.7% absolute CIDEr scores
on the YouCookII dataset.
| no_new_dataset | 0.950273 |
2303.12641 | Frederik Pahde | Frederik Pahde, Maximilian Dreyer, Wojciech Samek, Sebastian
Lapuschkin | Reveal to Revise: An Explainable AI Life Cycle for Iterative Bias
Correction of Deep Models | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | State-of-the-art machine learning models often learn spurious correlations
embedded in the training data. This poses risks when deploying these models for
high-stake decision-making, such as in medical applications like skin cancer
detection. To tackle this problem, we propose Reveal to Revise (R2R), a
framework entailing the entire eXplainable Artificial Intelligence (XAI) life
cycle, enabling practitioners to iteratively identify, mitigate, and
(re-)evaluate spurious model behavior with a minimal amount of human
interaction. In the first step (1), R2R reveals model weaknesses by finding
outliers in attributions or through inspection of latent concepts learned by
the model. Secondly (2), the responsible artifacts are detected and spatially
localized in the input data, which is then leveraged to (3) revise the model
behavior. Concretely, we apply the methods of RRR, CDEP and ClArC for model
correction, and (4) (re-)evaluate the model's performance and remaining
sensitivity towards the artifact. Using two medical benchmark datasets for
Melanoma detection and bone age estimation, we apply our R2R framework to VGG,
ResNet and EfficientNet architectures and thereby reveal and correct real
dataset-intrinsic artifacts, as well as synthetic variants in a controlled
setting. Completing the XAI life cycle, we demonstrate multiple R2R iterations
to mitigate different biases. Code is available on
https://github.com/maxdreyer/Reveal2Revise.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 15:23:09 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 07:43:17 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Pahde",
"Frederik",
""
],
[
"Dreyer",
"Maximilian",
""
],
[
"Samek",
"Wojciech",
""
],
[
"Lapuschkin",
"Sebastian",
""
]
] | TITLE: Reveal to Revise: An Explainable AI Life Cycle for Iterative Bias
Correction of Deep Models
ABSTRACT: State-of-the-art machine learning models often learn spurious correlations
embedded in the training data. This poses risks when deploying these models for
high-stake decision-making, such as in medical applications like skin cancer
detection. To tackle this problem, we propose Reveal to Revise (R2R), a
framework entailing the entire eXplainable Artificial Intelligence (XAI) life
cycle, enabling practitioners to iteratively identify, mitigate, and
(re-)evaluate spurious model behavior with a minimal amount of human
interaction. In the first step (1), R2R reveals model weaknesses by finding
outliers in attributions or through inspection of latent concepts learned by
the model. Secondly (2), the responsible artifacts are detected and spatially
localized in the input data, which is then leveraged to (3) revise the model
behavior. Concretely, we apply the methods of RRR, CDEP and ClArC for model
correction, and (4) (re-)evaluate the model's performance and remaining
sensitivity towards the artifact. Using two medical benchmark datasets for
Melanoma detection and bone age estimation, we apply our R2R framework to VGG,
ResNet and EfficientNet architectures and thereby reveal and correct real
dataset-intrinsic artifacts, as well as synthetic variants in a controlled
setting. Completing the XAI life cycle, we demonstrate multiple R2R iterations
to mitigate different biases. Code is available on
https://github.com/maxdreyer/Reveal2Revise.
| no_new_dataset | 0.948251 |
2303.12810 | Shrivats Agrawal | Shrivats Agrawal | Are LLMs the Master of All Trades? : Exploring Domain-Agnostic Reasoning
Skills of LLMs | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The potential of large language models (LLMs) to reason like humans has been
a highly contested topic in Machine Learning communities. However, the
reasoning abilities of humans are multifaceted and can be seen in various
forms, including analogical, spatial and moral reasoning, among others. This
fact raises the question whether LLMs can perform equally well across all these
different domains. This research work aims to investigate the performance of
LLMs on different reasoning tasks by conducting experiments that directly use
or draw inspirations from existing datasets on analogical and spatial
reasoning. Additionally, to evaluate the ability of LLMs to reason like human,
their performance is evaluted on more open-ended, natural language questions.
My findings indicate that LLMs excel at analogical and moral reasoning, yet
struggle to perform as proficiently on spatial reasoning tasks. I believe these
experiments are crucial for informing the future development of LLMs,
particularly in contexts that require diverse reasoning proficiencies. By
shedding light on the reasoning abilities of LLMs, this study aims to push
forward our understanding of how they can better emulate the cognitive
abilities of humans.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 22:53:44 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Agrawal",
"Shrivats",
""
]
] | TITLE: Are LLMs the Master of All Trades? : Exploring Domain-Agnostic Reasoning
Skills of LLMs
ABSTRACT: The potential of large language models (LLMs) to reason like humans has been
a highly contested topic in Machine Learning communities. However, the
reasoning abilities of humans are multifaceted and can be seen in various
forms, including analogical, spatial and moral reasoning, among others. This
fact raises the question whether LLMs can perform equally well across all these
different domains. This research work aims to investigate the performance of
LLMs on different reasoning tasks by conducting experiments that directly use
or draw inspirations from existing datasets on analogical and spatial
reasoning. Additionally, to evaluate the ability of LLMs to reason like human,
their performance is evaluted on more open-ended, natural language questions.
My findings indicate that LLMs excel at analogical and moral reasoning, yet
struggle to perform as proficiently on spatial reasoning tasks. I believe these
experiments are crucial for informing the future development of LLMs,
particularly in contexts that require diverse reasoning proficiencies. By
shedding light on the reasoning abilities of LLMs, this study aims to push
forward our understanding of how they can better emulate the cognitive
abilities of humans.
| no_new_dataset | 0.9357 |
2303.12961 | Michael Wornow | Michael Wornow, Yizhe Xu, Rahul Thapa, Birju Patel, Ethan Steinberg,
Scott Fleming, Michael A. Pfeffer, Jason Fries, Nigam H. Shah | The Shaky Foundations of Clinical Foundation Models: A Survey of Large
Language Models and Foundation Models for EMRs | Reformatted figures, updated contributions | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The successes of foundation models such as ChatGPT and AlphaFold have spurred
significant interest in building similar models for electronic medical records
(EMRs) to improve patient care and hospital operations. However, recent hype
has obscured critical gaps in our understanding of these models' capabilities.
We review over 80 foundation models trained on non-imaging EMR data (i.e.
clinical text and/or structured data) and create a taxonomy delineating their
architectures, training data, and potential use cases. We find that most models
are trained on small, narrowly-scoped clinical datasets (e.g. MIMIC-III) or
broad, public biomedical corpora (e.g. PubMed) and are evaluated on tasks that
do not provide meaningful insights on their usefulness to health systems. In
light of these findings, we propose an improved evaluation framework for
measuring the benefits of clinical foundation models that is more closely
grounded to metrics that matter in healthcare.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 23:54:14 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 21:50:03 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wornow",
"Michael",
""
],
[
"Xu",
"Yizhe",
""
],
[
"Thapa",
"Rahul",
""
],
[
"Patel",
"Birju",
""
],
[
"Steinberg",
"Ethan",
""
],
[
"Fleming",
"Scott",
""
],
[
"Pfeffer",
"Michael A.",
""
],
[
"Fries",
"Jason",
""
],
[
"Shah",
"Nigam H.",
""
]
] | TITLE: The Shaky Foundations of Clinical Foundation Models: A Survey of Large
Language Models and Foundation Models for EMRs
ABSTRACT: The successes of foundation models such as ChatGPT and AlphaFold have spurred
significant interest in building similar models for electronic medical records
(EMRs) to improve patient care and hospital operations. However, recent hype
has obscured critical gaps in our understanding of these models' capabilities.
We review over 80 foundation models trained on non-imaging EMR data (i.e.
clinical text and/or structured data) and create a taxonomy delineating their
architectures, training data, and potential use cases. We find that most models
are trained on small, narrowly-scoped clinical datasets (e.g. MIMIC-III) or
broad, public biomedical corpora (e.g. PubMed) and are evaluated on tasks that
do not provide meaningful insights on their usefulness to health systems. In
light of these findings, we propose an improved evaluation framework for
measuring the benefits of clinical foundation models that is more closely
grounded to metrics that matter in healthcare.
| no_new_dataset | 0.947914 |
2303.13416 | Thong Nguyen | Thong Nguyen, Sean MacAvaney, Andrew Yates | A Unified Framework for Learned Sparse Retrieval | null | ECIR 2023 | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Learned sparse retrieval (LSR) is a family of first-stage retrieval methods
that are trained to generate sparse lexical representations of queries and
documents for use with an inverted index. Many LSR methods have been recently
introduced, with Splade models achieving state-of-the-art performance on
MSMarco. Despite similarities in their model architectures, many LSR methods
show substantial differences in effectiveness and efficiency. Differences in
the experimental setups and configurations used make it difficult to compare
the methods and derive insights. In this work, we analyze existing LSR methods
and identify key components to establish an LSR framework that unifies all LSR
methods under the same perspective. We then reproduce all prominent methods
using a common codebase and re-train them in the same environment, which allows
us to quantify how components of the framework affect effectiveness and
efficiency. We find that (1) including document term weighting is most
important for a method's effectiveness, (2) including query weighting has a
small positive impact, and (3) document expansion and query expansion have a
cancellation effect. As a result, we show how removing query expansion from a
state-of-the-art model can reduce latency significantly while maintaining
effectiveness on MSMarco and TripClick benchmarks. Our code is publicly
available at https://github.com/thongnt99/learned-sparse-retrieval
| [
{
"version": "v1",
"created": "Thu, 23 Mar 2023 16:38:18 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Nguyen",
"Thong",
""
],
[
"MacAvaney",
"Sean",
""
],
[
"Yates",
"Andrew",
""
]
] | TITLE: A Unified Framework for Learned Sparse Retrieval
ABSTRACT: Learned sparse retrieval (LSR) is a family of first-stage retrieval methods
that are trained to generate sparse lexical representations of queries and
documents for use with an inverted index. Many LSR methods have been recently
introduced, with Splade models achieving state-of-the-art performance on
MSMarco. Despite similarities in their model architectures, many LSR methods
show substantial differences in effectiveness and efficiency. Differences in
the experimental setups and configurations used make it difficult to compare
the methods and derive insights. In this work, we analyze existing LSR methods
and identify key components to establish an LSR framework that unifies all LSR
methods under the same perspective. We then reproduce all prominent methods
using a common codebase and re-train them in the same environment, which allows
us to quantify how components of the framework affect effectiveness and
efficiency. We find that (1) including document term weighting is most
important for a method's effectiveness, (2) including query weighting has a
small positive impact, and (3) document expansion and query expansion have a
cancellation effect. As a result, we show how removing query expansion from a
state-of-the-art model can reduce latency significantly while maintaining
effectiveness on MSMarco and TripClick benchmarks. Our code is publicly
available at https://github.com/thongnt99/learned-sparse-retrieval
| no_new_dataset | 0.946941 |
2303.13939 | Qinglin Zhang | Qinglin Zhang, Chong Deng, Jiaqing Liu, Hai Yu, Qian Chen, Wen Wang,
Zhijie Yan, Jinglin Liu, Yi Ren, Zhou Zhao | MUG: A General Meeting Understanding and Generation Benchmark | Paper accepted to the 2023 IEEE International Conference on
Acoustics, Speech, and Signal Processing (ICASSP 2023), Rhodes, Greece | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Listening to long video/audio recordings from video conferencing and online
courses for acquiring information is extremely inefficient. Even after ASR
systems transcribe recordings into long-form spoken language documents, reading
ASR transcripts only partly speeds up seeking information. It has been observed
that a range of NLP applications, such as keyphrase extraction, topic
segmentation, and summarization, significantly improve users' efficiency in
grasping important information. The meeting scenario is among the most valuable
scenarios for deploying these spoken language processing (SLP) capabilities.
However, the lack of large-scale public meeting datasets annotated for these
SLP tasks severely hinders their advancement. To prompt SLP advancement, we
establish a large-scale general Meeting Understanding and Generation Benchmark
(MUG) to benchmark the performance of a wide range of SLP tasks, including
topic segmentation, topic-level and session-level extractive summarization and
topic title generation, keyphrase extraction, and action item detection. To
facilitate the MUG benchmark, we construct and release a large-scale meeting
dataset for comprehensive long-form SLP development, the AliMeeting4MUG Corpus,
which consists of 654 recorded Mandarin meeting sessions with diverse topic
coverage, with manual annotations for SLP tasks on manual transcripts of
meeting recordings. To the best of our knowledge, the AliMeeting4MUG Corpus is
so far the largest meeting corpus in scale and facilitates most SLP tasks. In
this paper, we provide a detailed introduction of this corpus, SLP tasks and
evaluation methods, baseline systems and their performance.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 11:52:25 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 03:51:52 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Qinglin",
""
],
[
"Deng",
"Chong",
""
],
[
"Liu",
"Jiaqing",
""
],
[
"Yu",
"Hai",
""
],
[
"Chen",
"Qian",
""
],
[
"Wang",
"Wen",
""
],
[
"Yan",
"Zhijie",
""
],
[
"Liu",
"Jinglin",
""
],
[
"Ren",
"Yi",
""
],
[
"Zhao",
"Zhou",
""
]
] | TITLE: MUG: A General Meeting Understanding and Generation Benchmark
ABSTRACT: Listening to long video/audio recordings from video conferencing and online
courses for acquiring information is extremely inefficient. Even after ASR
systems transcribe recordings into long-form spoken language documents, reading
ASR transcripts only partly speeds up seeking information. It has been observed
that a range of NLP applications, such as keyphrase extraction, topic
segmentation, and summarization, significantly improve users' efficiency in
grasping important information. The meeting scenario is among the most valuable
scenarios for deploying these spoken language processing (SLP) capabilities.
However, the lack of large-scale public meeting datasets annotated for these
SLP tasks severely hinders their advancement. To prompt SLP advancement, we
establish a large-scale general Meeting Understanding and Generation Benchmark
(MUG) to benchmark the performance of a wide range of SLP tasks, including
topic segmentation, topic-level and session-level extractive summarization and
topic title generation, keyphrase extraction, and action item detection. To
facilitate the MUG benchmark, we construct and release a large-scale meeting
dataset for comprehensive long-form SLP development, the AliMeeting4MUG Corpus,
which consists of 654 recorded Mandarin meeting sessions with diverse topic
coverage, with manual annotations for SLP tasks on manual transcripts of
meeting recordings. To the best of our knowledge, the AliMeeting4MUG Corpus is
so far the largest meeting corpus in scale and facilitates most SLP tasks. In
this paper, we provide a detailed introduction of this corpus, SLP tasks and
evaluation methods, baseline systems and their performance.
| new_dataset | 0.96741 |
2303.14194 | Ritam Majumdar | Ritam Majumdar, Shirish Karande, Lovekesh Vig | DeepEpiSolver: Unravelling Inverse problems in Covid, HIV, Ebola and
Disease Transmission | Publication accepted at International Conference for Learning
Representations 2023: First Workshop in Machine Learning and Global Health | null | null | null | cs.LG cs.AI q-bio.QM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The spread of many infectious diseases is modeled using variants of the SIR
compartmental model, which is a coupled differential equation. The coefficients
of the SIR model determine the spread trajectories of disease, on whose basis
proactive measures can be taken. Hence, the coefficient estimates must be both
fast and accurate. Shaier et al. in the paper "Disease Informed Neural
Networks" used Physics Informed Neural Networks (PINNs) to estimate the
parameters of the SIR model. There are two drawbacks to this approach. First,
the training time for PINNs is high, with certain diseases taking close to 90
hrs to train. Second, PINNs don't generalize for a new SIDR trajectory, and
learning its corresponding SIR parameters requires retraining the PINN from
scratch. In this work, we aim to eliminate both of these drawbacks. We generate
a dataset between the parameters of ODE and the spread trajectories by solving
the forward problem for a large distribution of parameters using the LSODA
algorithm. We then use a neural network to learn the mapping between spread
trajectories and coefficients of SIDR in an offline manner. This allows us to
learn the parameters of a new spread trajectory without having to retrain,
enabling generalization at test time. We observe a speed-up of 3-4 orders of
magnitude with accuracy comparable to that of PINNs for 11 highly infectious
diseases. Further finetuning of neural network inferred ODE coefficients using
PINN further leads to 2-3 orders improvement of estimated coefficients.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 11:47:16 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Majumdar",
"Ritam",
""
],
[
"Karande",
"Shirish",
""
],
[
"Vig",
"Lovekesh",
""
]
] | TITLE: DeepEpiSolver: Unravelling Inverse problems in Covid, HIV, Ebola and
Disease Transmission
ABSTRACT: The spread of many infectious diseases is modeled using variants of the SIR
compartmental model, which is a coupled differential equation. The coefficients
of the SIR model determine the spread trajectories of disease, on whose basis
proactive measures can be taken. Hence, the coefficient estimates must be both
fast and accurate. Shaier et al. in the paper "Disease Informed Neural
Networks" used Physics Informed Neural Networks (PINNs) to estimate the
parameters of the SIR model. There are two drawbacks to this approach. First,
the training time for PINNs is high, with certain diseases taking close to 90
hrs to train. Second, PINNs don't generalize for a new SIDR trajectory, and
learning its corresponding SIR parameters requires retraining the PINN from
scratch. In this work, we aim to eliminate both of these drawbacks. We generate
a dataset between the parameters of ODE and the spread trajectories by solving
the forward problem for a large distribution of parameters using the LSODA
algorithm. We then use a neural network to learn the mapping between spread
trajectories and coefficients of SIDR in an offline manner. This allows us to
learn the parameters of a new spread trajectory without having to retrain,
enabling generalization at test time. We observe a speed-up of 3-4 orders of
magnitude with accuracy comparable to that of PINNs for 11 highly infectious
diseases. Further finetuning of neural network inferred ODE coefficients using
PINN further leads to 2-3 orders improvement of estimated coefficients.
| no_new_dataset | 0.943608 |
2303.14197 | Wenqing Li | Yue Wang, Wending Li, Michail Maniatakos, Saif Eddin Jabari | Optimal Smoothing Distribution Exploration for Backdoor Neutralization
in Deep Learning-based Traffic Systems | null | null | null | null | cs.LG cs.AI cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep Reinforcement Learning (DRL) enhances the efficiency of Autonomous
Vehicles (AV), but also makes them susceptible to backdoor attacks that can
result in traffic congestion or collisions. Backdoor functionality is typically
incorporated by contaminating training datasets with covert malicious data to
maintain high precision on genuine inputs while inducing the desired
(malicious) outputs for specific inputs chosen by adversaries. Current defenses
against backdoors mainly focus on image classification using image-based
features, which cannot be readily transferred to the regression task of
DRL-based AV controllers since the inputs are continuous sensor data, i.e., the
combinations of velocity and distance of AV and its surrounding vehicles. Our
proposed method adds well-designed noise to the input to neutralize backdoors.
The approach involves learning an optimal smoothing (noise) distribution to
preserve the normal functionality of genuine inputs while neutralizing
backdoors. By doing so, the resulting model is expected to be more resilient
against backdoor attacks while maintaining high accuracy on genuine inputs. The
effectiveness of the proposed method is verified on a simulated traffic system
based on a microscopic traffic simulator, where experimental results showcase
that the smoothed traffic controller can neutralize all trigger samples and
maintain the performance of relieving traffic congestion
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 16:13:04 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Wang",
"Yue",
""
],
[
"Li",
"Wending",
""
],
[
"Maniatakos",
"Michail",
""
],
[
"Jabari",
"Saif Eddin",
""
]
] | TITLE: Optimal Smoothing Distribution Exploration for Backdoor Neutralization
in Deep Learning-based Traffic Systems
ABSTRACT: Deep Reinforcement Learning (DRL) enhances the efficiency of Autonomous
Vehicles (AV), but also makes them susceptible to backdoor attacks that can
result in traffic congestion or collisions. Backdoor functionality is typically
incorporated by contaminating training datasets with covert malicious data to
maintain high precision on genuine inputs while inducing the desired
(malicious) outputs for specific inputs chosen by adversaries. Current defenses
against backdoors mainly focus on image classification using image-based
features, which cannot be readily transferred to the regression task of
DRL-based AV controllers since the inputs are continuous sensor data, i.e., the
combinations of velocity and distance of AV and its surrounding vehicles. Our
proposed method adds well-designed noise to the input to neutralize backdoors.
The approach involves learning an optimal smoothing (noise) distribution to
preserve the normal functionality of genuine inputs while neutralizing
backdoors. By doing so, the resulting model is expected to be more resilient
against backdoor attacks while maintaining high accuracy on genuine inputs. The
effectiveness of the proposed method is verified on a simulated traffic system
based on a microscopic traffic simulator, where experimental results showcase
that the smoothed traffic controller can neutralize all trigger samples and
maintain the performance of relieving traffic congestion
| no_new_dataset | 0.944228 |
2303.14220 | Cl\'ement Chadebec | Cl\'ement Chadebec and St\'ephanie Allassonni\`ere | Variational Inference for Longitudinal Data Using Normalizing Flows | null | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a new latent variable generative model able to handle
high dimensional longitudinal data and relying on variational inference. The
time dependency between the observations of an input sequence is modelled using
normalizing flows over the associated latent variables. The proposed method can
be used to generate either fully synthetic longitudinal sequences or
trajectories that are conditioned on several data in a sequence and
demonstrates good robustness properties to missing data. We test the model on 6
datasets of different complexity and show that it can achieve better likelihood
estimates than some competitors as well as more reliable missing data
imputation. A code is made available at
\url{https://github.com/clementchadebec/variational_inference_for_longitudinal_data}.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 18:29:18 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Chadebec",
"Clément",
""
],
[
"Allassonnière",
"Stéphanie",
""
]
] | TITLE: Variational Inference for Longitudinal Data Using Normalizing Flows
ABSTRACT: This paper introduces a new latent variable generative model able to handle
high dimensional longitudinal data and relying on variational inference. The
time dependency between the observations of an input sequence is modelled using
normalizing flows over the associated latent variables. The proposed method can
be used to generate either fully synthetic longitudinal sequences or
trajectories that are conditioned on several data in a sequence and
demonstrates good robustness properties to missing data. We test the model on 6
datasets of different complexity and show that it can achieve better likelihood
estimates than some competitors as well as more reliable missing data
imputation. A code is made available at
\url{https://github.com/clementchadebec/variational_inference_for_longitudinal_data}.
| no_new_dataset | 0.948394 |
2303.14222 | Zoe Tieges Dr | Oliver Vinzelberg, Mark David Jenkins, Gordon Morison, David McMinn
and Zoe Tieges | Lay Text Summarisation Using Natural Language Processing: A Narrative
Literature Review | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Summarisation of research results in plain language is crucial for promoting
public understanding of research findings. The use of Natural Language
Processing to generate lay summaries has the potential to relieve researchers'
workload and bridge the gap between science and society. The aim of this
narrative literature review is to describe and compare the different text
summarisation approaches used to generate lay summaries. We searched the
databases Web of Science, Google Scholar, IEEE Xplore, Association for
Computing Machinery Digital Library and arXiv for articles published until 6
May 2022. We included original studies on automatic text summarisation methods
to generate lay summaries. We screened 82 articles and included eight relevant
papers published between 2020 and 2021, all using the same dataset. The results
show that transformer-based methods such as Bidirectional Encoder
Representations from Transformers (BERT) and Pre-training with Extracted
Gap-sentences for Abstractive Summarization (PEGASUS) dominate the landscape of
lay text summarisation, with all but one study using these methods. A
combination of extractive and abstractive summarisation methods in a hybrid
approach was found to be most effective. Furthermore, pre-processing approaches
to input text (e.g. applying extractive summarisation) or determining which
sections of a text to include, appear critical. Evaluation metrics such as
Recall-Oriented Understudy for Gisting Evaluation (ROUGE) were used, which do
not consider readability. To conclude, automatic lay text summarisation is
under-explored. Future research should consider long document lay text
summarisation, including clinical trial reports, and the development of
evaluation metrics that consider readability of the lay summary.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 18:30:50 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Vinzelberg",
"Oliver",
""
],
[
"Jenkins",
"Mark David",
""
],
[
"Morison",
"Gordon",
""
],
[
"McMinn",
"David",
""
],
[
"Tieges",
"Zoe",
""
]
] | TITLE: Lay Text Summarisation Using Natural Language Processing: A Narrative
Literature Review
ABSTRACT: Summarisation of research results in plain language is crucial for promoting
public understanding of research findings. The use of Natural Language
Processing to generate lay summaries has the potential to relieve researchers'
workload and bridge the gap between science and society. The aim of this
narrative literature review is to describe and compare the different text
summarisation approaches used to generate lay summaries. We searched the
databases Web of Science, Google Scholar, IEEE Xplore, Association for
Computing Machinery Digital Library and arXiv for articles published until 6
May 2022. We included original studies on automatic text summarisation methods
to generate lay summaries. We screened 82 articles and included eight relevant
papers published between 2020 and 2021, all using the same dataset. The results
show that transformer-based methods such as Bidirectional Encoder
Representations from Transformers (BERT) and Pre-training with Extracted
Gap-sentences for Abstractive Summarization (PEGASUS) dominate the landscape of
lay text summarisation, with all but one study using these methods. A
combination of extractive and abstractive summarisation methods in a hybrid
approach was found to be most effective. Furthermore, pre-processing approaches
to input text (e.g. applying extractive summarisation) or determining which
sections of a text to include, appear critical. Evaluation metrics such as
Recall-Oriented Understudy for Gisting Evaluation (ROUGE) were used, which do
not consider readability. To conclude, automatic lay text summarisation is
under-explored. Future research should consider long document lay text
summarisation, including clinical trial reports, and the development of
evaluation metrics that consider readability of the lay summary.
| no_new_dataset | 0.947817 |
2303.14243 | Heng Yu | Heng Yu, Joel Julin, Zoltan A. Milacski, Koichiro Niinuma, Laszlo A.
Jeni | DyLiN: Making Light Field Networks Dynamic | CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Light Field Networks, the re-formulations of radiance fields to oriented
rays, are magnitudes faster than their coordinate network counterparts, and
provide higher fidelity with respect to representing 3D structures from 2D
observations. They would be well suited for generic scene representation and
manipulation, but suffer from one problem: they are limited to holistic and
static scenes. In this paper, we propose the Dynamic Light Field Network
(DyLiN) method that can handle non-rigid deformations, including topological
changes. We learn a deformation field from input rays to canonical rays, and
lift them into a higher dimensional space to handle discontinuities. We further
introduce CoDyLiN, which augments DyLiN with controllable attribute inputs. We
train both models via knowledge distillation from pretrained dynamic radiance
fields. We evaluated DyLiN using both synthetic and real world datasets that
include various non-rigid deformations. DyLiN qualitatively outperformed and
quantitatively matched state-of-the-art methods in terms of visual fidelity,
while being 25 - 71x computationally faster. We also tested CoDyLiN on
attribute annotated data and it surpassed its teacher model. Project page:
https://dylin2023.github.io .
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 19:05:48 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Yu",
"Heng",
""
],
[
"Julin",
"Joel",
""
],
[
"Milacski",
"Zoltan A.",
""
],
[
"Niinuma",
"Koichiro",
""
],
[
"Jeni",
"Laszlo A.",
""
]
] | TITLE: DyLiN: Making Light Field Networks Dynamic
ABSTRACT: Light Field Networks, the re-formulations of radiance fields to oriented
rays, are magnitudes faster than their coordinate network counterparts, and
provide higher fidelity with respect to representing 3D structures from 2D
observations. They would be well suited for generic scene representation and
manipulation, but suffer from one problem: they are limited to holistic and
static scenes. In this paper, we propose the Dynamic Light Field Network
(DyLiN) method that can handle non-rigid deformations, including topological
changes. We learn a deformation field from input rays to canonical rays, and
lift them into a higher dimensional space to handle discontinuities. We further
introduce CoDyLiN, which augments DyLiN with controllable attribute inputs. We
train both models via knowledge distillation from pretrained dynamic radiance
fields. We evaluated DyLiN using both synthetic and real world datasets that
include various non-rigid deformations. DyLiN qualitatively outperformed and
quantitatively matched state-of-the-art methods in terms of visual fidelity,
while being 25 - 71x computationally faster. We also tested CoDyLiN on
attribute annotated data and it surpassed its teacher model. Project page:
https://dylin2023.github.io .
| no_new_dataset | 0.949482 |
2303.14247 | Bruno Arcanjo Mr | Bruno Arcanjo, Bruno Ferrarini, Michael Milford, Klaus D.
McDonald-Maier and Shoaib Ehsan | A-MuSIC: An Adaptive Ensemble System For Visual Place Recognition In
Changing Environments | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual place recognition (VPR) is an essential component of robot navigation
and localization systems that allows them to identify a place using only image
data. VPR is challenging due to the significant changes in a place's appearance
under different illumination throughout the day, with seasonal weather and when
observed from different viewpoints. Currently, no single VPR technique excels
in every environmental condition, each exhibiting unique benefits and
shortcomings. As a result, VPR systems combining multiple techniques achieve
more reliable VPR performance in changing environments, at the cost of higher
computational loads. Addressing this shortcoming, we propose an adaptive VPR
system dubbed Adaptive Multi-Self Identification and Correction (A-MuSIC). We
start by developing a method to collect information of the runtime performance
of a VPR technique by analysing the frame-to-frame continuity of matched
queries. We then demonstrate how to operate the method on a static ensemble of
techniques, generating data on which techniques are contributing the most for
the current environment. A-MuSIC uses the collected information to both select
a minimal subset of techniques and to decide when a re-selection is required
during navigation. A-MuSIC matches or beats state-of-the-art VPR performance
across all tested benchmark datasets while maintaining its computational load
on par with individual techniques.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 19:25:22 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Arcanjo",
"Bruno",
""
],
[
"Ferrarini",
"Bruno",
""
],
[
"Milford",
"Michael",
""
],
[
"McDonald-Maier",
"Klaus D.",
""
],
[
"Ehsan",
"Shoaib",
""
]
] | TITLE: A-MuSIC: An Adaptive Ensemble System For Visual Place Recognition In
Changing Environments
ABSTRACT: Visual place recognition (VPR) is an essential component of robot navigation
and localization systems that allows them to identify a place using only image
data. VPR is challenging due to the significant changes in a place's appearance
under different illumination throughout the day, with seasonal weather and when
observed from different viewpoints. Currently, no single VPR technique excels
in every environmental condition, each exhibiting unique benefits and
shortcomings. As a result, VPR systems combining multiple techniques achieve
more reliable VPR performance in changing environments, at the cost of higher
computational loads. Addressing this shortcoming, we propose an adaptive VPR
system dubbed Adaptive Multi-Self Identification and Correction (A-MuSIC). We
start by developing a method to collect information of the runtime performance
of a VPR technique by analysing the frame-to-frame continuity of matched
queries. We then demonstrate how to operate the method on a static ensemble of
techniques, generating data on which techniques are contributing the most for
the current environment. A-MuSIC uses the collected information to both select
a minimal subset of techniques and to decide when a re-selection is required
during navigation. A-MuSIC matches or beats state-of-the-art VPR performance
across all tested benchmark datasets while maintaining its computational load
on par with individual techniques.
| no_new_dataset | 0.945701 |
2303.14254 | Xiyuan Zhang | Xiyuan Zhang, Ranak Roy Chowdhury, Jingbo Shang, Rajesh Gupta, Dezhi
Hong | Towards Diverse and Coherent Augmentation for Time-Series Forecasting | ICASSP 2023 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time-series data augmentation mitigates the issue of insufficient training
data for deep learning models. Yet, existing augmentation methods are mainly
designed for classification, where class labels can be preserved even if
augmentation alters the temporal dynamics. We note that augmentation designed
for forecasting requires diversity as well as coherence with the original
temporal dynamics. As time-series data generated by real-life physical
processes exhibit characteristics in both the time and frequency domains, we
propose to combine Spectral and Time Augmentation (STAug) for generating more
diverse and coherent samples. Specifically, in the frequency domain, we use the
Empirical Mode Decomposition to decompose a time series and reassemble the
subcomponents with random weights. This way, we generate diverse samples while
being coherent with the original temporal relationships as they contain the
same set of base components. In the time domain, we adapt a mix-up strategy
that generates diverse as well as linearly in-between coherent samples.
Experiments on five real-world time-series datasets demonstrate that STAug
outperforms the base models without data augmentation as well as
state-of-the-art augmentation methods.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 19:40:34 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Zhang",
"Xiyuan",
""
],
[
"Chowdhury",
"Ranak Roy",
""
],
[
"Shang",
"Jingbo",
""
],
[
"Gupta",
"Rajesh",
""
],
[
"Hong",
"Dezhi",
""
]
] | TITLE: Towards Diverse and Coherent Augmentation for Time-Series Forecasting
ABSTRACT: Time-series data augmentation mitigates the issue of insufficient training
data for deep learning models. Yet, existing augmentation methods are mainly
designed for classification, where class labels can be preserved even if
augmentation alters the temporal dynamics. We note that augmentation designed
for forecasting requires diversity as well as coherence with the original
temporal dynamics. As time-series data generated by real-life physical
processes exhibit characteristics in both the time and frequency domains, we
propose to combine Spectral and Time Augmentation (STAug) for generating more
diverse and coherent samples. Specifically, in the frequency domain, we use the
Empirical Mode Decomposition to decompose a time series and reassemble the
subcomponents with random weights. This way, we generate diverse samples while
being coherent with the original temporal relationships as they contain the
same set of base components. In the time domain, we adapt a mix-up strategy
that generates diverse as well as linearly in-between coherent samples.
Experiments on five real-world time-series datasets demonstrate that STAug
outperforms the base models without data augmentation as well as
state-of-the-art augmentation methods.
| no_new_dataset | 0.948298 |
2303.14256 | Wilhelm Hasselbring | David Georg Reichelt, Stefan K\"uhne, Wilhelm Hasselbring | Automated Identification of Performance Changes at Code Level | 10 pages, 2022 IEEE 22nd International Conference on Software
Quality, Reliability and Security (QRS) | null | 10.1109/QRS57517.2022.00096 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To develop software with optimal performance, even small performance changes
need to be identified. Identifying performance changes is challenging since the
performance of software is influenced by non-deterministic factors. Therefore,
not every performance change is measurable with reasonable effort. In this
work, we discuss which performance changes are measurable at code level with
reasonable measurement effort and how to identify them. We present (1) an
analysis of the boundaries of measuring performance changes, (2) an approach
for determining a configuration for reproducible performance change
identification, and (3) an evaluation comparing of how well our approach is
able to identify performance changes in the application server Jetty compared
with the usage of Jetty's own performance regression benchmarks. Thereby, we
find (1) that small performance differences are only measurable by fine-grained
measurement workloads, (2) that performance changes caused by the change of one
operation can be identified using a unit-test-sized workload definition and a
suitable configuration, and (3) that using our approach identifies small
performance regressions more efficiently than using Jetty's performance
regression benchmarks.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 19:51:52 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Reichelt",
"David Georg",
""
],
[
"Kühne",
"Stefan",
""
],
[
"Hasselbring",
"Wilhelm",
""
]
] | TITLE: Automated Identification of Performance Changes at Code Level
ABSTRACT: To develop software with optimal performance, even small performance changes
need to be identified. Identifying performance changes is challenging since the
performance of software is influenced by non-deterministic factors. Therefore,
not every performance change is measurable with reasonable effort. In this
work, we discuss which performance changes are measurable at code level with
reasonable measurement effort and how to identify them. We present (1) an
analysis of the boundaries of measuring performance changes, (2) an approach
for determining a configuration for reproducible performance change
identification, and (3) an evaluation comparing of how well our approach is
able to identify performance changes in the application server Jetty compared
with the usage of Jetty's own performance regression benchmarks. Thereby, we
find (1) that small performance differences are only measurable by fine-grained
measurement workloads, (2) that performance changes caused by the change of one
operation can be identified using a unit-test-sized workload definition and a
suitable configuration, and (3) that using our approach identifies small
performance regressions more efficiently than using Jetty's performance
regression benchmarks.
| no_new_dataset | 0.946745 |
2303.14272 | Shiwali Mohan | Wiktor Piotrowski and Roni Stern and Yoni Sher and Jacob Le and
Matthew Klenk and Johan deKleer and Shiwali Mohan | Learning to Operate in Open Worlds by Adapting Planning Models | To appears in the Proceedings of the 22nd International Conference on
Autonomous Agents and Multiagent Systems (AAMAS 2023) | null | null | null | cs.AI cs.LG cs.SC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Planning agents are ill-equipped to act in novel situations in which their
domain model no longer accurately represents the world. We introduce an
approach for such agents operating in open worlds that detects the presence of
novelties and effectively adapts their domain models and consequent action
selection. It uses observations of action execution and measures their
divergence from what is expected, according to the environment model, to infer
existence of a novelty. Then, it revises the model through a heuristics-guided
search over model changes. We report empirical evaluations on the CartPole
problem, a standard Reinforcement Learning (RL) benchmark. The results show
that our approach can deal with a class of novelties very quickly and in an
interpretable fashion.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 21:04:16 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Piotrowski",
"Wiktor",
""
],
[
"Stern",
"Roni",
""
],
[
"Sher",
"Yoni",
""
],
[
"Le",
"Jacob",
""
],
[
"Klenk",
"Matthew",
""
],
[
"deKleer",
"Johan",
""
],
[
"Mohan",
"Shiwali",
""
]
] | TITLE: Learning to Operate in Open Worlds by Adapting Planning Models
ABSTRACT: Planning agents are ill-equipped to act in novel situations in which their
domain model no longer accurately represents the world. We introduce an
approach for such agents operating in open worlds that detects the presence of
novelties and effectively adapts their domain models and consequent action
selection. It uses observations of action execution and measures their
divergence from what is expected, according to the environment model, to infer
existence of a novelty. Then, it revises the model through a heuristics-guided
search over model changes. We report empirical evaluations on the CartPole
problem, a standard Reinforcement Learning (RL) benchmark. The results show
that our approach can deal with a class of novelties very quickly and in an
interpretable fashion.
| no_new_dataset | 0.946547 |
2303.14279 | Ilias Triantafyllopoulos | Ilias Triantafyllopoulos, Georgios Paraskevopoulos, Alexandros
Potamianos | Depression detection in social media posts using affective and social
norm features | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | We propose a deep architecture for depression detection from social media
posts. The proposed architecture builds upon BERT to extract language
representations from social media posts and combines these representations
using an attentive bidirectional GRU network. We incorporate affective
information, by augmenting the text representations with features extracted
from a pretrained emotion classifier. Motivated by psychological literature we
propose to incorporate profanity and morality features of posts and words in
our architecture using a late fusion scheme. Our analysis indicates that
morality and profanity can be important features for depression detection. We
apply our model for depression detection on Reddit posts on the Pirina dataset,
and further consider the setting of detecting depressed users, given multiple
posts per user, proposed in the Reddit RSDD dataset. The inclusion of the
proposed features yields state-of-the-art results in both settings, namely
2.65% and 6.73% absolute improvement in F1 score respectively. Index Terms:
Depression detection, BERT, Feature fusion, Emotion recognition, profanity,
morality
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 21:26:27 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Triantafyllopoulos",
"Ilias",
""
],
[
"Paraskevopoulos",
"Georgios",
""
],
[
"Potamianos",
"Alexandros",
""
]
] | TITLE: Depression detection in social media posts using affective and social
norm features
ABSTRACT: We propose a deep architecture for depression detection from social media
posts. The proposed architecture builds upon BERT to extract language
representations from social media posts and combines these representations
using an attentive bidirectional GRU network. We incorporate affective
information, by augmenting the text representations with features extracted
from a pretrained emotion classifier. Motivated by psychological literature we
propose to incorporate profanity and morality features of posts and words in
our architecture using a late fusion scheme. Our analysis indicates that
morality and profanity can be important features for depression detection. We
apply our model for depression detection on Reddit posts on the Pirina dataset,
and further consider the setting of detecting depressed users, given multiple
posts per user, proposed in the Reddit RSDD dataset. The inclusion of the
proposed features yields state-of-the-art results in both settings, namely
2.65% and 6.73% absolute improvement in F1 score respectively. Index Terms:
Depression detection, BERT, Feature fusion, Emotion recognition, profanity,
morality
| no_new_dataset | 0.945801 |
2303.14291 | Ryan-Rhys Griffiths | Ryan-Rhys Griffiths | Applications of Gaussian Processes at Extreme Lengthscales: From
Molecules to Black Holes | PhD Thesis submitted at the University of Cambridge, August 2022. The
thesis is based on a number of previous works also available on arXiv (see
Introduction) | null | 10.17863/CAM.93643 | null | stat.ML astro-ph.HE astro-ph.IM cond-mat.mtrl-sci cs.LG physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | In many areas of the observational and experimental sciences data is scarce.
Data observation in high-energy astrophysics is disrupted by celestial
occlusions and limited telescope time while data derived from laboratory
experiments in synthetic chemistry and materials science is time and
cost-intensive to collect. On the other hand, knowledge about the
data-generation mechanism is often available in the sciences, such as the
measurement error of a piece of laboratory apparatus. Both characteristics,
small data and knowledge of the underlying physics, make Gaussian processes
(GPs) ideal candidates for fitting such datasets. GPs can make predictions with
consideration of uncertainty, for example in the virtual screening of molecules
and materials, and can also make inferences about incomplete data such as the
latent emission signature from a black hole accretion disc. Furthermore, GPs
are currently the workhorse model for Bayesian optimisation, a methodology
foreseen to be a guide for laboratory experiments in scientific discovery
campaigns. The first contribution of this thesis is to use GP modelling to
reason about the latent emission signature from the Seyfert galaxy Markarian
335, and by extension, to reason about the applicability of various theoretical
models of black hole accretion discs. The second contribution is to extend the
GP framework to molecular and chemical reaction representations and to provide
an open-source software library to enable the framework to be used by
scientists. The third contribution is to leverage GPs to discover novel and
performant photoswitch molecules. The fourth contribution is to introduce a
Bayesian optimisation scheme capable of modelling aleatoric uncertainty to
facilitate the identification of material compositions that possess intrinsic
robustness to large scale fabrication processes.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 22:20:14 GMT"
}
] | 2023-03-28T00:00:00 | [
[
"Griffiths",
"Ryan-Rhys",
""
]
] | TITLE: Applications of Gaussian Processes at Extreme Lengthscales: From
Molecules to Black Holes
ABSTRACT: In many areas of the observational and experimental sciences data is scarce.
Data observation in high-energy astrophysics is disrupted by celestial
occlusions and limited telescope time while data derived from laboratory
experiments in synthetic chemistry and materials science is time and
cost-intensive to collect. On the other hand, knowledge about the
data-generation mechanism is often available in the sciences, such as the
measurement error of a piece of laboratory apparatus. Both characteristics,
small data and knowledge of the underlying physics, make Gaussian processes
(GPs) ideal candidates for fitting such datasets. GPs can make predictions with
consideration of uncertainty, for example in the virtual screening of molecules
and materials, and can also make inferences about incomplete data such as the
latent emission signature from a black hole accretion disc. Furthermore, GPs
are currently the workhorse model for Bayesian optimisation, a methodology
foreseen to be a guide for laboratory experiments in scientific discovery
campaigns. The first contribution of this thesis is to use GP modelling to
reason about the latent emission signature from the Seyfert galaxy Markarian
335, and by extension, to reason about the applicability of various theoretical
models of black hole accretion discs. The second contribution is to extend the
GP framework to molecular and chemical reaction representations and to provide
an open-source software library to enable the framework to be used by
scientists. The third contribution is to leverage GPs to discover novel and
performant photoswitch molecules. The fourth contribution is to introduce a
Bayesian optimisation scheme capable of modelling aleatoric uncertainty to
facilitate the identification of material compositions that possess intrinsic
robustness to large scale fabrication processes.
| no_new_dataset | 0.949902 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.