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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2303.10898 | Min Zhang | Min Zhang, Jintang Xue, Pranav Kadam, Hardik Prajapati, Shan Liu,
C.-C. Jay Kuo | A Tiny Machine Learning Model for Point Cloud Object Classification | 13 pages, 4 figures | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The design of a tiny machine learning model, which can be deployed in mobile
and edge devices, for point cloud object classification is investigated in this
work. To achieve this objective, we replace the multi-scale representation of a
point cloud object with a single-scale representation for complexity reduction,
and exploit rich 3D geometric information of a point cloud object for
performance improvement. The proposed solution is named Green-PointHop due to
its low computational complexity. We evaluate the performance of Green-PointHop
on ModelNet40 and ScanObjectNN two datasets. Green-PointHop has a model size of
64K parameters. It demands 2.3M floating-point operations (FLOPs) to classify a
ModelNet40 object of 1024 down-sampled points. Its classification performance
gaps against the state-of-the-art DGCNN method are 3% and 7% for ModelNet40 and
ScanObjectNN, respectively. On the other hand, the model size and inference
complexity of DGCNN are 42X and 1203X of those of Green-PointHop, respectively.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 06:35:46 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Zhang",
"Min",
""
],
[
"Xue",
"Jintang",
""
],
[
"Kadam",
"Pranav",
""
],
[
"Prajapati",
"Hardik",
""
],
[
"Liu",
"Shan",
""
],
[
"Kuo",
"C. -C. Jay",
""
]
] | TITLE: A Tiny Machine Learning Model for Point Cloud Object Classification
ABSTRACT: The design of a tiny machine learning model, which can be deployed in mobile
and edge devices, for point cloud object classification is investigated in this
work. To achieve this objective, we replace the multi-scale representation of a
point cloud object with a single-scale representation for complexity reduction,
and exploit rich 3D geometric information of a point cloud object for
performance improvement. The proposed solution is named Green-PointHop due to
its low computational complexity. We evaluate the performance of Green-PointHop
on ModelNet40 and ScanObjectNN two datasets. Green-PointHop has a model size of
64K parameters. It demands 2.3M floating-point operations (FLOPs) to classify a
ModelNet40 object of 1024 down-sampled points. Its classification performance
gaps against the state-of-the-art DGCNN method are 3% and 7% for ModelNet40 and
ScanObjectNN, respectively. On the other hand, the model size and inference
complexity of DGCNN are 42X and 1203X of those of Green-PointHop, respectively.
| no_new_dataset | 0.948442 |
2303.10901 | Mohsen Amini Salehi | Ali Mokhtari, Drake Rawls, Tony Huynh, Jeremiah Green, Mohsen Amini
Salehi | E2C: A Visual Simulator to Reinforce Education of Heterogeneous
Computing Systems | Accepted in Edupar '23, as part of IPDPS '23 Conference. arXiv admin
note: text overlap with arXiv:2212.11333 | null | null | null | cs.DC cs.AR cs.OS | http://creativecommons.org/publicdomain/zero/1.0/ | With the increasing popularity of accelerator technologies (e.g., GPUs and
TPUs) and the emergence of domain-specific computing via ASICs and FPGA, the
matter of heterogeneity and understanding its ramifications on the performance
has become more critical than ever before. However, it is challenging to
effectively educate students about the potential impacts of heterogeneity on
the performance of distributed systems; and on the logic of resource allocation
methods to efficiently utilize the resources. Making use of the real
infrastructure for benchmarking the performance of heterogeneous machines, for
different applications, with respect to different objectives, and under various
workload intensities is cost- and time-prohibitive. To reinforce the quality of
learning about various dimensions of heterogeneity, and to decrease the
widening gap in education, we develop an open-source simulation tool, called
E2C, that can help students researchers to study any type of heterogeneous (or
homogeneous) computing system and measure its performance under various
configurations. E2C is equipped with an intuitive graphical user interface
(GUI) that enables its users to easily examine system-level solutions
(scheduling, load balancing, scalability, etc.) in a controlled environment
within a short time. E2C is a discrete event simulator that offers the
following features: (i) simulating a heterogeneous computing system; (ii)
implementing a newly developed scheduling method and plugging it into the
system, (iii) measuring energy consumption and other output-related metrics;
and (iv) powerful visual aspects to ease the learning curve for students. We
used E2C as an assignment in the Distributed and Cloud Computing course. Our
anonymous survey study indicates that students rated E2C with the score of 8.7
out of 10 for its usefulness in understanding the concepts of scheduling in
heterogeneous computing.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 06:38:10 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Mokhtari",
"Ali",
""
],
[
"Rawls",
"Drake",
""
],
[
"Huynh",
"Tony",
""
],
[
"Green",
"Jeremiah",
""
],
[
"Salehi",
"Mohsen Amini",
""
]
] | TITLE: E2C: A Visual Simulator to Reinforce Education of Heterogeneous
Computing Systems
ABSTRACT: With the increasing popularity of accelerator technologies (e.g., GPUs and
TPUs) and the emergence of domain-specific computing via ASICs and FPGA, the
matter of heterogeneity and understanding its ramifications on the performance
has become more critical than ever before. However, it is challenging to
effectively educate students about the potential impacts of heterogeneity on
the performance of distributed systems; and on the logic of resource allocation
methods to efficiently utilize the resources. Making use of the real
infrastructure for benchmarking the performance of heterogeneous machines, for
different applications, with respect to different objectives, and under various
workload intensities is cost- and time-prohibitive. To reinforce the quality of
learning about various dimensions of heterogeneity, and to decrease the
widening gap in education, we develop an open-source simulation tool, called
E2C, that can help students researchers to study any type of heterogeneous (or
homogeneous) computing system and measure its performance under various
configurations. E2C is equipped with an intuitive graphical user interface
(GUI) that enables its users to easily examine system-level solutions
(scheduling, load balancing, scalability, etc.) in a controlled environment
within a short time. E2C is a discrete event simulator that offers the
following features: (i) simulating a heterogeneous computing system; (ii)
implementing a newly developed scheduling method and plugging it into the
system, (iii) measuring energy consumption and other output-related metrics;
and (iv) powerful visual aspects to ease the learning curve for students. We
used E2C as an assignment in the Distributed and Cloud Computing course. Our
anonymous survey study indicates that students rated E2C with the score of 8.7
out of 10 for its usefulness in understanding the concepts of scheduling in
heterogeneous computing.
| no_new_dataset | 0.944485 |
2303.10912 | Fan Cui | Fan Cui, Liyong Guo, Quandong Wang, Peng Gao, Yujun Wang | Exploring Representation Learning for Small-Footprint Keyword Spotting | null | null | 10.21437/Interspeech.2022-10558 | null | cs.SD cs.CL cs.LG eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we investigate representation learning for low-resource
keyword spotting (KWS). The main challenges of KWS are limited labeled data and
limited available device resources. To address those challenges, we explore
representation learning for KWS by self-supervised contrastive learning and
self-training with pretrained model. First, local-global contrastive siamese
networks (LGCSiam) are designed to learn similar utterance-level
representations for similar audio samplers by proposed local-global contrastive
loss without requiring ground-truth. Second, a self-supervised pretrained
Wav2Vec 2.0 model is applied as a constraint module (WVC) to force the KWS
model to learn frame-level acoustic representations. By the LGCSiam and WVC
modules, the proposed small-footprint KWS model can be pretrained with
unlabeled data. Experiments on speech commands dataset show that the
self-training WVC module and the self-supervised LGCSiam module significantly
improve accuracy, especially in the case of training on a small labeled
dataset.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 07:09:26 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Cui",
"Fan",
""
],
[
"Guo",
"Liyong",
""
],
[
"Wang",
"Quandong",
""
],
[
"Gao",
"Peng",
""
],
[
"Wang",
"Yujun",
""
]
] | TITLE: Exploring Representation Learning for Small-Footprint Keyword Spotting
ABSTRACT: In this paper, we investigate representation learning for low-resource
keyword spotting (KWS). The main challenges of KWS are limited labeled data and
limited available device resources. To address those challenges, we explore
representation learning for KWS by self-supervised contrastive learning and
self-training with pretrained model. First, local-global contrastive siamese
networks (LGCSiam) are designed to learn similar utterance-level
representations for similar audio samplers by proposed local-global contrastive
loss without requiring ground-truth. Second, a self-supervised pretrained
Wav2Vec 2.0 model is applied as a constraint module (WVC) to force the KWS
model to learn frame-level acoustic representations. By the LGCSiam and WVC
modules, the proposed small-footprint KWS model can be pretrained with
unlabeled data. Experiments on speech commands dataset show that the
self-training WVC module and the self-supervised LGCSiam module significantly
improve accuracy, especially in the case of training on a small labeled
dataset.
| no_new_dataset | 0.948585 |
2303.10913 | Fanhai Zeng | Lei Ma, Rong xin Li, Fanhai Zeng, Ling Guo, and George Em Karniadakis | Bi-orthogonal fPINN: A physics-informed neural network method for
solving time-dependent stochastic fractional PDEs | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fractional partial differential equations (FPDEs) can effectively represent
anomalous transport and nonlocal interactions. However, inherent uncertainties
arise naturally in real applications due to random forcing or unknown material
properties. Mathematical models considering nonlocal interactions with
uncertainty quantification can be formulated as stochastic fractional partial
differential equations (SFPDEs). There are many challenges in solving SFPDEs
numerically, especially for long-time integration since such problems are
high-dimensional and nonlocal. Here, we combine the bi-orthogonal (BO) method
for representing stochastic processes with physics-informed neural networks
(PINNs) for solving partial differential equations to formulate the
bi-orthogonal PINN method (BO-fPINN) for solving time-dependent SFPDEs.
Specifically, we introduce a deep neural network for the stochastic solution of
the time-dependent SFPDEs, and include the BO constraints in the loss function
following a weak formulation. Since automatic differentiation is not currently
applicable to fractional derivatives, we employ discretization on a grid to to
compute the fractional derivatives of the neural network output. The weak
formulation loss function of the BO-fPINN method can overcome some drawbacks of
the BO methods and thus can be used to solve SFPDEs with eigenvalue crossings.
Moreover, the BO-fPINN method can be used for inverse SFPDEs with the same
framework and same computational complexity as for forward problems. We
demonstrate the effectiveness of the BO-fPINN method for different benchmark
problems. The results demonstrate the flexibility and efficiency of the
proposed method, especially for inverse problems.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 07:09:51 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Ma",
"Lei",
""
],
[
"Li",
"Rong xin",
""
],
[
"Zeng",
"Fanhai",
""
],
[
"Guo",
"Ling",
""
],
[
"Karniadakis",
"George Em",
""
]
] | TITLE: Bi-orthogonal fPINN: A physics-informed neural network method for
solving time-dependent stochastic fractional PDEs
ABSTRACT: Fractional partial differential equations (FPDEs) can effectively represent
anomalous transport and nonlocal interactions. However, inherent uncertainties
arise naturally in real applications due to random forcing or unknown material
properties. Mathematical models considering nonlocal interactions with
uncertainty quantification can be formulated as stochastic fractional partial
differential equations (SFPDEs). There are many challenges in solving SFPDEs
numerically, especially for long-time integration since such problems are
high-dimensional and nonlocal. Here, we combine the bi-orthogonal (BO) method
for representing stochastic processes with physics-informed neural networks
(PINNs) for solving partial differential equations to formulate the
bi-orthogonal PINN method (BO-fPINN) for solving time-dependent SFPDEs.
Specifically, we introduce a deep neural network for the stochastic solution of
the time-dependent SFPDEs, and include the BO constraints in the loss function
following a weak formulation. Since automatic differentiation is not currently
applicable to fractional derivatives, we employ discretization on a grid to to
compute the fractional derivatives of the neural network output. The weak
formulation loss function of the BO-fPINN method can overcome some drawbacks of
the BO methods and thus can be used to solve SFPDEs with eigenvalue crossings.
Moreover, the BO-fPINN method can be used for inverse SFPDEs with the same
framework and same computational complexity as for forward problems. We
demonstrate the effectiveness of the BO-fPINN method for different benchmark
problems. The results demonstrate the flexibility and efficiency of the
proposed method, especially for inverse problems.
| no_new_dataset | 0.951594 |
2303.10930 | Andreas Wiedholz | Andreas Wiedholz, Stefanie Wucherer, Simon Dietrich | Semantic 3D scene segmentation for robotic assembly process execution | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Adapting robot programmes to changes in the environment is a well-known
industry problem, and it is the reason why many tedious tasks are not automated
in small and medium-sized enterprises (SMEs). A semantic world model of a
robot's previously unknown environment created from point clouds is one way for
these companies to automate assembly tasks that are typically performed by
humans. The semantic segmentation of point clouds for robot manipulators or
cobots in industrial environments has received little attention due to a lack
of suitable datasets. This paper describes a pipeline for creating synthetic
point clouds for specific use cases in order to train a model for point cloud
semantic segmentation. We show that models trained with our data achieve high
per-class accuracy (> 90%) for semantic point cloud segmentation on unseen
real-world data. Our approach is applicable not only to the 3D camera used in
training data generation but also to other depth cameras based on different
technologies. The application tested in this work is a industry-related
peg-in-the-hole process. With our approach the necessity of user assistance
during a robot's commissioning can be reduced to a minimum.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 08:04:52 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wiedholz",
"Andreas",
""
],
[
"Wucherer",
"Stefanie",
""
],
[
"Dietrich",
"Simon",
""
]
] | TITLE: Semantic 3D scene segmentation for robotic assembly process execution
ABSTRACT: Adapting robot programmes to changes in the environment is a well-known
industry problem, and it is the reason why many tedious tasks are not automated
in small and medium-sized enterprises (SMEs). A semantic world model of a
robot's previously unknown environment created from point clouds is one way for
these companies to automate assembly tasks that are typically performed by
humans. The semantic segmentation of point clouds for robot manipulators or
cobots in industrial environments has received little attention due to a lack
of suitable datasets. This paper describes a pipeline for creating synthetic
point clouds for specific use cases in order to train a model for point cloud
semantic segmentation. We show that models trained with our data achieve high
per-class accuracy (> 90%) for semantic point cloud segmentation on unseen
real-world data. Our approach is applicable not only to the 3D camera used in
training data generation but also to other depth cameras based on different
technologies. The application tested in this work is a industry-related
peg-in-the-hole process. With our approach the necessity of user assistance
during a robot's commissioning can be reduced to a minimum.
| no_new_dataset | 0.951414 |
2303.10941 | Haoyu Wang | Haoyu Wang, Shaoli Huang, Fang Zhao, Chun Yuan, Ying Shan | HMC: Hierarchical Mesh Coarsening for Skeleton-free Motion Retargeting | null | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a simple yet effective method for skeleton-free motion
retargeting. Previous methods transfer motion between high-resolution meshes,
failing to preserve the inherent local-part motions in the mesh. Addressing
this issue, our proposed method learns the correspondence in a coarse-to-fine
fashion by integrating the retargeting process with a mesh-coarsening pipeline.
First, we propose a mesh-coarsening module that coarsens the mesh
representations for better motion transfer. This module improves the ability to
handle small-part motion and preserves the local motion interdependence between
neighboring mesh vertices. Furthermore, we leverage a hierarchical refinement
procedure to complement missing mesh details by gradually improving the
low-resolution mesh output with a higher-resolution one. We evaluate our method
on several well-known 3D character datasets, and it yields an average
improvement of 25% on point-wise mesh euclidean distance (PMD) against the
start-of-art method. Moreover, our qualitative results show that our method is
significantly helpful in preserving the moving consistency of different body
parts on the target character due to disentangling body-part structures and
mesh details in a hierarchical way.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 08:52:34 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wang",
"Haoyu",
""
],
[
"Huang",
"Shaoli",
""
],
[
"Zhao",
"Fang",
""
],
[
"Yuan",
"Chun",
""
],
[
"Shan",
"Ying",
""
]
] | TITLE: HMC: Hierarchical Mesh Coarsening for Skeleton-free Motion Retargeting
ABSTRACT: We present a simple yet effective method for skeleton-free motion
retargeting. Previous methods transfer motion between high-resolution meshes,
failing to preserve the inherent local-part motions in the mesh. Addressing
this issue, our proposed method learns the correspondence in a coarse-to-fine
fashion by integrating the retargeting process with a mesh-coarsening pipeline.
First, we propose a mesh-coarsening module that coarsens the mesh
representations for better motion transfer. This module improves the ability to
handle small-part motion and preserves the local motion interdependence between
neighboring mesh vertices. Furthermore, we leverage a hierarchical refinement
procedure to complement missing mesh details by gradually improving the
low-resolution mesh output with a higher-resolution one. We evaluate our method
on several well-known 3D character datasets, and it yields an average
improvement of 25% on point-wise mesh euclidean distance (PMD) against the
start-of-art method. Moreover, our qualitative results show that our method is
significantly helpful in preserving the moving consistency of different body
parts on the target character due to disentangling body-part structures and
mesh details in a hierarchical way.
| no_new_dataset | 0.95018 |
2303.10942 | Bolaji Yusuf | Bolaji Yusuf, Aditya Gourav, Ankur Gandhe, Ivan Bulyko | On-the-fly Text Retrieval for End-to-End ASR Adaptation | Accepted to ICASSP 2023; Appendix added to include ablations that
could not fit into the conference 4-page limit | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | End-to-end speech recognition models are improved by incorporating external
text sources, typically by fusion with an external language model. Such
language models have to be retrained whenever the corpus of interest changes.
Furthermore, since they store the entire corpus in their parameters, rare words
can be challenging to recall. In this work, we propose augmenting a
transducer-based ASR model with a retrieval language model, which directly
retrieves from an external text corpus plausible completions for a partial ASR
hypothesis. These completions are then integrated into subsequent predictions
by an adapter, which is trained once, so that the corpus of interest can be
switched without incurring the computational overhead of retraining. Our
experiments show that the proposed model significantly improves the performance
of a transducer baseline on a pair of question-answering datasets. Further, it
outperforms shallow fusion on recognition of named entities by about 7
relative; when the two are combined, the relative improvement increases to 13%.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 08:54:40 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Yusuf",
"Bolaji",
""
],
[
"Gourav",
"Aditya",
""
],
[
"Gandhe",
"Ankur",
""
],
[
"Bulyko",
"Ivan",
""
]
] | TITLE: On-the-fly Text Retrieval for End-to-End ASR Adaptation
ABSTRACT: End-to-end speech recognition models are improved by incorporating external
text sources, typically by fusion with an external language model. Such
language models have to be retrained whenever the corpus of interest changes.
Furthermore, since they store the entire corpus in their parameters, rare words
can be challenging to recall. In this work, we propose augmenting a
transducer-based ASR model with a retrieval language model, which directly
retrieves from an external text corpus plausible completions for a partial ASR
hypothesis. These completions are then integrated into subsequent predictions
by an adapter, which is trained once, so that the corpus of interest can be
switched without incurring the computational overhead of retraining. Our
experiments show that the proposed model significantly improves the performance
of a transducer baseline on a pair of question-answering datasets. Further, it
outperforms shallow fusion on recognition of named entities by about 7
relative; when the two are combined, the relative improvement increases to 13%.
| no_new_dataset | 0.94868 |
2303.10945 | Junyang Chen | Junyang Chen, Xiaoyu Xian, Zhijing Yang, Tianshui Chen, Yongyi Lu,
Yukai Shi, Jinshan Pan, Liang Lin | Open-World Pose Transfer via Sequential Test-Time Adaption | We call for a solid pose transfer model that can handle open-world
instances beyond a specific dataset | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Pose transfer aims to transfer a given person into a specified posture, has
recently attracted considerable attention. A typical pose transfer framework
usually employs representative datasets to train a discriminative model, which
is often violated by out-of-distribution (OOD) instances. Recently, test-time
adaption (TTA) offers a feasible solution for OOD data by using a pre-trained
model that learns essential features with self-supervision. However, those
methods implicitly make an assumption that all test distributions have a
unified signal that can be learned directly. In open-world conditions, the pose
transfer task raises various independent signals: OOD appearance and skeleton,
which need to be extracted and distributed in speciality. To address this
point, we develop a SEquential Test-time Adaption (SETA). In the test-time
phrase, SETA extracts and distributes external appearance texture by augmenting
OOD data for self-supervised training. To make non-Euclidean similarity among
different postures explicit, SETA uses the image representations derived from a
person re-identification (Re-ID) model for similarity computation. By
addressing implicit posture representation in the test-time sequentially, SETA
greatly improves the generalization performance of current pose transfer
models. In our experiment, we first show that pose transfer can be applied to
open-world applications, including Tiktok reenactment and celebrity motion
synthesis.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:01:23 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Chen",
"Junyang",
""
],
[
"Xian",
"Xiaoyu",
""
],
[
"Yang",
"Zhijing",
""
],
[
"Chen",
"Tianshui",
""
],
[
"Lu",
"Yongyi",
""
],
[
"Shi",
"Yukai",
""
],
[
"Pan",
"Jinshan",
""
],
[
"Lin",
"Liang",
""
]
] | TITLE: Open-World Pose Transfer via Sequential Test-Time Adaption
ABSTRACT: Pose transfer aims to transfer a given person into a specified posture, has
recently attracted considerable attention. A typical pose transfer framework
usually employs representative datasets to train a discriminative model, which
is often violated by out-of-distribution (OOD) instances. Recently, test-time
adaption (TTA) offers a feasible solution for OOD data by using a pre-trained
model that learns essential features with self-supervision. However, those
methods implicitly make an assumption that all test distributions have a
unified signal that can be learned directly. In open-world conditions, the pose
transfer task raises various independent signals: OOD appearance and skeleton,
which need to be extracted and distributed in speciality. To address this
point, we develop a SEquential Test-time Adaption (SETA). In the test-time
phrase, SETA extracts and distributes external appearance texture by augmenting
OOD data for self-supervised training. To make non-Euclidean similarity among
different postures explicit, SETA uses the image representations derived from a
person re-identification (Re-ID) model for similarity computation. By
addressing implicit posture representation in the test-time sequentially, SETA
greatly improves the generalization performance of current pose transfer
models. In our experiment, we first show that pose transfer can be applied to
open-world applications, including Tiktok reenactment and celebrity motion
synthesis.
| no_new_dataset | 0.947186 |
2303.10949 | Haibin Yu | Haibin Yu, Yuxuan Hu, Yao Qian, Ma Jin, Linquan Liu, Shujie Liu, Yu
Shi, Yanmin Qian, Edward Lin, Michael Zeng | Code-Switching Text Generation and Injection in Mandarin-English ASR | Accepted by ICASSP 2023 | null | null | null | eess.AS cs.CL cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Code-switching speech refers to a means of expression by mixing two or more
languages within a single utterance. Automatic Speech Recognition (ASR) with
End-to-End (E2E) modeling for such speech can be a challenging task due to the
lack of data. In this study, we investigate text generation and injection for
improving the performance of an industry commonly-used streaming model,
Transformer-Transducer (T-T), in Mandarin-English code-switching speech
recognition. We first propose a strategy to generate code-switching text data
and then investigate injecting generated text into T-T model explicitly by
Text-To-Speech (TTS) conversion or implicitly by tying speech and text latent
spaces. Experimental results on the T-T model trained with a dataset containing
1,800 hours of real Mandarin-English code-switched speech show that our
approaches to inject generated code-switching text significantly boost the
performance of T-T models, i.e., 16% relative Token-based Error Rate (TER)
reduction averaged on three evaluation sets, and the approach of tying speech
and text latent spaces is superior to that of TTS conversion on the evaluation
set which contains more homogeneous data with the training set.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:13:27 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Yu",
"Haibin",
""
],
[
"Hu",
"Yuxuan",
""
],
[
"Qian",
"Yao",
""
],
[
"Jin",
"Ma",
""
],
[
"Liu",
"Linquan",
""
],
[
"Liu",
"Shujie",
""
],
[
"Shi",
"Yu",
""
],
[
"Qian",
"Yanmin",
""
],
[
"Lin",
"Edward",
""
],
[
"Zeng",
"Michael",
""
]
] | TITLE: Code-Switching Text Generation and Injection in Mandarin-English ASR
ABSTRACT: Code-switching speech refers to a means of expression by mixing two or more
languages within a single utterance. Automatic Speech Recognition (ASR) with
End-to-End (E2E) modeling for such speech can be a challenging task due to the
lack of data. In this study, we investigate text generation and injection for
improving the performance of an industry commonly-used streaming model,
Transformer-Transducer (T-T), in Mandarin-English code-switching speech
recognition. We first propose a strategy to generate code-switching text data
and then investigate injecting generated text into T-T model explicitly by
Text-To-Speech (TTS) conversion or implicitly by tying speech and text latent
spaces. Experimental results on the T-T model trained with a dataset containing
1,800 hours of real Mandarin-English code-switched speech show that our
approaches to inject generated code-switching text significantly boost the
performance of T-T models, i.e., 16% relative Token-based Error Rate (TER)
reduction averaged on three evaluation sets, and the approach of tying speech
and text latent spaces is superior to that of TTS conversion on the evaluation
set which contains more homogeneous data with the training set.
| new_dataset | 0.959724 |
2303.10951 | Junjie Ye | Junjie Ye, Changhong Fu, Ziang Cao, Shan An, Guangze Zheng, Bowen Li | Tracker Meets Night: A Transformer Enhancer for UAV Tracking | null | IEEE Robotics and Automation Letters, 2022, vol. 7 No. 2 | 10.1109/LRA.2022.3146911 | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most previous progress in object tracking is realized in daytime scenes with
favorable illumination. State-of-the-arts can hardly carry on their superiority
at night so far, thereby considerably blocking the broadening of visual
tracking-related unmanned aerial vehicle (UAV) applications. To realize
reliable UAV tracking at night, a spatial-channel Transformer-based low-light
enhancer (namely SCT), which is trained in a novel task-inspired manner, is
proposed and plugged prior to tracking approaches. To achieve semantic-level
low-light enhancement targeting the high-level task, the novel spatial-channel
attention module is proposed to model global information while preserving local
context. In the enhancement process, SCT denoises and illuminates nighttime
images simultaneously through a robust non-linear curve projection. Moreover,
to provide a comprehensive evaluation, we construct a challenging nighttime
tracking benchmark, namely DarkTrack2021, which contains 110 challenging
sequences with over 100 K frames in total. Evaluations on both the public
UAVDark135 benchmark and the newly constructed DarkTrack2021 benchmark show
that the task-inspired design enables SCT with significant performance gains
for nighttime UAV tracking compared with other top-ranked low-light enhancers.
Real-world tests on a typical UAV platform further verify the practicability of
the proposed approach. The DarkTrack2021 benchmark and the code of the proposed
approach are publicly available at https://github.com/vision4robotics/SCT.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:18:52 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Ye",
"Junjie",
""
],
[
"Fu",
"Changhong",
""
],
[
"Cao",
"Ziang",
""
],
[
"An",
"Shan",
""
],
[
"Zheng",
"Guangze",
""
],
[
"Li",
"Bowen",
""
]
] | TITLE: Tracker Meets Night: A Transformer Enhancer for UAV Tracking
ABSTRACT: Most previous progress in object tracking is realized in daytime scenes with
favorable illumination. State-of-the-arts can hardly carry on their superiority
at night so far, thereby considerably blocking the broadening of visual
tracking-related unmanned aerial vehicle (UAV) applications. To realize
reliable UAV tracking at night, a spatial-channel Transformer-based low-light
enhancer (namely SCT), which is trained in a novel task-inspired manner, is
proposed and plugged prior to tracking approaches. To achieve semantic-level
low-light enhancement targeting the high-level task, the novel spatial-channel
attention module is proposed to model global information while preserving local
context. In the enhancement process, SCT denoises and illuminates nighttime
images simultaneously through a robust non-linear curve projection. Moreover,
to provide a comprehensive evaluation, we construct a challenging nighttime
tracking benchmark, namely DarkTrack2021, which contains 110 challenging
sequences with over 100 K frames in total. Evaluations on both the public
UAVDark135 benchmark and the newly constructed DarkTrack2021 benchmark show
that the task-inspired design enables SCT with significant performance gains
for nighttime UAV tracking compared with other top-ranked low-light enhancers.
Real-world tests on a typical UAV platform further verify the practicability of
the proposed approach. The DarkTrack2021 benchmark and the code of the proposed
approach are publicly available at https://github.com/vision4robotics/SCT.
| new_dataset | 0.949106 |
2303.10960 | Heather Doig | Heather Doig, Oscar Pizarro and Stefan B. Williams | Improved Benthic Classification using Resolution Scaling and SymmNet
Unsupervised Domain Adaptation | 7 pages, 6 figures. Accepted to IEEE International Conference on
Robotics and Automation (ICRA) 2023, London UK | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous Underwater Vehicles (AUVs) conduct regular visual surveys of
marine environments to characterise and monitor the composition and diversity
of the benthos. The use of machine learning classifiers for this task is
limited by the low numbers of annotations available and the many fine-grained
classes involved. In addition to these challenges, there are domain shifts
between image sets acquired during different AUV surveys due to changes in
camera systems, imaging altitude, illumination and water column properties
leading to a drop in classification performance for images from a different
survey where some or all these elements may have changed. This paper proposes a
framework to improve the performance of a benthic morphospecies classifier when
used to classify images from a different survey compared to the training data.
We adapt the SymmNet state-of-the-art Unsupervised Domain Adaptation method
with an efficient bilinear pooling layer and image scaling to normalise spatial
resolution, and show improved classification accuracy. We test our approach on
two datasets with images from AUV surveys with different imaging payloads and
locations. The results show that generic domain adaptation can be enhanced to
produce a significant increase in accuracy for images from an AUV survey that
differs from the training images.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:33:47 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Doig",
"Heather",
""
],
[
"Pizarro",
"Oscar",
""
],
[
"Williams",
"Stefan B.",
""
]
] | TITLE: Improved Benthic Classification using Resolution Scaling and SymmNet
Unsupervised Domain Adaptation
ABSTRACT: Autonomous Underwater Vehicles (AUVs) conduct regular visual surveys of
marine environments to characterise and monitor the composition and diversity
of the benthos. The use of machine learning classifiers for this task is
limited by the low numbers of annotations available and the many fine-grained
classes involved. In addition to these challenges, there are domain shifts
between image sets acquired during different AUV surveys due to changes in
camera systems, imaging altitude, illumination and water column properties
leading to a drop in classification performance for images from a different
survey where some or all these elements may have changed. This paper proposes a
framework to improve the performance of a benthic morphospecies classifier when
used to classify images from a different survey compared to the training data.
We adapt the SymmNet state-of-the-art Unsupervised Domain Adaptation method
with an efficient bilinear pooling layer and image scaling to normalise spatial
resolution, and show improved classification accuracy. We test our approach on
two datasets with images from AUV surveys with different imaging payloads and
locations. The results show that generic domain adaptation can be enhanced to
produce a significant increase in accuracy for images from an AUV survey that
differs from the training images.
| no_new_dataset | 0.954095 |
2303.10962 | Kenneth Blomqvist | Kenneth Blomqvist, Francesco Milano, Jen Jen Chung, Lionel Ott, Roland
Siegwart | Neural Implicit Vision-Language Feature Fields | null | null | null | null | cs.RO cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recently, groundbreaking results have been presented on open-vocabulary
semantic image segmentation. Such methods segment each pixel in an image into
arbitrary categories provided at run-time in the form of text prompts, as
opposed to a fixed set of classes defined at training time. In this work, we
present a zero-shot volumetric open-vocabulary semantic scene segmentation
method. Our method builds on the insight that we can fuse image features from a
vision-language model into a neural implicit representation. We show that the
resulting feature field can be segmented into different classes by assigning
points to natural language text prompts. The implicit volumetric representation
enables us to segment the scene both in 3D and 2D by rendering feature maps
from any given viewpoint of the scene. We show that our method works on noisy
real-world data and can run in real-time on live sensor data dynamically
adjusting to text prompts. We also present quantitative comparisons on the
ScanNet dataset.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:38:09 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Blomqvist",
"Kenneth",
""
],
[
"Milano",
"Francesco",
""
],
[
"Chung",
"Jen Jen",
""
],
[
"Ott",
"Lionel",
""
],
[
"Siegwart",
"Roland",
""
]
] | TITLE: Neural Implicit Vision-Language Feature Fields
ABSTRACT: Recently, groundbreaking results have been presented on open-vocabulary
semantic image segmentation. Such methods segment each pixel in an image into
arbitrary categories provided at run-time in the form of text prompts, as
opposed to a fixed set of classes defined at training time. In this work, we
present a zero-shot volumetric open-vocabulary semantic scene segmentation
method. Our method builds on the insight that we can fuse image features from a
vision-language model into a neural implicit representation. We show that the
resulting feature field can be segmented into different classes by assigning
points to natural language text prompts. The implicit volumetric representation
enables us to segment the scene both in 3D and 2D by rendering feature maps
from any given viewpoint of the scene. We show that our method works on noisy
real-world data and can run in real-time on live sensor data dynamically
adjusting to text prompts. We also present quantitative comparisons on the
ScanNet dataset.
| no_new_dataset | 0.950824 |
2303.10971 | Dongliang Cao | Dongliang Cao, Florian Bernard | Self-Supervised Learning for Multimodal Non-Rigid 3D Shape Matching | accepted by CVPR 2023 | null | null | null | cs.CV cs.AI cs.CG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The matching of 3D shapes has been extensively studied for shapes represented
as surface meshes, as well as for shapes represented as point clouds. While
point clouds are a common representation of raw real-world 3D data (e.g. from
laser scanners), meshes encode rich and expressive topological information, but
their creation typically requires some form of (often manual) curation. In
turn, methods that purely rely on point clouds are unable to meet the matching
quality of mesh-based methods that utilise the additional topological
structure. In this work we close this gap by introducing a self-supervised
multimodal learning strategy that combines mesh-based functional map
regularisation with a contrastive loss that couples mesh and point cloud data.
Our shape matching approach allows to obtain intramodal correspondences for
triangle meshes, complete point clouds, and partially observed point clouds, as
well as correspondences across these data modalities. We demonstrate that our
method achieves state-of-the-art results on several challenging benchmark
datasets even in comparison to recent supervised methods, and that our method
reaches previously unseen cross-dataset generalisation ability.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:47:02 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Cao",
"Dongliang",
""
],
[
"Bernard",
"Florian",
""
]
] | TITLE: Self-Supervised Learning for Multimodal Non-Rigid 3D Shape Matching
ABSTRACT: The matching of 3D shapes has been extensively studied for shapes represented
as surface meshes, as well as for shapes represented as point clouds. While
point clouds are a common representation of raw real-world 3D data (e.g. from
laser scanners), meshes encode rich and expressive topological information, but
their creation typically requires some form of (often manual) curation. In
turn, methods that purely rely on point clouds are unable to meet the matching
quality of mesh-based methods that utilise the additional topological
structure. In this work we close this gap by introducing a self-supervised
multimodal learning strategy that combines mesh-based functional map
regularisation with a contrastive loss that couples mesh and point cloud data.
Our shape matching approach allows to obtain intramodal correspondences for
triangle meshes, complete point clouds, and partially observed point clouds, as
well as correspondences across these data modalities. We demonstrate that our
method achieves state-of-the-art results on several challenging benchmark
datasets even in comparison to recent supervised methods, and that our method
reaches previously unseen cross-dataset generalisation ability.
| no_new_dataset | 0.950088 |
2303.10975 | Wang Zhe | Zhe Wang, Siqi Fan, Xiaoliang Huo, Tongda Xu, Yan Wang, Jingjing Liu,
Yilun Chen, Ya-Qin Zhang | VIMI: Vehicle-Infrastructure Multi-view Intermediate Fusion for
Camera-based 3D Object Detection | 8 pages, 9 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In autonomous driving, Vehicle-Infrastructure Cooperative 3D Object Detection
(VIC3D) makes use of multi-view cameras from both vehicles and traffic
infrastructure, providing a global vantage point with rich semantic context of
road conditions beyond a single vehicle viewpoint. Two major challenges prevail
in VIC3D: 1) inherent calibration noise when fusing multi-view images, caused
by time asynchrony across cameras; 2) information loss when projecting 2D
features into 3D space. To address these issues, We propose a novel 3D object
detection framework, Vehicles-Infrastructure Multi-view Intermediate fusion
(VIMI). First, to fully exploit the holistic perspectives from both vehicles
and infrastructure, we propose a Multi-scale Cross Attention (MCA) module that
fuses infrastructure and vehicle features on selective multi-scales to correct
the calibration noise introduced by camera asynchrony. Then, we design a
Camera-aware Channel Masking (CCM) module that uses camera parameters as priors
to augment the fused features. We further introduce a Feature Compression (FC)
module with channel and spatial compression blocks to reduce the size of
transmitted features for enhanced efficiency. Experiments show that VIMI
achieves 15.61% overall AP_3D and 21.44% AP_BEV on the new VIC3D dataset,
DAIR-V2X-C, significantly outperforming state-of-the-art early fusion and late
fusion methods with comparable transmission cost.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 09:56:17 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wang",
"Zhe",
""
],
[
"Fan",
"Siqi",
""
],
[
"Huo",
"Xiaoliang",
""
],
[
"Xu",
"Tongda",
""
],
[
"Wang",
"Yan",
""
],
[
"Liu",
"Jingjing",
""
],
[
"Chen",
"Yilun",
""
],
[
"Zhang",
"Ya-Qin",
""
]
] | TITLE: VIMI: Vehicle-Infrastructure Multi-view Intermediate Fusion for
Camera-based 3D Object Detection
ABSTRACT: In autonomous driving, Vehicle-Infrastructure Cooperative 3D Object Detection
(VIC3D) makes use of multi-view cameras from both vehicles and traffic
infrastructure, providing a global vantage point with rich semantic context of
road conditions beyond a single vehicle viewpoint. Two major challenges prevail
in VIC3D: 1) inherent calibration noise when fusing multi-view images, caused
by time asynchrony across cameras; 2) information loss when projecting 2D
features into 3D space. To address these issues, We propose a novel 3D object
detection framework, Vehicles-Infrastructure Multi-view Intermediate fusion
(VIMI). First, to fully exploit the holistic perspectives from both vehicles
and infrastructure, we propose a Multi-scale Cross Attention (MCA) module that
fuses infrastructure and vehicle features on selective multi-scales to correct
the calibration noise introduced by camera asynchrony. Then, we design a
Camera-aware Channel Masking (CCM) module that uses camera parameters as priors
to augment the fused features. We further introduce a Feature Compression (FC)
module with channel and spatial compression blocks to reduce the size of
transmitted features for enhanced efficiency. Experiments show that VIMI
achieves 15.61% overall AP_3D and 21.44% AP_BEV on the new VIC3D dataset,
DAIR-V2X-C, significantly outperforming state-of-the-art early fusion and late
fusion methods with comparable transmission cost.
| no_new_dataset | 0.940079 |
2303.10993 | T. Konstantin Rusch | T. Konstantin Rusch, Michael M. Bronstein, Siddhartha Mishra | A Survey on Oversmoothing in Graph Neural Networks | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Node features of graph neural networks (GNNs) tend to become more similar
with the increase of the network depth. This effect is known as over-smoothing,
which we axiomatically define as the exponential convergence of suitable
similarity measures on the node features. Our definition unifies previous
approaches and gives rise to new quantitative measures of over-smoothing.
Moreover, we empirically demonstrate this behavior for several over-smoothing
measures on different graphs (small-, medium-, and large-scale). We also review
several approaches for mitigating over-smoothing and empirically test their
effectiveness on real-world graph datasets. Through illustrative examples, we
demonstrate that mitigating over-smoothing is a necessary but not sufficient
condition for building deep GNNs that are expressive on a wide range of graph
learning tasks. Finally, we extend our definition of over-smoothing to the
rapidly emerging field of continuous-time GNNs.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 10:21:29 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Rusch",
"T. Konstantin",
""
],
[
"Bronstein",
"Michael M.",
""
],
[
"Mishra",
"Siddhartha",
""
]
] | TITLE: A Survey on Oversmoothing in Graph Neural Networks
ABSTRACT: Node features of graph neural networks (GNNs) tend to become more similar
with the increase of the network depth. This effect is known as over-smoothing,
which we axiomatically define as the exponential convergence of suitable
similarity measures on the node features. Our definition unifies previous
approaches and gives rise to new quantitative measures of over-smoothing.
Moreover, we empirically demonstrate this behavior for several over-smoothing
measures on different graphs (small-, medium-, and large-scale). We also review
several approaches for mitigating over-smoothing and empirically test their
effectiveness on real-world graph datasets. Through illustrative examples, we
demonstrate that mitigating over-smoothing is a necessary but not sufficient
condition for building deep GNNs that are expressive on a wide range of graph
learning tasks. Finally, we extend our definition of over-smoothing to the
rapidly emerging field of continuous-time GNNs.
| no_new_dataset | 0.95222 |
2303.10999 | Nathan Hubens | Nathan Hubens, Victor Delvigne, Matei Mancas, Bernard Gosselin, Marius
Preda, Titus Zaharia | Induced Feature Selection by Structured Pruning | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The advent of sparsity inducing techniques in neural networks has been of a
great help in the last few years. Indeed, those methods allowed to find lighter
and faster networks, able to perform more efficiently in resource-constrained
environment such as mobile devices or highly requested servers. Such a sparsity
is generally imposed on the weights of neural networks, reducing the footprint
of the architecture. In this work, we go one step further by imposing sparsity
jointly on the weights and on the input data. This can be achieved following a
three-step process: 1) impose a certain structured sparsity on the weights of
the network; 2) track back input features corresponding to zeroed blocks of
weight; 3) remove useless weights and input features and retrain the network.
Performing pruning both on the network and on input data not only allows for
extreme reduction in terms of parameters and operations but can also serve as
an interpretation process. Indeed, with the help of data pruning, we now have
information about which input feature is useful for the network to keep its
performance. Experiments conducted on a variety of architectures and datasets:
MLP validated on MNIST, CIFAR10/100 and ConvNets (VGG16 and ResNet18),
validated on CIFAR10/100 and CALTECH101 respectively, show that it is possible
to achieve additional gains in terms of total parameters and in FLOPs by
performing pruning on input data, while also increasing accuracy.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 10:29:35 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Hubens",
"Nathan",
""
],
[
"Delvigne",
"Victor",
""
],
[
"Mancas",
"Matei",
""
],
[
"Gosselin",
"Bernard",
""
],
[
"Preda",
"Marius",
""
],
[
"Zaharia",
"Titus",
""
]
] | TITLE: Induced Feature Selection by Structured Pruning
ABSTRACT: The advent of sparsity inducing techniques in neural networks has been of a
great help in the last few years. Indeed, those methods allowed to find lighter
and faster networks, able to perform more efficiently in resource-constrained
environment such as mobile devices or highly requested servers. Such a sparsity
is generally imposed on the weights of neural networks, reducing the footprint
of the architecture. In this work, we go one step further by imposing sparsity
jointly on the weights and on the input data. This can be achieved following a
three-step process: 1) impose a certain structured sparsity on the weights of
the network; 2) track back input features corresponding to zeroed blocks of
weight; 3) remove useless weights and input features and retrain the network.
Performing pruning both on the network and on input data not only allows for
extreme reduction in terms of parameters and operations but can also serve as
an interpretation process. Indeed, with the help of data pruning, we now have
information about which input feature is useful for the network to keep its
performance. Experiments conducted on a variety of architectures and datasets:
MLP validated on MNIST, CIFAR10/100 and ConvNets (VGG16 and ResNet18),
validated on CIFAR10/100 and CALTECH101 respectively, show that it is possible
to achieve additional gains in terms of total parameters and in FLOPs by
performing pruning on input data, while also increasing accuracy.
| no_new_dataset | 0.944177 |
2303.11009 | Binbin Wang | Binbin Wang, Mingming Li, Zhixiong Zeng, Jingwei Zhuo, Songlin Wang,
Sulong Xu, Bo Long, Weipeng Yan | Learning Multi-Stage Multi-Grained Semantic Embeddings for E-Commerce
Search | null | null | 10.1145/3543873.3584638 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieving relevant items that match users' queries from billion-scale corpus
forms the core of industrial e-commerce search systems, in which
embedding-based retrieval (EBR) methods are prevailing. These methods adopt a
two-tower framework to learn embedding vectors for query and item separately
and thus leverage efficient approximate nearest neighbor (ANN) search to
retrieve relevant items. However, existing EBR methods usually ignore
inconsistent user behaviors in industrial multi-stage search systems, resulting
in insufficient retrieval efficiency with a low commercial return. To tackle
this challenge, we propose to improve EBR methods by learning Multi-level
Multi-Grained Semantic Embeddings(MMSE). We propose the multi-stage information
mining to exploit the ordered, clicked, unclicked and random sampled items in
practical user behavior data, and then capture query-item similarity via a
post-fusion strategy. We then propose multi-grained learning objectives that
integrate the retrieval loss with global comparison ability and the ranking
loss with local comparison ability to generate semantic embeddings. Both
experiments on a real-world billion-scale dataset and online A/B tests verify
the effectiveness of MMSE in achieving significant performance improvements on
metrics such as offline recall and online conversion rate (CVR).
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 10:39:25 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wang",
"Binbin",
""
],
[
"Li",
"Mingming",
""
],
[
"Zeng",
"Zhixiong",
""
],
[
"Zhuo",
"Jingwei",
""
],
[
"Wang",
"Songlin",
""
],
[
"Xu",
"Sulong",
""
],
[
"Long",
"Bo",
""
],
[
"Yan",
"Weipeng",
""
]
] | TITLE: Learning Multi-Stage Multi-Grained Semantic Embeddings for E-Commerce
Search
ABSTRACT: Retrieving relevant items that match users' queries from billion-scale corpus
forms the core of industrial e-commerce search systems, in which
embedding-based retrieval (EBR) methods are prevailing. These methods adopt a
two-tower framework to learn embedding vectors for query and item separately
and thus leverage efficient approximate nearest neighbor (ANN) search to
retrieve relevant items. However, existing EBR methods usually ignore
inconsistent user behaviors in industrial multi-stage search systems, resulting
in insufficient retrieval efficiency with a low commercial return. To tackle
this challenge, we propose to improve EBR methods by learning Multi-level
Multi-Grained Semantic Embeddings(MMSE). We propose the multi-stage information
mining to exploit the ordered, clicked, unclicked and random sampled items in
practical user behavior data, and then capture query-item similarity via a
post-fusion strategy. We then propose multi-grained learning objectives that
integrate the retrieval loss with global comparison ability and the ranking
loss with local comparison ability to generate semantic embeddings. Both
experiments on a real-world billion-scale dataset and online A/B tests verify
the effectiveness of MMSE in achieving significant performance improvements on
metrics such as offline recall and online conversion rate (CVR).
| no_new_dataset | 0.941493 |
2303.11011 | Xinglong Luo | Xinglong Luo, Kunming Luo, Ao Luo, Zhengning Wang, Ping Tan,
Shuaicheng Liu | Learning Optical Flow from Event Camera with Rendered Dataset | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of estimating optical flow from event cameras. One
important issue is how to build a high-quality event-flow dataset with accurate
event values and flow labels. Previous datasets are created by either capturing
real scenes by event cameras or synthesizing from images with pasted foreground
objects. The former case can produce real event values but with calculated flow
labels, which are sparse and inaccurate. The later case can generate dense flow
labels but the interpolated events are prone to errors. In this work, we
propose to render a physically correct event-flow dataset using computer
graphics models. In particular, we first create indoor and outdoor 3D scenes by
Blender with rich scene content variations. Second, diverse camera motions are
included for the virtual capturing, producing images and accurate flow labels.
Third, we render high-framerate videos between images for accurate events. The
rendered dataset can adjust the density of events, based on which we further
introduce an adaptive density module (ADM). Experiments show that our proposed
dataset can facilitate event-flow learning, whereas previous approaches when
trained on our dataset can improve their performances constantly by a
relatively large margin. In addition, event-flow pipelines when equipped with
our ADM can further improve performances.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 10:44:32 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Luo",
"Xinglong",
""
],
[
"Luo",
"Kunming",
""
],
[
"Luo",
"Ao",
""
],
[
"Wang",
"Zhengning",
""
],
[
"Tan",
"Ping",
""
],
[
"Liu",
"Shuaicheng",
""
]
] | TITLE: Learning Optical Flow from Event Camera with Rendered Dataset
ABSTRACT: We study the problem of estimating optical flow from event cameras. One
important issue is how to build a high-quality event-flow dataset with accurate
event values and flow labels. Previous datasets are created by either capturing
real scenes by event cameras or synthesizing from images with pasted foreground
objects. The former case can produce real event values but with calculated flow
labels, which are sparse and inaccurate. The later case can generate dense flow
labels but the interpolated events are prone to errors. In this work, we
propose to render a physically correct event-flow dataset using computer
graphics models. In particular, we first create indoor and outdoor 3D scenes by
Blender with rich scene content variations. Second, diverse camera motions are
included for the virtual capturing, producing images and accurate flow labels.
Third, we render high-framerate videos between images for accurate events. The
rendered dataset can adjust the density of events, based on which we further
introduce an adaptive density module (ADM). Experiments show that our proposed
dataset can facilitate event-flow learning, whereas previous approaches when
trained on our dataset can improve their performances constantly by a
relatively large margin. In addition, event-flow pipelines when equipped with
our ADM can further improve performances.
| new_dataset | 0.970212 |
2303.11034 | Haohao Sun | Haohao Sun, Yilong Zhang, Peng Chen, Haixia Wang, Ronghua Liang | Internal Structure Attention Network for Fingerprint Presentation Attack
Detection from Optical Coherence Tomography | 12 pages, 14 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As a non-invasive optical imaging technique, optical coherence tomography
(OCT) has proven promising for automatic fingerprint recognition system (AFRS)
applications. Diverse approaches have been proposed for OCT-based fingerprint
presentation attack detection (PAD). However, considering the complexity and
variety of PA samples, it is extremely challenging to increase the
generalization ability with the limited PA dataset. To solve the challenge,
this paper presents a novel supervised learning-based PAD method, denoted as
ISAPAD, which applies prior knowledge to guide network training and enhance the
generalization ability. The proposed dual-branch architecture can not only
learns global features from the OCT image, but also concentrate on layered
structure feature which comes from the internal structure attention module
(ISAM). The simple yet effective ISAM enables the proposed network to obtain
layered segmentation features belonging only to Bonafide from noisy OCT volume
data directly. Combined with effective training strategies and PAD score
generation rules, ISAPAD obtains optimal PAD performance in limited training
data. Domain generalization experiments and visualization analysis validate the
effectiveness of the proposed method for OCT PAD.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 11:36:09 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Sun",
"Haohao",
""
],
[
"Zhang",
"Yilong",
""
],
[
"Chen",
"Peng",
""
],
[
"Wang",
"Haixia",
""
],
[
"Liang",
"Ronghua",
""
]
] | TITLE: Internal Structure Attention Network for Fingerprint Presentation Attack
Detection from Optical Coherence Tomography
ABSTRACT: As a non-invasive optical imaging technique, optical coherence tomography
(OCT) has proven promising for automatic fingerprint recognition system (AFRS)
applications. Diverse approaches have been proposed for OCT-based fingerprint
presentation attack detection (PAD). However, considering the complexity and
variety of PA samples, it is extremely challenging to increase the
generalization ability with the limited PA dataset. To solve the challenge,
this paper presents a novel supervised learning-based PAD method, denoted as
ISAPAD, which applies prior knowledge to guide network training and enhance the
generalization ability. The proposed dual-branch architecture can not only
learns global features from the OCT image, but also concentrate on layered
structure feature which comes from the internal structure attention module
(ISAM). The simple yet effective ISAM enables the proposed network to obtain
layered segmentation features belonging only to Bonafide from noisy OCT volume
data directly. Combined with effective training strategies and PAD score
generation rules, ISAPAD obtains optimal PAD performance in limited training
data. Domain generalization experiments and visualization analysis validate the
effectiveness of the proposed method for OCT PAD.
| no_new_dataset | 0.949201 |
2303.11040 | Yinpeng Dong | Yinpeng Dong, Caixin Kang, Jinlai Zhang, Zijian Zhu, Yikai Wang, Xiao
Yang, Hang Su, Xingxing Wei, Jun Zhu | Benchmarking Robustness of 3D Object Detection to Common Corruptions in
Autonomous Driving | CVPR 2023 | null | null | null | cs.CV cs.AI cs.CR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | 3D object detection is an important task in autonomous driving to perceive
the surroundings. Despite the excellent performance, the existing 3D detectors
lack the robustness to real-world corruptions caused by adverse weathers,
sensor noises, etc., provoking concerns about the safety and reliability of
autonomous driving systems. To comprehensively and rigorously benchmark the
corruption robustness of 3D detectors, in this paper we design 27 types of
common corruptions for both LiDAR and camera inputs considering real-world
driving scenarios. By synthesizing these corruptions on public datasets, we
establish three corruption robustness benchmarks -- KITTI-C, nuScenes-C, and
Waymo-C. Then, we conduct large-scale experiments on 24 diverse 3D object
detection models to evaluate their corruption robustness. Based on the
evaluation results, we draw several important findings, including: 1)
motion-level corruptions are the most threatening ones that lead to significant
performance drop of all models; 2) LiDAR-camera fusion models demonstrate
better robustness; 3) camera-only models are extremely vulnerable to image
corruptions, showing the indispensability of LiDAR point clouds. We release the
benchmarks and codes at https://github.com/kkkcx/3D_Corruptions_AD. We hope
that our benchmarks and findings can provide insights for future research on
developing robust 3D object detection models.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 11:45:54 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Dong",
"Yinpeng",
""
],
[
"Kang",
"Caixin",
""
],
[
"Zhang",
"Jinlai",
""
],
[
"Zhu",
"Zijian",
""
],
[
"Wang",
"Yikai",
""
],
[
"Yang",
"Xiao",
""
],
[
"Su",
"Hang",
""
],
[
"Wei",
"Xingxing",
""
],
[
"Zhu",
"Jun",
""
]
] | TITLE: Benchmarking Robustness of 3D Object Detection to Common Corruptions in
Autonomous Driving
ABSTRACT: 3D object detection is an important task in autonomous driving to perceive
the surroundings. Despite the excellent performance, the existing 3D detectors
lack the robustness to real-world corruptions caused by adverse weathers,
sensor noises, etc., provoking concerns about the safety and reliability of
autonomous driving systems. To comprehensively and rigorously benchmark the
corruption robustness of 3D detectors, in this paper we design 27 types of
common corruptions for both LiDAR and camera inputs considering real-world
driving scenarios. By synthesizing these corruptions on public datasets, we
establish three corruption robustness benchmarks -- KITTI-C, nuScenes-C, and
Waymo-C. Then, we conduct large-scale experiments on 24 diverse 3D object
detection models to evaluate their corruption robustness. Based on the
evaluation results, we draw several important findings, including: 1)
motion-level corruptions are the most threatening ones that lead to significant
performance drop of all models; 2) LiDAR-camera fusion models demonstrate
better robustness; 3) camera-only models are extremely vulnerable to image
corruptions, showing the indispensability of LiDAR point clouds. We release the
benchmarks and codes at https://github.com/kkkcx/3D_Corruptions_AD. We hope
that our benchmarks and findings can provide insights for future research on
developing robust 3D object detection models.
| no_new_dataset | 0.838084 |
2303.11053 | Vishwa Prakash HV | Aadityan Ganesh, Prajakta Nimbhorkar, Pratik Ghosal and Vishwa Prakash
HV | Fair Healthcare Rationing to Maximize Dynamic Utilities | null | null | null | null | cs.MA | http://creativecommons.org/licenses/by-sa/4.0/ | Allocation of scarce healthcare resources under limited logistic and
infrastructural facilities is a major issue in the modern society. We consider
the problem of allocation of healthcare resources like vaccines to people or
hospital beds to patients in an online manner. Our model takes into account the
arrival of resources on a day-to-day basis, different categories of agents, the
possible unavailability of agents on certain days, and the utility associated
with each allotment as well as its variation over time.
We propose a model where priorities for various categories are modelled in
terms of utilities of agents. We give online and offline algorithms to compute
an allocation that respects eligibility of agents into different categories,
and incentivizes agents not to hide their eligibility for some category. The
offline algorithm gives an optimal allocation while the on-line algorithm gives
an approximation to the optimal allocation in terms of total utility. Our
algorithms are efficient, and maintain fairness among different categories of
agents. Our models have applications in other areas like refugee settlement and
visa allocation. We evaluate the performance of our algorithms on real-life and
synthetic datasets. The experimental results show that the online algorithm is
fast and performs better than the given theoretical bound in terms of total
utility. Moreover, the experimental results confirm that our utility-based
model correctly captures the priorities of categories
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 12:07:21 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Ganesh",
"Aadityan",
""
],
[
"Nimbhorkar",
"Prajakta",
""
],
[
"Ghosal",
"Pratik",
""
],
[
"HV",
"Vishwa Prakash",
""
]
] | TITLE: Fair Healthcare Rationing to Maximize Dynamic Utilities
ABSTRACT: Allocation of scarce healthcare resources under limited logistic and
infrastructural facilities is a major issue in the modern society. We consider
the problem of allocation of healthcare resources like vaccines to people or
hospital beds to patients in an online manner. Our model takes into account the
arrival of resources on a day-to-day basis, different categories of agents, the
possible unavailability of agents on certain days, and the utility associated
with each allotment as well as its variation over time.
We propose a model where priorities for various categories are modelled in
terms of utilities of agents. We give online and offline algorithms to compute
an allocation that respects eligibility of agents into different categories,
and incentivizes agents not to hide their eligibility for some category. The
offline algorithm gives an optimal allocation while the on-line algorithm gives
an approximation to the optimal allocation in terms of total utility. Our
algorithms are efficient, and maintain fairness among different categories of
agents. Our models have applications in other areas like refugee settlement and
visa allocation. We evaluate the performance of our algorithms on real-life and
synthetic datasets. The experimental results show that the online algorithm is
fast and performs better than the given theoretical bound in terms of total
utility. Moreover, the experimental results confirm that our utility-based
model correctly captures the priorities of categories
| no_new_dataset | 0.948394 |
2303.11066 | Yuhao Chen | Yuhao Chen, Xin Tan, Borui Zhao, Zhaowei Chen, Renjie Song, Jiajun
Liang, Xuequan Lu | Boosting Semi-Supervised Learning by Exploiting All Unlabeled Data | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-supervised learning (SSL) has attracted enormous attention due to its
vast potential of mitigating the dependence on large labeled datasets. The
latest methods (e.g., FixMatch) use a combination of consistency regularization
and pseudo-labeling to achieve remarkable successes. However, these methods all
suffer from the waste of complicated examples since all pseudo-labels have to
be selected by a high threshold to filter out noisy ones. Hence, the examples
with ambiguous predictions will not contribute to the training phase. For
better leveraging all unlabeled examples, we propose two novel techniques:
Entropy Meaning Loss (EML) and Adaptive Negative Learning (ANL). EML
incorporates the prediction distribution of non-target classes into the
optimization objective to avoid competition with target class, and thus
generating more high-confidence predictions for selecting pseudo-label. ANL
introduces the additional negative pseudo-label for all unlabeled data to
leverage low-confidence examples. It adaptively allocates this label by
dynamically evaluating the top-k performance of the model. EML and ANL do not
introduce any additional parameter and hyperparameter. We integrate these
techniques with FixMatch, and develop a simple yet powerful framework called
FullMatch. Extensive experiments on several common SSL benchmarks
(CIFAR-10/100, SVHN, STL-10 and ImageNet) demonstrate that FullMatch exceeds
FixMatch by a large margin. Integrated with FlexMatch (an advanced
FixMatch-based framework), we achieve state-of-the-art performance. Source code
is at https://github.com/megvii-research/FullMatch.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 12:44:11 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Chen",
"Yuhao",
""
],
[
"Tan",
"Xin",
""
],
[
"Zhao",
"Borui",
""
],
[
"Chen",
"Zhaowei",
""
],
[
"Song",
"Renjie",
""
],
[
"Liang",
"Jiajun",
""
],
[
"Lu",
"Xuequan",
""
]
] | TITLE: Boosting Semi-Supervised Learning by Exploiting All Unlabeled Data
ABSTRACT: Semi-supervised learning (SSL) has attracted enormous attention due to its
vast potential of mitigating the dependence on large labeled datasets. The
latest methods (e.g., FixMatch) use a combination of consistency regularization
and pseudo-labeling to achieve remarkable successes. However, these methods all
suffer from the waste of complicated examples since all pseudo-labels have to
be selected by a high threshold to filter out noisy ones. Hence, the examples
with ambiguous predictions will not contribute to the training phase. For
better leveraging all unlabeled examples, we propose two novel techniques:
Entropy Meaning Loss (EML) and Adaptive Negative Learning (ANL). EML
incorporates the prediction distribution of non-target classes into the
optimization objective to avoid competition with target class, and thus
generating more high-confidence predictions for selecting pseudo-label. ANL
introduces the additional negative pseudo-label for all unlabeled data to
leverage low-confidence examples. It adaptively allocates this label by
dynamically evaluating the top-k performance of the model. EML and ANL do not
introduce any additional parameter and hyperparameter. We integrate these
techniques with FixMatch, and develop a simple yet powerful framework called
FullMatch. Extensive experiments on several common SSL benchmarks
(CIFAR-10/100, SVHN, STL-10 and ImageNet) demonstrate that FullMatch exceeds
FixMatch by a large margin. Integrated with FlexMatch (an advanced
FixMatch-based framework), we achieve state-of-the-art performance. Source code
is at https://github.com/megvii-research/FullMatch.
| no_new_dataset | 0.945751 |
2303.11074 | Saffron Huang | Saffron Huang and Divya Siddarth | Generative AI and the Digital Commons | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by-sa/4.0/ | Many generative foundation models (or GFMs) are trained on publicly available
data and use public infrastructure, but 1) may degrade the "digital commons"
that they depend on, and 2) do not have processes in place to return value
captured to data producers and stakeholders. Existing conceptions of data
rights and protection (focusing largely on individually-owned data and
associated privacy concerns) and copyright or licensing-based models offer some
instructive priors, but are ill-suited for the issues that may arise from
models trained on commons-based data. We outline the risks posed by GFMs and
why they are relevant to the digital commons, and propose numerous
governance-based solutions that include investments in standardized
dataset/model disclosure and other kinds of transparency when it comes to
generative models' training and capabilities, consortia-based funding for
monitoring/standards/auditing organizations, requirements or norms for GFM
companies to contribute high quality data to the commons, and structures for
shared ownership based on individual or community provision of fine-tuning
data.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:01:48 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Huang",
"Saffron",
""
],
[
"Siddarth",
"Divya",
""
]
] | TITLE: Generative AI and the Digital Commons
ABSTRACT: Many generative foundation models (or GFMs) are trained on publicly available
data and use public infrastructure, but 1) may degrade the "digital commons"
that they depend on, and 2) do not have processes in place to return value
captured to data producers and stakeholders. Existing conceptions of data
rights and protection (focusing largely on individually-owned data and
associated privacy concerns) and copyright or licensing-based models offer some
instructive priors, but are ill-suited for the issues that may arise from
models trained on commons-based data. We outline the risks posed by GFMs and
why they are relevant to the digital commons, and propose numerous
governance-based solutions that include investments in standardized
dataset/model disclosure and other kinds of transparency when it comes to
generative models' training and capabilities, consortia-based funding for
monitoring/standards/auditing organizations, requirements or norms for GFM
companies to contribute high quality data to the commons, and structures for
shared ownership based on individual or community provision of fine-tuning
data.
| no_new_dataset | 0.948917 |
2303.11076 | Kamil Faber | Kamil Faber, Dominik Zurek, Marcin Pietron, Nathalie Japkowicz,
Antonio Vergari, Roberto Corizzo | From MNIST to ImageNet and Back: Benchmarking Continual Curriculum
Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Continual learning (CL) is one of the most promising trends in recent machine
learning research. Its goal is to go beyond classical assumptions in machine
learning and develop models and learning strategies that present high
robustness in dynamic environments. The landscape of CL research is fragmented
into several learning evaluation protocols, comprising different learning
tasks, datasets, and evaluation metrics. Additionally, the benchmarks adopted
so far are still distant from the complexity of real-world scenarios, and are
usually tailored to highlight capabilities specific to certain strategies. In
such a landscape, it is hard to objectively assess strategies. In this work, we
fill this gap for CL on image data by introducing two novel CL benchmarks that
involve multiple heterogeneous tasks from six image datasets, with varying
levels of complexity and quality. Our aim is to fairly evaluate current
state-of-the-art CL strategies on a common ground that is closer to complex
real-world scenarios. We additionally structure our benchmarks so that tasks
are presented in increasing and decreasing order of complexity -- according to
a curriculum -- in order to evaluate if current CL models are able to exploit
structure across tasks. We devote particular emphasis to providing the CL
community with a rigorous and reproducible evaluation protocol for measuring
the ability of a model to generalize and not to forget while learning.
Furthermore, we provide an extensive experimental evaluation showing that
popular CL strategies, when challenged with our benchmarks, yield sub-par
performance, high levels of forgetting, and present a limited ability to
effectively leverage curriculum task ordering. We believe that these results
highlight the need for rigorous comparisons in future CL works as well as pave
the way to design new CL strategies that are able to deal with more complex
scenarios.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 18:11:19 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Faber",
"Kamil",
""
],
[
"Zurek",
"Dominik",
""
],
[
"Pietron",
"Marcin",
""
],
[
"Japkowicz",
"Nathalie",
""
],
[
"Vergari",
"Antonio",
""
],
[
"Corizzo",
"Roberto",
""
]
] | TITLE: From MNIST to ImageNet and Back: Benchmarking Continual Curriculum
Learning
ABSTRACT: Continual learning (CL) is one of the most promising trends in recent machine
learning research. Its goal is to go beyond classical assumptions in machine
learning and develop models and learning strategies that present high
robustness in dynamic environments. The landscape of CL research is fragmented
into several learning evaluation protocols, comprising different learning
tasks, datasets, and evaluation metrics. Additionally, the benchmarks adopted
so far are still distant from the complexity of real-world scenarios, and are
usually tailored to highlight capabilities specific to certain strategies. In
such a landscape, it is hard to objectively assess strategies. In this work, we
fill this gap for CL on image data by introducing two novel CL benchmarks that
involve multiple heterogeneous tasks from six image datasets, with varying
levels of complexity and quality. Our aim is to fairly evaluate current
state-of-the-art CL strategies on a common ground that is closer to complex
real-world scenarios. We additionally structure our benchmarks so that tasks
are presented in increasing and decreasing order of complexity -- according to
a curriculum -- in order to evaluate if current CL models are able to exploit
structure across tasks. We devote particular emphasis to providing the CL
community with a rigorous and reproducible evaluation protocol for measuring
the ability of a model to generalize and not to forget while learning.
Furthermore, we provide an extensive experimental evaluation showing that
popular CL strategies, when challenged with our benchmarks, yield sub-par
performance, high levels of forgetting, and present a limited ability to
effectively leverage curriculum task ordering. We believe that these results
highlight the need for rigorous comparisons in future CL works as well as pave
the way to design new CL strategies that are able to deal with more complex
scenarios.
| no_new_dataset | 0.940735 |
2303.11078 | Lianyu Wang | Lianyu Wang, Meng Wang, Daoqiang Zhang, Huazhu Fu | Model Barrier: A Compact Un-Transferable Isolation Domain for Model
Intellectual Property Protection | null | null | null | null | cs.AI cs.CY | http://creativecommons.org/licenses/by/4.0/ | As scientific and technological advancements result from human intellectual
labor and computational costs, protecting model intellectual property (IP) has
become increasingly important to encourage model creators and owners. Model IP
protection involves preventing the use of well-trained models on unauthorized
domains. To address this issue, we propose a novel approach called Compact
Un-Transferable Isolation Domain (CUTI-domain), which acts as a barrier to
block illegal transfers from authorized to unauthorized domains. Specifically,
CUTI-domain blocks cross-domain transfers by highlighting the private style
features of the authorized domain, leading to recognition failure on
unauthorized domains with irrelevant private style features. Moreover, we
provide two solutions for using CUTI-domain depending on whether the
unauthorized domain is known or not: target-specified CUTI-domain and
target-free CUTI-domain. Our comprehensive experimental results on four digit
datasets, CIFAR10 & STL10, and VisDA-2017 dataset demonstrate that CUTI-domain
can be easily implemented as a plug-and-play module with different backbones,
providing an efficient solution for model IP protection.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:07:11 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wang",
"Lianyu",
""
],
[
"Wang",
"Meng",
""
],
[
"Zhang",
"Daoqiang",
""
],
[
"Fu",
"Huazhu",
""
]
] | TITLE: Model Barrier: A Compact Un-Transferable Isolation Domain for Model
Intellectual Property Protection
ABSTRACT: As scientific and technological advancements result from human intellectual
labor and computational costs, protecting model intellectual property (IP) has
become increasingly important to encourage model creators and owners. Model IP
protection involves preventing the use of well-trained models on unauthorized
domains. To address this issue, we propose a novel approach called Compact
Un-Transferable Isolation Domain (CUTI-domain), which acts as a barrier to
block illegal transfers from authorized to unauthorized domains. Specifically,
CUTI-domain blocks cross-domain transfers by highlighting the private style
features of the authorized domain, leading to recognition failure on
unauthorized domains with irrelevant private style features. Moreover, we
provide two solutions for using CUTI-domain depending on whether the
unauthorized domain is known or not: target-specified CUTI-domain and
target-free CUTI-domain. Our comprehensive experimental results on four digit
datasets, CIFAR10 & STL10, and VisDA-2017 dataset demonstrate that CUTI-domain
can be easily implemented as a plug-and-play module with different backbones,
providing an efficient solution for model IP protection.
| no_new_dataset | 0.945399 |
2303.11079 | Vladimir Dvorkin | Vladimir Dvorkin and Audun Botterud | Differentially Private Algorithms for Synthetic Power System Datasets | null | null | null | null | cs.CR cs.DS cs.LG math.OC | http://creativecommons.org/licenses/by/4.0/ | While power systems research relies on the availability of real-world network
datasets, data owners (e.g., system operators) are hesitant to share data due
to security and privacy risks. To control these risks, we develop
privacy-preserving algorithms for the synthetic generation of optimization and
machine learning datasets. Taking a real-world dataset as input, the algorithms
output its noisy, synthetic version, which preserves the accuracy of the real
data on a specific downstream model or even a large population of those. We
control the privacy loss using Laplace and Exponential mechanisms of
differential privacy and preserve data accuracy using a post-processing convex
optimization. We apply the algorithms to generate synthetic network parameters
and wind power data.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:38:58 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Dvorkin",
"Vladimir",
""
],
[
"Botterud",
"Audun",
""
]
] | TITLE: Differentially Private Algorithms for Synthetic Power System Datasets
ABSTRACT: While power systems research relies on the availability of real-world network
datasets, data owners (e.g., system operators) are hesitant to share data due
to security and privacy risks. To control these risks, we develop
privacy-preserving algorithms for the synthetic generation of optimization and
machine learning datasets. Taking a real-world dataset as input, the algorithms
output its noisy, synthetic version, which preserves the accuracy of the real
data on a specific downstream model or even a large population of those. We
control the privacy loss using Laplace and Exponential mechanisms of
differential privacy and preserve data accuracy using a post-processing convex
optimization. We apply the algorithms to generate synthetic network parameters
and wind power data.
| no_new_dataset | 0.943452 |
2303.11082 | Simon Razniewski | Blerta Veseli, Sneha Singhania, Simon Razniewski, Gerhard Weikum | Evaluating Language Models for Knowledge Base Completion | Data and code available at https://github.com/bveseli/LMsForKBC | ESWC 2023 | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Structured knowledge bases (KBs) are a foundation of many intelligent
applications, yet are notoriously incomplete. Language models (LMs) have
recently been proposed for unsupervised knowledge base completion (KBC), yet,
despite encouraging initial results, questions regarding their suitability
remain open. Existing evaluations often fall short because they only evaluate
on popular subjects, or sample already existing facts from KBs. In this work,
we introduce a novel, more challenging benchmark dataset, and a methodology
tailored for a realistic assessment of the KBC potential of LMs. For automated
assessment, we curate a dataset called WD-KNOWN, which provides an unbiased
random sample of Wikidata, containing over 3.9 million facts. In a second step,
we perform a human evaluation on predictions that are not yet in the KB, as
only this provides real insights into the added value over existing KBs. Our
key finding is that biases in dataset conception of previous benchmarks lead to
a systematic overestimate of LM performance for KBC. However, our results also
reveal strong areas of LMs. We could, for example, perform a significant
completion of Wikidata on the relations nativeLanguage, by a factor of ~21
(from 260k to 5.8M) at 82% precision, usedLanguage, by a factor of ~2.1 (from
2.1M to 6.6M) at 82% precision, and citizenOf by a factor of ~0.3 (from 4.2M to
5.3M) at 90% precision. Moreover, we find that LMs possess surprisingly strong
generalization capabilities: even on relations where most facts were not
directly observed in LM training, prediction quality can be high.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:14:59 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Veseli",
"Blerta",
""
],
[
"Singhania",
"Sneha",
""
],
[
"Razniewski",
"Simon",
""
],
[
"Weikum",
"Gerhard",
""
]
] | TITLE: Evaluating Language Models for Knowledge Base Completion
ABSTRACT: Structured knowledge bases (KBs) are a foundation of many intelligent
applications, yet are notoriously incomplete. Language models (LMs) have
recently been proposed for unsupervised knowledge base completion (KBC), yet,
despite encouraging initial results, questions regarding their suitability
remain open. Existing evaluations often fall short because they only evaluate
on popular subjects, or sample already existing facts from KBs. In this work,
we introduce a novel, more challenging benchmark dataset, and a methodology
tailored for a realistic assessment of the KBC potential of LMs. For automated
assessment, we curate a dataset called WD-KNOWN, which provides an unbiased
random sample of Wikidata, containing over 3.9 million facts. In a second step,
we perform a human evaluation on predictions that are not yet in the KB, as
only this provides real insights into the added value over existing KBs. Our
key finding is that biases in dataset conception of previous benchmarks lead to
a systematic overestimate of LM performance for KBC. However, our results also
reveal strong areas of LMs. We could, for example, perform a significant
completion of Wikidata on the relations nativeLanguage, by a factor of ~21
(from 260k to 5.8M) at 82% precision, usedLanguage, by a factor of ~2.1 (from
2.1M to 6.6M) at 82% precision, and citizenOf by a factor of ~0.3 (from 4.2M to
5.3M) at 90% precision. Moreover, we find that LMs possess surprisingly strong
generalization capabilities: even on relations where most facts were not
directly observed in LM training, prediction quality can be high.
| new_dataset | 0.970688 |
2303.11090 | Guoliang Wang | Guoliang Wang, Yanlei Shang and Yong Chen | Scene Graph Based Fusion Network For Image-Text Retrieval | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A critical challenge to image-text retrieval is how to learn accurate
correspondences between images and texts. Most existing methods mainly focus on
coarse-grained correspondences based on co-occurrences of semantic objects,
while failing to distinguish the fine-grained local correspondences. In this
paper, we propose a novel Scene Graph based Fusion Network (dubbed SGFN), which
enhances the images'/texts' features through intra- and cross-modal fusion for
image-text retrieval. To be specific, we design an intra-modal hierarchical
attention fusion to incorporate semantic contexts, such as objects, attributes,
and relationships, into images'/texts' feature vectors via scene graphs, and a
cross-modal attention fusion to combine the contextual semantics and local
fusion via contextual vectors. Extensive experiments on public datasets
Flickr30K and MSCOCO show that our SGFN performs better than quite a few SOTA
image-text retrieval methods.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:22:56 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Wang",
"Guoliang",
""
],
[
"Shang",
"Yanlei",
""
],
[
"Chen",
"Yong",
""
]
] | TITLE: Scene Graph Based Fusion Network For Image-Text Retrieval
ABSTRACT: A critical challenge to image-text retrieval is how to learn accurate
correspondences between images and texts. Most existing methods mainly focus on
coarse-grained correspondences based on co-occurrences of semantic objects,
while failing to distinguish the fine-grained local correspondences. In this
paper, we propose a novel Scene Graph based Fusion Network (dubbed SGFN), which
enhances the images'/texts' features through intra- and cross-modal fusion for
image-text retrieval. To be specific, we design an intra-modal hierarchical
attention fusion to incorporate semantic contexts, such as objects, attributes,
and relationships, into images'/texts' feature vectors via scene graphs, and a
cross-modal attention fusion to combine the contextual semantics and local
fusion via contextual vectors. Extensive experiments on public datasets
Flickr30K and MSCOCO show that our SGFN performs better than quite a few SOTA
image-text retrieval methods.
| no_new_dataset | 0.949716 |
2303.11100 | Duan Furong | Furong Duan, Tao Zhu, Jinqiang Wang, Liming Chen, Huansheng Ning,
Yaping Wan | A Multi-Task Deep Learning Approach for Sensor-based Human Activity
Recognition and Segmentation | 14 pages, 14 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sensor-based human activity segmentation and recognition are two important
and challenging problems in many real-world applications and they have drawn
increasing attention from the deep learning community in recent years. Most of
the existing deep learning works were designed based on pre-segmented sensor
streams and they have treated activity segmentation and recognition as two
separate tasks. In practice, performing data stream segmentation is very
challenging. We believe that both activity segmentation and recognition may
convey unique information which can complement each other to improve the
performance of the two tasks. In this paper, we firstly proposes a new
multitask deep neural network to solve the two tasks simultaneously. The
proposed neural network adopts selective convolution and features multiscale
windows to segment activities of long or short time durations. First, multiple
windows of different scales are generated to center on each unit of the feature
sequence. Then, the model is trained to predict, for each window, the activity
class and the offset to the true activity boundaries. Finally, overlapping
windows are filtered out by non-maximum suppression, and adjacent windows of
the same activity are concatenated to complete the segmentation task. Extensive
experiments were conducted on eight popular benchmarking datasets, and the
results show that our proposed method outperforms the state-of-the-art methods
both for activity recognition and segmentation.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 13:34:28 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Duan",
"Furong",
""
],
[
"Zhu",
"Tao",
""
],
[
"Wang",
"Jinqiang",
""
],
[
"Chen",
"Liming",
""
],
[
"Ning",
"Huansheng",
""
],
[
"Wan",
"Yaping",
""
]
] | TITLE: A Multi-Task Deep Learning Approach for Sensor-based Human Activity
Recognition and Segmentation
ABSTRACT: Sensor-based human activity segmentation and recognition are two important
and challenging problems in many real-world applications and they have drawn
increasing attention from the deep learning community in recent years. Most of
the existing deep learning works were designed based on pre-segmented sensor
streams and they have treated activity segmentation and recognition as two
separate tasks. In practice, performing data stream segmentation is very
challenging. We believe that both activity segmentation and recognition may
convey unique information which can complement each other to improve the
performance of the two tasks. In this paper, we firstly proposes a new
multitask deep neural network to solve the two tasks simultaneously. The
proposed neural network adopts selective convolution and features multiscale
windows to segment activities of long or short time durations. First, multiple
windows of different scales are generated to center on each unit of the feature
sequence. Then, the model is trained to predict, for each window, the activity
class and the offset to the true activity boundaries. Finally, overlapping
windows are filtered out by non-maximum suppression, and adjacent windows of
the same activity are concatenated to complete the segmentation task. Extensive
experiments were conducted on eight popular benchmarking datasets, and the
results show that our proposed method outperforms the state-of-the-art methods
both for activity recognition and segmentation.
| no_new_dataset | 0.946843 |
2303.11120 | Gianluca Scarpellini | Francesco Giuliari, Gianluca Scarpellini, Stuart James, Yiming Wang,
Alessio Del Bue | Positional Diffusion: Ordering Unordered Sets with Diffusion
Probabilistic Models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Positional reasoning is the process of ordering unsorted parts contained in a
set into a consistent structure. We present Positional Diffusion, a
plug-and-play graph formulation with Diffusion Probabilistic Models to address
positional reasoning. We use the forward process to map elements' positions in
a set to random positions in a continuous space. Positional Diffusion learns to
reverse the noising process and recover the original positions through an
Attention-based Graph Neural Network. We conduct extensive experiments with
benchmark datasets including two puzzle datasets, three sentence ordering
datasets, and one visual storytelling dataset, demonstrating that our method
outperforms long-lasting research on puzzle solving with up to +18% compared to
the second-best deep learning method, and performs on par against the
state-of-the-art methods on sentence ordering and visual storytelling. Our work
highlights the suitability of diffusion models for ordering problems and
proposes a novel formulation and method for solving various ordering tasks.
Project website at https://iit-pavis.github.io/Positional_Diffusion/
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 14:01:01 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Giuliari",
"Francesco",
""
],
[
"Scarpellini",
"Gianluca",
""
],
[
"James",
"Stuart",
""
],
[
"Wang",
"Yiming",
""
],
[
"Del Bue",
"Alessio",
""
]
] | TITLE: Positional Diffusion: Ordering Unordered Sets with Diffusion
Probabilistic Models
ABSTRACT: Positional reasoning is the process of ordering unsorted parts contained in a
set into a consistent structure. We present Positional Diffusion, a
plug-and-play graph formulation with Diffusion Probabilistic Models to address
positional reasoning. We use the forward process to map elements' positions in
a set to random positions in a continuous space. Positional Diffusion learns to
reverse the noising process and recover the original positions through an
Attention-based Graph Neural Network. We conduct extensive experiments with
benchmark datasets including two puzzle datasets, three sentence ordering
datasets, and one visual storytelling dataset, demonstrating that our method
outperforms long-lasting research on puzzle solving with up to +18% compared to
the second-best deep learning method, and performs on par against the
state-of-the-art methods on sentence ordering and visual storytelling. Our work
highlights the suitability of diffusion models for ordering problems and
proposes a novel formulation and method for solving various ordering tasks.
Project website at https://iit-pavis.github.io/Positional_Diffusion/
| new_dataset | 0.960137 |
2303.11135 | Ziquan Liu | Ziquan Liu, Yi Xu, Xiangyang Ji, Antoni B. Chan | TWINS: A Fine-Tuning Framework for Improved Transferability of
Adversarial Robustness and Generalization | CVPR2023 | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent years have seen the ever-increasing importance of pre-trained models
and their downstream training in deep learning research and applications. At
the same time, the defense for adversarial examples has been mainly
investigated in the context of training from random initialization on simple
classification tasks. To better exploit the potential of pre-trained models in
adversarial robustness, this paper focuses on the fine-tuning of an
adversarially pre-trained model in various classification tasks. Existing
research has shown that since the robust pre-trained model has already learned
a robust feature extractor, the crucial question is how to maintain the
robustness in the pre-trained model when learning the downstream task. We study
the model-based and data-based approaches for this goal and find that the two
common approaches cannot achieve the objective of improving both generalization
and adversarial robustness. Thus, we propose a novel statistics-based approach,
Two-WIng NormliSation (TWINS) fine-tuning framework, which consists of two
neural networks where one of them keeps the population means and variances of
pre-training data in the batch normalization layers. Besides the robust
information transfer, TWINS increases the effective learning rate without
hurting the training stability since the relationship between a weight norm and
its gradient norm in standard batch normalization layer is broken, resulting in
a faster escape from the sub-optimal initialization and alleviating the robust
overfitting. Finally, TWINS is shown to be effective on a wide range of image
classification datasets in terms of both generalization and robustness. Our
code is available at https://github.com/ziquanliu/CVPR2023-TWINS.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 14:12:55 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Liu",
"Ziquan",
""
],
[
"Xu",
"Yi",
""
],
[
"Ji",
"Xiangyang",
""
],
[
"Chan",
"Antoni B.",
""
]
] | TITLE: TWINS: A Fine-Tuning Framework for Improved Transferability of
Adversarial Robustness and Generalization
ABSTRACT: Recent years have seen the ever-increasing importance of pre-trained models
and their downstream training in deep learning research and applications. At
the same time, the defense for adversarial examples has been mainly
investigated in the context of training from random initialization on simple
classification tasks. To better exploit the potential of pre-trained models in
adversarial robustness, this paper focuses on the fine-tuning of an
adversarially pre-trained model in various classification tasks. Existing
research has shown that since the robust pre-trained model has already learned
a robust feature extractor, the crucial question is how to maintain the
robustness in the pre-trained model when learning the downstream task. We study
the model-based and data-based approaches for this goal and find that the two
common approaches cannot achieve the objective of improving both generalization
and adversarial robustness. Thus, we propose a novel statistics-based approach,
Two-WIng NormliSation (TWINS) fine-tuning framework, which consists of two
neural networks where one of them keeps the population means and variances of
pre-training data in the batch normalization layers. Besides the robust
information transfer, TWINS increases the effective learning rate without
hurting the training stability since the relationship between a weight norm and
its gradient norm in standard batch normalization layer is broken, resulting in
a faster escape from the sub-optimal initialization and alleviating the robust
overfitting. Finally, TWINS is shown to be effective on a wide range of image
classification datasets in terms of both generalization and robustness. Our
code is available at https://github.com/ziquanliu/CVPR2023-TWINS.
| no_new_dataset | 0.946151 |
2303.11137 | Yu Cao | Yu Cao, Xiangqiao Meng, P.Y. Mok, Xueting Liu, Tong-Yee Lee, Ping Li | AnimeDiffusion: Anime Face Line Drawing Colorization via Diffusion
Models | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It is a time-consuming and tedious work for manually colorizing anime line
drawing images, which is an essential stage in cartoon animation creation
pipeline. Reference-based line drawing colorization is a challenging task that
relies on the precise cross-domain long-range dependency modelling between the
line drawing and reference image. Existing learning methods still utilize
generative adversarial networks (GANs) as one key module of their model
architecture. In this paper, we propose a novel method called AnimeDiffusion
using diffusion models that performs anime face line drawing colorization
automatically. To the best of our knowledge, this is the first diffusion model
tailored for anime content creation. In order to solve the huge training
consumption problem of diffusion models, we design a hybrid training strategy,
first pre-training a diffusion model with classifier-free guidance and then
fine-tuning it with image reconstruction guidance. We find that with a few
iterations of fine-tuning, the model shows wonderful colorization performance,
as illustrated in Fig. 1. For training AnimeDiffusion, we conduct an anime face
line drawing colorization benchmark dataset, which contains 31696 training data
and 579 testing data. We hope this dataset can fill the gap of no available
high resolution anime face dataset for colorization method evaluation. Through
multiple quantitative metrics evaluated on our dataset and a user study, we
demonstrate AnimeDiffusion outperforms state-of-the-art GANs-based models for
anime face line drawing colorization. We also collaborate with professional
artists to test and apply our AnimeDiffusion for their creation work. We
release our code on https://github.com/xq-meng/AnimeDiffusion.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 14:15:23 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Cao",
"Yu",
""
],
[
"Meng",
"Xiangqiao",
""
],
[
"Mok",
"P. Y.",
""
],
[
"Liu",
"Xueting",
""
],
[
"Lee",
"Tong-Yee",
""
],
[
"Li",
"Ping",
""
]
] | TITLE: AnimeDiffusion: Anime Face Line Drawing Colorization via Diffusion
Models
ABSTRACT: It is a time-consuming and tedious work for manually colorizing anime line
drawing images, which is an essential stage in cartoon animation creation
pipeline. Reference-based line drawing colorization is a challenging task that
relies on the precise cross-domain long-range dependency modelling between the
line drawing and reference image. Existing learning methods still utilize
generative adversarial networks (GANs) as one key module of their model
architecture. In this paper, we propose a novel method called AnimeDiffusion
using diffusion models that performs anime face line drawing colorization
automatically. To the best of our knowledge, this is the first diffusion model
tailored for anime content creation. In order to solve the huge training
consumption problem of diffusion models, we design a hybrid training strategy,
first pre-training a diffusion model with classifier-free guidance and then
fine-tuning it with image reconstruction guidance. We find that with a few
iterations of fine-tuning, the model shows wonderful colorization performance,
as illustrated in Fig. 1. For training AnimeDiffusion, we conduct an anime face
line drawing colorization benchmark dataset, which contains 31696 training data
and 579 testing data. We hope this dataset can fill the gap of no available
high resolution anime face dataset for colorization method evaluation. Through
multiple quantitative metrics evaluated on our dataset and a user study, we
demonstrate AnimeDiffusion outperforms state-of-the-art GANs-based models for
anime face line drawing colorization. We also collaborate with professional
artists to test and apply our AnimeDiffusion for their creation work. We
release our code on https://github.com/xq-meng/AnimeDiffusion.
| new_dataset | 0.970576 |
2303.11169 | Ming Li | Ming Li, Xinming Huang and Ziming Zhang | Self-supervised Geometric Features Discovery via Interpretable Attentio
for Vehicle Re-Identification and Beyond (Complete Version) | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To learn distinguishable patterns, most of recent works in vehicle
re-identification (ReID) struggled to redevelop official benchmarks to provide
various supervisions, which requires prohibitive human labors. In this paper,
we seek to achieve the similar goal but do not involve more human efforts. To
this end, we introduce a novel framework, which successfully encodes both
geometric local features and global representations to distinguish vehicle
instances, optimized only by the supervision from official ID labels.
Specifically, given our insight that objects in ReID share similar geometric
characteristics, we propose to borrow self-supervised representation learning
to facilitate geometric features discovery. To condense these features, we
introduce an interpretable attention module, with the core of local maxima
aggregation instead of fully automatic learning, whose mechanism is completely
understandable and whose response map is physically reasonable. To the best of
our knowledge, we are the first that perform self-supervised learning to
discover geometric features. We conduct comprehensive experiments on three most
popular datasets for vehicle ReID, i.e., VeRi-776, CityFlow-ReID, and
VehicleID. We report our state-of-the-art (SOTA) performances and promising
visualization results. We also show the excellent scalability of our approach
on other ReID related tasks, i.e., person ReID and multi-target multi-camera
(MTMC) vehicle tracking. The code is available at https://github.com/
ming1993li/Self-supervised-Geometric.
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2023 08:21:00 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Li",
"Ming",
""
],
[
"Huang",
"Xinming",
""
],
[
"Zhang",
"Ziming",
""
]
] | TITLE: Self-supervised Geometric Features Discovery via Interpretable Attentio
for Vehicle Re-Identification and Beyond (Complete Version)
ABSTRACT: To learn distinguishable patterns, most of recent works in vehicle
re-identification (ReID) struggled to redevelop official benchmarks to provide
various supervisions, which requires prohibitive human labors. In this paper,
we seek to achieve the similar goal but do not involve more human efforts. To
this end, we introduce a novel framework, which successfully encodes both
geometric local features and global representations to distinguish vehicle
instances, optimized only by the supervision from official ID labels.
Specifically, given our insight that objects in ReID share similar geometric
characteristics, we propose to borrow self-supervised representation learning
to facilitate geometric features discovery. To condense these features, we
introduce an interpretable attention module, with the core of local maxima
aggregation instead of fully automatic learning, whose mechanism is completely
understandable and whose response map is physically reasonable. To the best of
our knowledge, we are the first that perform self-supervised learning to
discover geometric features. We conduct comprehensive experiments on three most
popular datasets for vehicle ReID, i.e., VeRi-776, CityFlow-ReID, and
VehicleID. We report our state-of-the-art (SOTA) performances and promising
visualization results. We also show the excellent scalability of our approach
on other ReID related tasks, i.e., person ReID and multi-target multi-camera
(MTMC) vehicle tracking. The code is available at https://github.com/
ming1993li/Self-supervised-Geometric.
| no_new_dataset | 0.946448 |
2303.11172 | Samin Poudel | Samin Poudel and Marwan Bikdash | Explaining the Performance of Collaborative Filtering Methods With
Optimal Data Characteristics | null | International Journal on Cybernetics & Informatics (IJCI) Vol. 12,
No.2, April 2023 | 10.5121/ijci.2023.120221 | null | cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | The performance of a Collaborative Filtering (CF) method is based on the
properties of a User-Item Rating Matrix (URM). And the properties or Rating
Data Characteristics (RDC) of a URM are constantly changing. Recent studies
significantly explained the variation in the performances of CF methods
resulted due to the change in URM using six or more RDC. Here, we found that
the significant proportion of variation in the performances of different CF
techniques can be accounted to two RDC only. The two RDC are the number of
ratings per user or Information per User (IpU) and the number of ratings per
item or Information per Item (IpI). And the performances of CF algorithms are
quadratic to IpU (or IpI) for a square URM. The findings of this study are
based on seven well-established CF methods and three popular public recommender
datasets: 1M MovieLens, 25M MovieLens, and Yahoo! Music Rating datasets
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 02:34:05 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Poudel",
"Samin",
""
],
[
"Bikdash",
"Marwan",
""
]
] | TITLE: Explaining the Performance of Collaborative Filtering Methods With
Optimal Data Characteristics
ABSTRACT: The performance of a Collaborative Filtering (CF) method is based on the
properties of a User-Item Rating Matrix (URM). And the properties or Rating
Data Characteristics (RDC) of a URM are constantly changing. Recent studies
significantly explained the variation in the performances of CF methods
resulted due to the change in URM using six or more RDC. Here, we found that
the significant proportion of variation in the performances of different CF
techniques can be accounted to two RDC only. The two RDC are the number of
ratings per user or Information per User (IpU) and the number of ratings per
item or Information per Item (IpI). And the performances of CF algorithms are
quadratic to IpU (or IpI) for a square URM. The findings of this study are
based on seven well-established CF methods and three popular public recommender
datasets: 1M MovieLens, 25M MovieLens, and Yahoo! Music Rating datasets
| no_new_dataset | 0.9434 |
2303.11175 | Hadi Mansourifar | Hadi Mansourifar and Steven J. Simske | Towards CGAN-based Satellite Image Synthesis with Partial Pixel-Wise
Annotation | null | null | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | Conditional Generative Adversarial Nets (CGANs) need a significantly huge
dataset with a detailed pixel-wise annotation to generate high-quality images.
Unfortunately, any amount of missing pixel annotations may significantly impact
the result not only locally, but also in annotated areas. To the best of our
knowledge, such a challenge has never been investigated in the broader field of
GANs. In this paper, we take the first step in this direction to study the
problem of CGAN-based satellite image synthesis given partially annotated
images. We first define the problem of image synthesis using partially
annotated data, and we discuss a scenario in which we face such a challenge. We
then propose an effective solution called detail augmentation to address this
problem. To do so, we tested two different approaches to augment details to
compensate for missing pixel-wise annotations. In the first approach, we
augmented the original images with their Canny edges to using the CGAN to
compensate for the missing annotations. The second approach, however, attempted
to assign a color to all pixels with missing annotation. Eventually, a
different CGAN was trained to translate the new feature images into a final
output.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2023 21:53:33 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Mansourifar",
"Hadi",
""
],
[
"Simske",
"Steven J.",
""
]
] | TITLE: Towards CGAN-based Satellite Image Synthesis with Partial Pixel-Wise
Annotation
ABSTRACT: Conditional Generative Adversarial Nets (CGANs) need a significantly huge
dataset with a detailed pixel-wise annotation to generate high-quality images.
Unfortunately, any amount of missing pixel annotations may significantly impact
the result not only locally, but also in annotated areas. To the best of our
knowledge, such a challenge has never been investigated in the broader field of
GANs. In this paper, we take the first step in this direction to study the
problem of CGAN-based satellite image synthesis given partially annotated
images. We first define the problem of image synthesis using partially
annotated data, and we discuss a scenario in which we face such a challenge. We
then propose an effective solution called detail augmentation to address this
problem. To do so, we tested two different approaches to augment details to
compensate for missing pixel-wise annotations. In the first approach, we
augmented the original images with their Canny edges to using the CGAN to
compensate for the missing annotations. The second approach, however, attempted
to assign a color to all pixels with missing annotation. Eventually, a
different CGAN was trained to translate the new feature images into a final
output.
| no_new_dataset | 0.947039 |
2303.11181 | Naomi A. Arnold | Haaroon Yousaf, Naomi A. Arnold, Renaud Lambiotte, Timothy LaRock,
Richard G. Clegg, Peijie Zhong, Alhamza Alnaimi and Ben Steer | Non-Markovian paths and cycles in NFT trades | Abstract accepted as a workshop presentation at Complex Data Blocks,
a satellite of the 2022 Conference of Complex Systems | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent years have witnessed the availability of richer and richer datasets in
a variety of domains, where signals often have a multi-modal nature, blending
temporal, relational and semantic information. Within this context, several
works have shown that standard network models are sometimes not sufficient to
properly capture the complexity of real-world interacting systems. For this
reason, different attempts have been made to enrich the network language,
leading to the emerging field of higher-order networks. In this work, we
investigate the possibility of applying methods from higher-order networks to
extract information from the online trade of Non-fungible tokens (NFTs),
leveraging on their intrinsic temporal and non-Markovian nature. While NFTs as
a technology open up the realms for many exciting applications, its future is
marred by challenges of proof of ownership, scams, wash trading and possible
money laundering. We demonstrate that by investigating time-respecting
non-Markovian paths exhibited by NFT trades, we provide a practical path-based
approach to fraud detection.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 15:08:38 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Yousaf",
"Haaroon",
""
],
[
"Arnold",
"Naomi A.",
""
],
[
"Lambiotte",
"Renaud",
""
],
[
"LaRock",
"Timothy",
""
],
[
"Clegg",
"Richard G.",
""
],
[
"Zhong",
"Peijie",
""
],
[
"Alnaimi",
"Alhamza",
""
],
[
"Steer",
"Ben",
""
]
] | TITLE: Non-Markovian paths and cycles in NFT trades
ABSTRACT: Recent years have witnessed the availability of richer and richer datasets in
a variety of domains, where signals often have a multi-modal nature, blending
temporal, relational and semantic information. Within this context, several
works have shown that standard network models are sometimes not sufficient to
properly capture the complexity of real-world interacting systems. For this
reason, different attempts have been made to enrich the network language,
leading to the emerging field of higher-order networks. In this work, we
investigate the possibility of applying methods from higher-order networks to
extract information from the online trade of Non-fungible tokens (NFTs),
leveraging on their intrinsic temporal and non-Markovian nature. While NFTs as
a technology open up the realms for many exciting applications, its future is
marred by challenges of proof of ownership, scams, wash trading and possible
money laundering. We demonstrate that by investigating time-respecting
non-Markovian paths exhibited by NFT trades, we provide a practical path-based
approach to fraud detection.
| no_new_dataset | 0.947332 |
2303.11184 | Jiaqing Yuan | Jiaqing Yuan and Munindar P. Singh | Conversation Modeling to Predict Derailment | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Conversations among online users sometimes derail, i.e., break down into
personal attacks. Such derailment has a negative impact on the healthy growth
of cyberspace communities. The ability to predict whether ongoing conversations
are likely to derail could provide valuable real-time insight to interlocutors
and moderators. Prior approaches predict conversation derailment
retrospectively without the ability to forestall the derailment proactively.
Some works attempt to make dynamic prediction as the conversation develops, but
fail to incorporate multisource information, such as conversation structure and
distance to derailment.
We propose a hierarchical transformer-based framework that combines
utterance-level and conversation-level information to capture fine-grained
contextual semantics. We propose a domain-adaptive pretraining objective to
integrate conversational structure information and a multitask learning scheme
to leverage the distance from each utterance to derailment. An evaluation of
our framework on two conversation derailment datasets yields improvement over
F1 score for the prediction of derailment. These results demonstrate the
effectiveness of incorporating multisource information.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 15:10:45 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Yuan",
"Jiaqing",
""
],
[
"Singh",
"Munindar P.",
""
]
] | TITLE: Conversation Modeling to Predict Derailment
ABSTRACT: Conversations among online users sometimes derail, i.e., break down into
personal attacks. Such derailment has a negative impact on the healthy growth
of cyberspace communities. The ability to predict whether ongoing conversations
are likely to derail could provide valuable real-time insight to interlocutors
and moderators. Prior approaches predict conversation derailment
retrospectively without the ability to forestall the derailment proactively.
Some works attempt to make dynamic prediction as the conversation develops, but
fail to incorporate multisource information, such as conversation structure and
distance to derailment.
We propose a hierarchical transformer-based framework that combines
utterance-level and conversation-level information to capture fine-grained
contextual semantics. We propose a domain-adaptive pretraining objective to
integrate conversational structure information and a multitask learning scheme
to leverage the distance from each utterance to derailment. An evaluation of
our framework on two conversation derailment datasets yields improvement over
F1 score for the prediction of derailment. These results demonstrate the
effectiveness of incorporating multisource information.
| no_new_dataset | 0.946349 |
2303.11187 | Siyu Chen | Siyu Chen, Yitan Wang, Zhaoran Wang, Zhuoran Yang | A Unified Framework of Policy Learning for Contextual Bandit with
Confounding Bias and Missing Observations | 76 page, 5 figures | null | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by/4.0/ | We study the offline contextual bandit problem, where we aim to acquire an
optimal policy using observational data. However, this data usually contains
two deficiencies: (i) some variables that confound actions are not observed,
and (ii) missing observations exist in the collected data. Unobserved
confounders lead to a confounding bias and missing observations cause bias and
inefficiency problems. To overcome these challenges and learn the optimal
policy from the observed dataset, we present a new algorithm called
Causal-Adjusted Pessimistic (CAP) policy learning, which forms the reward
function as the solution of an integral equation system, builds a confidence
set, and greedily takes action with pessimism. With mild assumptions on the
data, we develop an upper bound to the suboptimality of CAP for the offline
contextual bandit problem.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 15:17:31 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Chen",
"Siyu",
""
],
[
"Wang",
"Yitan",
""
],
[
"Wang",
"Zhaoran",
""
],
[
"Yang",
"Zhuoran",
""
]
] | TITLE: A Unified Framework of Policy Learning for Contextual Bandit with
Confounding Bias and Missing Observations
ABSTRACT: We study the offline contextual bandit problem, where we aim to acquire an
optimal policy using observational data. However, this data usually contains
two deficiencies: (i) some variables that confound actions are not observed,
and (ii) missing observations exist in the collected data. Unobserved
confounders lead to a confounding bias and missing observations cause bias and
inefficiency problems. To overcome these challenges and learn the optimal
policy from the observed dataset, we present a new algorithm called
Causal-Adjusted Pessimistic (CAP) policy learning, which forms the reward
function as the solution of an integral equation system, builds a confidence
set, and greedily takes action with pessimism. With mild assumptions on the
data, we develop an upper bound to the suboptimality of CAP for the offline
contextual bandit problem.
| no_new_dataset | 0.947137 |
2303.11191 | Andr\'e Correia | Andr\'e Correia and Lu\'is A. Alexandre | A Survey of Demonstration Learning | 35 pages, 9 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | With the fast improvement of machine learning, reinforcement learning (RL)
has been used to automate human tasks in different areas. However, training
such agents is difficult and restricted to expert users. Moreover, it is mostly
limited to simulation environments due to the high cost and safety concerns of
interactions in the real world. Demonstration Learning is a paradigm in which
an agent learns to perform a task by imitating the behavior of an expert shown
in demonstrations. It is a relatively recent area in machine learning, but it
is gaining significant traction due to having tremendous potential for learning
complex behaviors from demonstrations. Learning from demonstration accelerates
the learning process by improving sample efficiency, while also reducing the
effort of the programmer. Due to learning without interacting with the
environment, demonstration learning would allow the automation of a wide range
of real world applications such as robotics and healthcare. This paper provides
a survey of demonstration learning, where we formally introduce the
demonstration problem along with its main challenges and provide a
comprehensive overview of the process of learning from demonstrations from the
creation of the demonstration data set, to learning methods from
demonstrations, and optimization by combining demonstration learning with
different machine learning methods. We also review the existing benchmarks and
identify their strengths and limitations. Additionally, we discuss the
advantages and disadvantages of the paradigm as well as its main applications.
Lastly, we discuss our perspective on open problems and research directions for
this rapidly growing field.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 15:22:10 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Correia",
"André",
""
],
[
"Alexandre",
"Luís A.",
""
]
] | TITLE: A Survey of Demonstration Learning
ABSTRACT: With the fast improvement of machine learning, reinforcement learning (RL)
has been used to automate human tasks in different areas. However, training
such agents is difficult and restricted to expert users. Moreover, it is mostly
limited to simulation environments due to the high cost and safety concerns of
interactions in the real world. Demonstration Learning is a paradigm in which
an agent learns to perform a task by imitating the behavior of an expert shown
in demonstrations. It is a relatively recent area in machine learning, but it
is gaining significant traction due to having tremendous potential for learning
complex behaviors from demonstrations. Learning from demonstration accelerates
the learning process by improving sample efficiency, while also reducing the
effort of the programmer. Due to learning without interacting with the
environment, demonstration learning would allow the automation of a wide range
of real world applications such as robotics and healthcare. This paper provides
a survey of demonstration learning, where we formally introduce the
demonstration problem along with its main challenges and provide a
comprehensive overview of the process of learning from demonstrations from the
creation of the demonstration data set, to learning methods from
demonstrations, and optimization by combining demonstration learning with
different machine learning methods. We also review the existing benchmarks and
identify their strengths and limitations. Additionally, we discuss the
advantages and disadvantages of the paradigm as well as its main applications.
Lastly, we discuss our perspective on open problems and research directions for
this rapidly growing field.
| no_new_dataset | 0.944485 |
2303.11224 | Tobias Weber | Tobias Weber, Michael Ingrisch, Bernd Bischl, David R\"ugamer | Cascaded Latent Diffusion Models for High-Resolution Chest X-ray
Synthesis | accepted at PAKDD 2023 | null | null | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | While recent advances in large-scale foundational models show promising
results, their application to the medical domain has not yet been explored in
detail. In this paper, we progress into the realms of large-scale modeling in
medical synthesis by proposing Cheff - a foundational cascaded latent diffusion
model, which generates highly-realistic chest radiographs providing
state-of-the-art quality on a 1-megapixel scale. We further propose MaCheX,
which is a unified interface for public chest datasets and forms the largest
open collection of chest X-rays up to date. With Cheff conditioned on
radiological reports, we further guide the synthesis process over text prompts
and unveil the research area of report-to-chest-X-ray generation.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 16:00:20 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Weber",
"Tobias",
""
],
[
"Ingrisch",
"Michael",
""
],
[
"Bischl",
"Bernd",
""
],
[
"Rügamer",
"David",
""
]
] | TITLE: Cascaded Latent Diffusion Models for High-Resolution Chest X-ray
Synthesis
ABSTRACT: While recent advances in large-scale foundational models show promising
results, their application to the medical domain has not yet been explored in
detail. In this paper, we progress into the realms of large-scale modeling in
medical synthesis by proposing Cheff - a foundational cascaded latent diffusion
model, which generates highly-realistic chest radiographs providing
state-of-the-art quality on a 1-megapixel scale. We further propose MaCheX,
which is a unified interface for public chest datasets and forms the largest
open collection of chest X-rays up to date. With Cheff conditioned on
radiological reports, we further guide the synthesis process over text prompts
and unveil the research area of report-to-chest-X-ray generation.
| no_new_dataset | 0.948058 |
2303.11235 | Jawad Tayyub | Tejaswini Medi, Jawad Tayyub, Muhammad Sarmad, Frank Lindseth and
Margret Keuper | FullFormer: Generating Shapes Inside Shapes | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Implicit generative models have been widely employed to model 3D data and
have recently proven to be successful in encoding and generating high-quality
3D shapes. This work builds upon these models and alleviates current
limitations by presenting the first implicit generative model that facilitates
the generation of complex 3D shapes with rich internal geometric details. To
achieve this, our model uses unsigned distance fields to represent nested 3D
surfaces allowing learning from non-watertight mesh data. We propose a
transformer-based autoregressive model for 3D shape generation that leverages
context-rich tokens from vector quantized shape embeddings. The generated
tokens are decoded into an unsigned distance field which is rendered into a
novel 3D shape exhibiting a rich internal structure. We demonstrate that our
model achieves state-of-the-art point cloud generation results on popular
classes of 'Cars', 'Planes', and 'Chairs' of the ShapeNet dataset.
Additionally, we curate a dataset that exclusively comprises shapes with
realistic internal details from the `Cars' class of ShapeNet and demonstrate
our method's efficacy in generating these shapes with internal geometry.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 16:19:23 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Medi",
"Tejaswini",
""
],
[
"Tayyub",
"Jawad",
""
],
[
"Sarmad",
"Muhammad",
""
],
[
"Lindseth",
"Frank",
""
],
[
"Keuper",
"Margret",
""
]
] | TITLE: FullFormer: Generating Shapes Inside Shapes
ABSTRACT: Implicit generative models have been widely employed to model 3D data and
have recently proven to be successful in encoding and generating high-quality
3D shapes. This work builds upon these models and alleviates current
limitations by presenting the first implicit generative model that facilitates
the generation of complex 3D shapes with rich internal geometric details. To
achieve this, our model uses unsigned distance fields to represent nested 3D
surfaces allowing learning from non-watertight mesh data. We propose a
transformer-based autoregressive model for 3D shape generation that leverages
context-rich tokens from vector quantized shape embeddings. The generated
tokens are decoded into an unsigned distance field which is rendered into a
novel 3D shape exhibiting a rich internal structure. We demonstrate that our
model achieves state-of-the-art point cloud generation results on popular
classes of 'Cars', 'Planes', and 'Chairs' of the ShapeNet dataset.
Additionally, we curate a dataset that exclusively comprises shapes with
realistic internal details from the `Cars' class of ShapeNet and demonstrate
our method's efficacy in generating these shapes with internal geometry.
| new_dataset | 0.963746 |
2303.11269 | Robert Dyer | Ali M. Keshk and Robert Dyer | Method Chaining Redux: An Empirical Study of Method Chaining in Java,
Kotlin, and Python | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by-nc-sa/4.0/ | There are possible benefits and drawbacks to chaining methods together, as is
often done in fluent APIs. A prior study investigated how Java developers chain
methods in over 2.7k open-source projects. That study observed, for the dataset
analyzed, that the use of method chaining in Java is popular and seems to be
increasing over time. That study however was limited to a smaller sample of
Java projects, and it is also not clear if the results generalize to other
languages. In this work, we first replicate the prior results by building a
similar dataset and our own analysis scripts. We then extend those results by
analyzing a much larger dataset of 89k Java projects and generalizing to other
programming languages by analyzing 26k Kotlin projects and 98k Python projects.
The results show chaining is more popular in Java and Kotlin than Python,
chaining use in Kotlin is not growing, and Python sees more use in non-testing
code.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 16:54:05 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Keshk",
"Ali M.",
""
],
[
"Dyer",
"Robert",
""
]
] | TITLE: Method Chaining Redux: An Empirical Study of Method Chaining in Java,
Kotlin, and Python
ABSTRACT: There are possible benefits and drawbacks to chaining methods together, as is
often done in fluent APIs. A prior study investigated how Java developers chain
methods in over 2.7k open-source projects. That study observed, for the dataset
analyzed, that the use of method chaining in Java is popular and seems to be
increasing over time. That study however was limited to a smaller sample of
Java projects, and it is also not clear if the results generalize to other
languages. In this work, we first replicate the prior results by building a
similar dataset and our own analysis scripts. We then extend those results by
analyzing a much larger dataset of 89k Java projects and generalizing to other
programming languages by analyzing 26k Kotlin projects and 98k Python projects.
The results show chaining is more popular in Java and Kotlin than Python,
chaining use in Kotlin is not growing, and Python sees more use in non-testing
code.
| new_dataset | 0.963472 |
2303.11274 | Xianxian Zeng | Xianxian Zeng, Yanjun Zheng | Cascading Hierarchical Networks with Multi-task Balanced Loss for
Fine-grained hashing | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | With the explosive growth in the number of fine-grained images in the
Internet era, it has become a challenging problem to perform fast and efficient
retrieval from large-scale fine-grained images. Among the many retrieval
methods, hashing methods are widely used due to their high efficiency and small
storage space occupation. Fine-grained hashing is more challenging than
traditional hashing problems due to the difficulties such as low inter-class
variances and high intra-class variances caused by the characteristics of
fine-grained images. To improve the retrieval accuracy of fine-grained hashing,
we propose a cascaded network to learn compact and highly semantic hash codes,
and introduce an attention-guided data augmentation method. We refer to this
network as a cascaded hierarchical data augmentation network. We also propose a
novel approach to coordinately balance the loss of multi-task learning. We do
extensive experiments on some common fine-grained visual classification
datasets. The experimental results demonstrate that our proposed method
outperforms several state-of-art hashing methods and can effectively improve
the accuracy of fine-grained retrieval. The source code is publicly available:
https://github.com/kaiba007/FG-CNET.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:08:48 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Zeng",
"Xianxian",
""
],
[
"Zheng",
"Yanjun",
""
]
] | TITLE: Cascading Hierarchical Networks with Multi-task Balanced Loss for
Fine-grained hashing
ABSTRACT: With the explosive growth in the number of fine-grained images in the
Internet era, it has become a challenging problem to perform fast and efficient
retrieval from large-scale fine-grained images. Among the many retrieval
methods, hashing methods are widely used due to their high efficiency and small
storage space occupation. Fine-grained hashing is more challenging than
traditional hashing problems due to the difficulties such as low inter-class
variances and high intra-class variances caused by the characteristics of
fine-grained images. To improve the retrieval accuracy of fine-grained hashing,
we propose a cascaded network to learn compact and highly semantic hash codes,
and introduce an attention-guided data augmentation method. We refer to this
network as a cascaded hierarchical data augmentation network. We also propose a
novel approach to coordinately balance the loss of multi-task learning. We do
extensive experiments on some common fine-grained visual classification
datasets. The experimental results demonstrate that our proposed method
outperforms several state-of-art hashing methods and can effectively improve
the accuracy of fine-grained retrieval. The source code is publicly available:
https://github.com/kaiba007/FG-CNET.
| no_new_dataset | 0.946892 |
2303.11288 | Zhelun Li | Chase Shimmin and Zhelun Li and Ema Smith | Rethinking SO(3)-equivariance with Bilinear Tensor Networks | null | null | null | null | cs.LG hep-ex hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many datasets in scientific and engineering applications are comprised of
objects which have specific geometric structure. A common example is data which
inhabits a representation of the group SO$(3)$ of 3D rotations: scalars,
vectors, tensors, \textit{etc}. One way for a neural network to exploit prior
knowledge of this structure is to enforce SO$(3)$-equivariance throughout its
layers, and several such architectures have been proposed. While general
methods for handling arbitrary SO$(3)$ representations exist, they
computationally intensive and complicated to implement. We show that by
judicious symmetry breaking, we can efficiently increase the expressiveness of
a network operating only on vector and order-2 tensor representations of
SO$(2)$. We demonstrate the method on an important problem from High Energy
Physics known as \textit{b-tagging}, where particle jets originating from
b-meson decays must be discriminated from an overwhelming QCD background. In
this task, we find that augmenting a standard architecture with our method
results in a \ensuremath{2.3\times} improvement in rejection score.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:23:15 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Shimmin",
"Chase",
""
],
[
"Li",
"Zhelun",
""
],
[
"Smith",
"Ema",
""
]
] | TITLE: Rethinking SO(3)-equivariance with Bilinear Tensor Networks
ABSTRACT: Many datasets in scientific and engineering applications are comprised of
objects which have specific geometric structure. A common example is data which
inhabits a representation of the group SO$(3)$ of 3D rotations: scalars,
vectors, tensors, \textit{etc}. One way for a neural network to exploit prior
knowledge of this structure is to enforce SO$(3)$-equivariance throughout its
layers, and several such architectures have been proposed. While general
methods for handling arbitrary SO$(3)$ representations exist, they
computationally intensive and complicated to implement. We show that by
judicious symmetry breaking, we can efficiently increase the expressiveness of
a network operating only on vector and order-2 tensor representations of
SO$(2)$. We demonstrate the method on an important problem from High Energy
Physics known as \textit{b-tagging}, where particle jets originating from
b-meson decays must be discriminated from an overwhelming QCD background. In
this task, we find that augmenting a standard architecture with our method
results in a \ensuremath{2.3\times} improvement in rejection score.
| no_new_dataset | 0.942242 |
2303.11296 | Christos Tzelepis | Simone Barattin, Christos Tzelepis, Ioannis Patras, Nicu Sebe | Attribute-preserving Face Dataset Anonymization via Latent Code
Optimization | Accepted for publication in CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work addresses the problem of anonymizing the identity of faces in a
dataset of images, such that the privacy of those depicted is not violated,
while at the same time the dataset is useful for downstream task such as for
training machine learning models. To the best of our knowledge, we are the
first to explicitly address this issue and deal with two major drawbacks of the
existing state-of-the-art approaches, namely that they (i) require the costly
training of additional, purpose-trained neural networks, and/or (ii) fail to
retain the facial attributes of the original images in the anonymized
counterparts, the preservation of which is of paramount importance for their
use in downstream tasks. We accordingly present a task-agnostic anonymization
procedure that directly optimizes the images' latent representation in the
latent space of a pre-trained GAN. By optimizing the latent codes directly, we
ensure both that the identity is of a desired distance away from the original
(with an identity obfuscation loss), whilst preserving the facial attributes
(using a novel feature-matching loss in FaRL's deep feature space). We
demonstrate through a series of both qualitative and quantitative experiments
that our method is capable of anonymizing the identity of the images whilst --
crucially -- better-preserving the facial attributes. We make the code and the
pre-trained models publicly available at: https://github.com/chi0tzp/FALCO.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:34:05 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Barattin",
"Simone",
""
],
[
"Tzelepis",
"Christos",
""
],
[
"Patras",
"Ioannis",
""
],
[
"Sebe",
"Nicu",
""
]
] | TITLE: Attribute-preserving Face Dataset Anonymization via Latent Code
Optimization
ABSTRACT: This work addresses the problem of anonymizing the identity of faces in a
dataset of images, such that the privacy of those depicted is not violated,
while at the same time the dataset is useful for downstream task such as for
training machine learning models. To the best of our knowledge, we are the
first to explicitly address this issue and deal with two major drawbacks of the
existing state-of-the-art approaches, namely that they (i) require the costly
training of additional, purpose-trained neural networks, and/or (ii) fail to
retain the facial attributes of the original images in the anonymized
counterparts, the preservation of which is of paramount importance for their
use in downstream tasks. We accordingly present a task-agnostic anonymization
procedure that directly optimizes the images' latent representation in the
latent space of a pre-trained GAN. By optimizing the latent codes directly, we
ensure both that the identity is of a desired distance away from the original
(with an identity obfuscation loss), whilst preserving the facial attributes
(using a novel feature-matching loss in FaRL's deep feature space). We
demonstrate through a series of both qualitative and quantitative experiments
that our method is capable of anonymizing the identity of the images whilst --
crucially -- better-preserving the facial attributes. We make the code and the
pre-trained models publicly available at: https://github.com/chi0tzp/FALCO.
| no_new_dataset | 0.949995 |
2303.11301 | Chen Yukang | Yukang Chen, Jianhui Liu, Xiangyu Zhang, Xiaojuan Qi, Jiaya Jia | VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking | In CVPR 2023, Code and models are available at
https://github.com/dvlab-research/VoxelNeXt | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 3D object detectors usually rely on hand-crafted proxies, e.g., anchors or
centers, and translate well-studied 2D frameworks to 3D. Thus, sparse voxel
features need to be densified and processed by dense prediction heads, which
inevitably costs extra computation. In this paper, we instead propose VoxelNext
for fully sparse 3D object detection. Our core insight is to predict objects
directly based on sparse voxel features, without relying on hand-crafted
proxies. Our strong sparse convolutional network VoxelNeXt detects and tracks
3D objects through voxel features entirely. It is an elegant and efficient
framework, with no need for sparse-to-dense conversion or NMS post-processing.
Our method achieves a better speed-accuracy trade-off than other mainframe
detectors on the nuScenes dataset. For the first time, we show that a fully
sparse voxel-based representation works decently for LIDAR 3D object detection
and tracking. Extensive experiments on nuScenes, Waymo, and Argoverse2
benchmarks validate the effectiveness of our approach. Without bells and
whistles, our model outperforms all existing LIDAR methods on the nuScenes
tracking test benchmark.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:40:44 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Chen",
"Yukang",
""
],
[
"Liu",
"Jianhui",
""
],
[
"Zhang",
"Xiangyu",
""
],
[
"Qi",
"Xiaojuan",
""
],
[
"Jia",
"Jiaya",
""
]
] | TITLE: VoxelNeXt: Fully Sparse VoxelNet for 3D Object Detection and Tracking
ABSTRACT: 3D object detectors usually rely on hand-crafted proxies, e.g., anchors or
centers, and translate well-studied 2D frameworks to 3D. Thus, sparse voxel
features need to be densified and processed by dense prediction heads, which
inevitably costs extra computation. In this paper, we instead propose VoxelNext
for fully sparse 3D object detection. Our core insight is to predict objects
directly based on sparse voxel features, without relying on hand-crafted
proxies. Our strong sparse convolutional network VoxelNeXt detects and tracks
3D objects through voxel features entirely. It is an elegant and efficient
framework, with no need for sparse-to-dense conversion or NMS post-processing.
Our method achieves a better speed-accuracy trade-off than other mainframe
detectors on the nuScenes dataset. For the first time, we show that a fully
sparse voxel-based representation works decently for LIDAR 3D object detection
and tracking. Extensive experiments on nuScenes, Waymo, and Argoverse2
benchmarks validate the effectiveness of our approach. Without bells and
whistles, our model outperforms all existing LIDAR methods on the nuScenes
tracking test benchmark.
| no_new_dataset | 0.949153 |
2303.11320 | Xi Chen | Xi Chen, Yau Shing Jonathan Cheung, Ser-Nam Lim, Hengshuang Zhao | ScribbleSeg: Scribble-based Interactive Image Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Interactive segmentation enables users to extract masks by providing simple
annotations to indicate the target, such as boxes, clicks, or scribbles. Among
these interaction formats, scribbles are the most flexible as they can be of
arbitrary shapes and sizes. This enables scribbles to provide more indications
of the target object. However, previous works mainly focus on click-based
configuration, and the scribble-based setting is rarely explored. In this work,
we attempt to formulate a standard protocol for scribble-based interactive
segmentation. Basically, we design diversified strategies to simulate scribbles
for training, propose a deterministic scribble generator for evaluation, and
construct a challenging benchmark. Besides, we build a strong framework
ScribbleSeg, consisting of a Prototype Adaption Module(PAM) and a Corrective
Refine Module (CRM), for the task. Extensive experiments show that ScribbleSeg
performs notably better than previous click-based methods. We hope this could
serve as a more powerful and general solution for interactive segmentation. Our
code will be made available.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:57:03 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Chen",
"Xi",
""
],
[
"Cheung",
"Yau Shing Jonathan",
""
],
[
"Lim",
"Ser-Nam",
""
],
[
"Zhao",
"Hengshuang",
""
]
] | TITLE: ScribbleSeg: Scribble-based Interactive Image Segmentation
ABSTRACT: Interactive segmentation enables users to extract masks by providing simple
annotations to indicate the target, such as boxes, clicks, or scribbles. Among
these interaction formats, scribbles are the most flexible as they can be of
arbitrary shapes and sizes. This enables scribbles to provide more indications
of the target object. However, previous works mainly focus on click-based
configuration, and the scribble-based setting is rarely explored. In this work,
we attempt to formulate a standard protocol for scribble-based interactive
segmentation. Basically, we design diversified strategies to simulate scribbles
for training, propose a deterministic scribble generator for evaluation, and
construct a challenging benchmark. Besides, we build a strong framework
ScribbleSeg, consisting of a Prototype Adaption Module(PAM) and a Corrective
Refine Module (CRM), for the task. Extensive experiments show that ScribbleSeg
performs notably better than previous click-based methods. We hope this could
serve as a more powerful and general solution for interactive segmentation. Our
code will be made available.
| no_new_dataset | 0.939858 |
2303.11327 | Chuang Gan | Yining Hong, Chunru Lin, Yilun Du, Zhenfang Chen, Joshua B. Tenenbaum,
Chuang Gan | 3D Concept Learning and Reasoning from Multi-View Images | CVPR 2023. Project page: https://vis-www.cs.umass.edu/3d-clr/ | null | null | null | cs.CV cs.AI cs.LG cs.RO | http://creativecommons.org/publicdomain/zero/1.0/ | Humans are able to accurately reason in 3D by gathering multi-view
observations of the surrounding world. Inspired by this insight, we introduce a
new large-scale benchmark for 3D multi-view visual question answering
(3DMV-VQA). This dataset is collected by an embodied agent actively moving and
capturing RGB images in an environment using the Habitat simulator. In total,
it consists of approximately 5k scenes, 600k images, paired with 50k questions.
We evaluate various state-of-the-art models for visual reasoning on our
benchmark and find that they all perform poorly. We suggest that a principled
approach for 3D reasoning from multi-view images should be to infer a compact
3D representation of the world from the multi-view images, which is further
grounded on open-vocabulary semantic concepts, and then to execute reasoning on
these 3D representations. As the first step towards this approach, we propose a
novel 3D concept learning and reasoning (3D-CLR) framework that seamlessly
combines these components via neural fields, 2D pre-trained vision-language
models, and neural reasoning operators. Experimental results suggest that our
framework outperforms baseline models by a large margin, but the challenge
remains largely unsolved. We further perform an in-depth analysis of the
challenges and highlight potential future directions.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:59:49 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Hong",
"Yining",
""
],
[
"Lin",
"Chunru",
""
],
[
"Du",
"Yilun",
""
],
[
"Chen",
"Zhenfang",
""
],
[
"Tenenbaum",
"Joshua B.",
""
],
[
"Gan",
"Chuang",
""
]
] | TITLE: 3D Concept Learning and Reasoning from Multi-View Images
ABSTRACT: Humans are able to accurately reason in 3D by gathering multi-view
observations of the surrounding world. Inspired by this insight, we introduce a
new large-scale benchmark for 3D multi-view visual question answering
(3DMV-VQA). This dataset is collected by an embodied agent actively moving and
capturing RGB images in an environment using the Habitat simulator. In total,
it consists of approximately 5k scenes, 600k images, paired with 50k questions.
We evaluate various state-of-the-art models for visual reasoning on our
benchmark and find that they all perform poorly. We suggest that a principled
approach for 3D reasoning from multi-view images should be to infer a compact
3D representation of the world from the multi-view images, which is further
grounded on open-vocabulary semantic concepts, and then to execute reasoning on
these 3D representations. As the first step towards this approach, we propose a
novel 3D concept learning and reasoning (3D-CLR) framework that seamlessly
combines these components via neural fields, 2D pre-trained vision-language
models, and neural reasoning operators. Experimental results suggest that our
framework outperforms baseline models by a large margin, but the challenge
remains largely unsolved. We further perform an in-depth analysis of the
challenges and highlight potential future directions.
| new_dataset | 0.968501 |
2303.11328 | Ruoshi Liu | Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tokmakov, Sergey
Zakharov, Carl Vondrick | Zero-1-to-3: Zero-shot One Image to 3D Object | Website: https://zero123.cs.columbia.edu/ | null | null | null | cs.CV cs.GR cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce Zero-1-to-3, a framework for changing the camera viewpoint of an
object given just a single RGB image. To perform novel view synthesis in this
under-constrained setting, we capitalize on the geometric priors that
large-scale diffusion models learn about natural images. Our conditional
diffusion model uses a synthetic dataset to learn controls of the relative
camera viewpoint, which allow new images to be generated of the same object
under a specified camera transformation. Even though it is trained on a
synthetic dataset, our model retains a strong zero-shot generalization ability
to out-of-distribution datasets as well as in-the-wild images, including
impressionist paintings. Our viewpoint-conditioned diffusion approach can
further be used for the task of 3D reconstruction from a single image.
Qualitative and quantitative experiments show that our method significantly
outperforms state-of-the-art single-view 3D reconstruction and novel view
synthesis models by leveraging Internet-scale pre-training.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 17:59:50 GMT"
}
] | 2023-03-21T00:00:00 | [
[
"Liu",
"Ruoshi",
""
],
[
"Wu",
"Rundi",
""
],
[
"Van Hoorick",
"Basile",
""
],
[
"Tokmakov",
"Pavel",
""
],
[
"Zakharov",
"Sergey",
""
],
[
"Vondrick",
"Carl",
""
]
] | TITLE: Zero-1-to-3: Zero-shot One Image to 3D Object
ABSTRACT: We introduce Zero-1-to-3, a framework for changing the camera viewpoint of an
object given just a single RGB image. To perform novel view synthesis in this
under-constrained setting, we capitalize on the geometric priors that
large-scale diffusion models learn about natural images. Our conditional
diffusion model uses a synthetic dataset to learn controls of the relative
camera viewpoint, which allow new images to be generated of the same object
under a specified camera transformation. Even though it is trained on a
synthetic dataset, our model retains a strong zero-shot generalization ability
to out-of-distribution datasets as well as in-the-wild images, including
impressionist paintings. Our viewpoint-conditioned diffusion approach can
further be used for the task of 3D reconstruction from a single image.
Qualitative and quantitative experiments show that our method significantly
outperforms state-of-the-art single-view 3D reconstruction and novel view
synthesis models by leveraging Internet-scale pre-training.
| no_new_dataset | 0.954137 |
1905.10575 | Zhaohong Deng | Xiang Ma, Liangzhe Chen, Zhaohong Deng, Peng Xu, Qisheng Yan, Kup-Sze
Choi, Shitong Wang | Deep Image Feature Learning with Fuzzy Rules | Accepted by IEEE Trans. Emerging Topics in Computational Intelligence | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The methods of extracting image features are the key to many image processing
tasks. At present, the most popular method is the deep neural network which can
automatically extract robust features through end-to-end training instead of
hand-crafted feature extraction. However, the deep neural network currently
faces many challenges: 1) its effectiveness is heavily dependent on large
datasets, so the computational complexity is very high; 2) it is usually
regarded as a black box model with poor interpretability. To meet the above
challenges, a more interpretable and scalable feature learning method, i.e.,
deep image feature learning with fuzzy rules (DIFL-FR), is proposed in the
paper, which combines the rule-based fuzzy modeling technique and the deep
stacked learning strategy. The method progressively learns image features
through a layer-by-layer manner based on fuzzy rules, so the feature learning
process can be better explained by the generated rules. More importantly, the
learning process of the method is only based on forward propagation without
back propagation and iterative learning, which results in the high learning
efficiency. In addition, the method is under the settings of unsupervised
learning and can be easily extended to scenes of supervised and semi-supervised
learning. Extensive experiments are conducted on image datasets of different
scales. The results obviously show the effectiveness of the proposed method.
| [
{
"version": "v1",
"created": "Sat, 25 May 2019 11:33:02 GMT"
},
{
"version": "v2",
"created": "Fri, 1 Nov 2019 00:03:26 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 06:10:49 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ma",
"Xiang",
""
],
[
"Chen",
"Liangzhe",
""
],
[
"Deng",
"Zhaohong",
""
],
[
"Xu",
"Peng",
""
],
[
"Yan",
"Qisheng",
""
],
[
"Choi",
"Kup-Sze",
""
],
[
"Wang",
"Shitong",
""
]
] | TITLE: Deep Image Feature Learning with Fuzzy Rules
ABSTRACT: The methods of extracting image features are the key to many image processing
tasks. At present, the most popular method is the deep neural network which can
automatically extract robust features through end-to-end training instead of
hand-crafted feature extraction. However, the deep neural network currently
faces many challenges: 1) its effectiveness is heavily dependent on large
datasets, so the computational complexity is very high; 2) it is usually
regarded as a black box model with poor interpretability. To meet the above
challenges, a more interpretable and scalable feature learning method, i.e.,
deep image feature learning with fuzzy rules (DIFL-FR), is proposed in the
paper, which combines the rule-based fuzzy modeling technique and the deep
stacked learning strategy. The method progressively learns image features
through a layer-by-layer manner based on fuzzy rules, so the feature learning
process can be better explained by the generated rules. More importantly, the
learning process of the method is only based on forward propagation without
back propagation and iterative learning, which results in the high learning
efficiency. In addition, the method is under the settings of unsupervised
learning and can be easily extended to scenes of supervised and semi-supervised
learning. Extensive experiments are conducted on image datasets of different
scales. The results obviously show the effectiveness of the proposed method.
| no_new_dataset | 0.9455 |
2003.13001 | HanQin Cai | HanQin Cai, Daniel Mckenzie, Wotao Yin, Zhenliang Zhang | Zeroth-Order Regularized Optimization (ZORO): Approximately Sparse
Gradients and Adaptive Sampling | null | SIAM Journal on Optimization 32, no. 2 (2022): 687-714 | 10.1137/21M1392966 | null | math.OC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of minimizing a high-dimensional objective function,
which may include a regularization term, using (possibly noisy) evaluations of
the function. Such optimization is also called derivative-free, zeroth-order,
or black-box optimization. We propose a new $\textbf{Z}$eroth-$\textbf{O}$rder
$\textbf{R}$egularized $\textbf{O}$ptimization method, dubbed ZORO. When the
underlying gradient is approximately sparse at an iterate, ZORO needs very few
objective function evaluations to obtain a new iterate that decreases the
objective function. We achieve this with an adaptive, randomized gradient
estimator, followed by an inexact proximal-gradient scheme. Under a novel
approximately sparse gradient assumption and various different convex settings,
we show the (theoretical and empirical) convergence rate of ZORO is only
logarithmically dependent on the problem dimension. Numerical experiments show
that ZORO outperforms the existing methods with similar assumptions, on both
synthetic and real datasets.
| [
{
"version": "v1",
"created": "Sun, 29 Mar 2020 11:01:17 GMT"
},
{
"version": "v2",
"created": "Fri, 22 May 2020 00:22:59 GMT"
},
{
"version": "v3",
"created": "Tue, 6 Oct 2020 04:33:46 GMT"
},
{
"version": "v4",
"created": "Tue, 19 Jan 2021 10:01:25 GMT"
},
{
"version": "v5",
"created": "Fri, 23 Jul 2021 11:16:43 GMT"
},
{
"version": "v6",
"created": "Wed, 1 Dec 2021 03:22:09 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Cai",
"HanQin",
""
],
[
"Mckenzie",
"Daniel",
""
],
[
"Yin",
"Wotao",
""
],
[
"Zhang",
"Zhenliang",
""
]
] | TITLE: Zeroth-Order Regularized Optimization (ZORO): Approximately Sparse
Gradients and Adaptive Sampling
ABSTRACT: We consider the problem of minimizing a high-dimensional objective function,
which may include a regularization term, using (possibly noisy) evaluations of
the function. Such optimization is also called derivative-free, zeroth-order,
or black-box optimization. We propose a new $\textbf{Z}$eroth-$\textbf{O}$rder
$\textbf{R}$egularized $\textbf{O}$ptimization method, dubbed ZORO. When the
underlying gradient is approximately sparse at an iterate, ZORO needs very few
objective function evaluations to obtain a new iterate that decreases the
objective function. We achieve this with an adaptive, randomized gradient
estimator, followed by an inexact proximal-gradient scheme. Under a novel
approximately sparse gradient assumption and various different convex settings,
we show the (theoretical and empirical) convergence rate of ZORO is only
logarithmically dependent on the problem dimension. Numerical experiments show
that ZORO outperforms the existing methods with similar assumptions, on both
synthetic and real datasets.
| no_new_dataset | 0.946695 |
2011.00925 | Mingzhou Yin | Mingzhou Yin, Andrea Iannelli, Roy S. Smith | Maximum Likelihood Estimation in Data-Driven Modeling and Control | null | IEEE Transactions on Automatic Control, vol. 68, no. 1, pp.
317-328, Jan. 2023 | 10.1109/TAC.2021.3137788 | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, various algorithms for data-driven simulation and control have been
proposed based on the Willems' fundamental lemma. However, when collected data
are noisy, these methods lead to ill-conditioned data-driven model structures.
In this work, we present a maximum likelihood framework to obtain an optimal
data-driven model, the signal matrix model, in the presence of output noise.
Data compression and noise level estimation schemes are also proposed to apply
the algorithm efficiently to large datasets and unknown noise level scenarios.
Two approaches in system identification and receding horizon control are
developed based on the derived optimal estimator. The first one identifies a
finite impulse response model. This approach improves the least-squares
estimator with less restrictive assumptions. The second one applies the signal
matrix model as the predictor in predictive control. The control performance is
shown to be better than existing data-driven predictive control algorithms,
especially under high noise levels. Both approaches demonstrate that the
derived estimator provides a promising framework to apply data-driven
algorithms to noisy data.
| [
{
"version": "v1",
"created": "Mon, 2 Nov 2020 12:11:45 GMT"
},
{
"version": "v2",
"created": "Sat, 7 Nov 2020 07:12:56 GMT"
},
{
"version": "v3",
"created": "Mon, 20 Dec 2021 21:17:18 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yin",
"Mingzhou",
""
],
[
"Iannelli",
"Andrea",
""
],
[
"Smith",
"Roy S.",
""
]
] | TITLE: Maximum Likelihood Estimation in Data-Driven Modeling and Control
ABSTRACT: Recently, various algorithms for data-driven simulation and control have been
proposed based on the Willems' fundamental lemma. However, when collected data
are noisy, these methods lead to ill-conditioned data-driven model structures.
In this work, we present a maximum likelihood framework to obtain an optimal
data-driven model, the signal matrix model, in the presence of output noise.
Data compression and noise level estimation schemes are also proposed to apply
the algorithm efficiently to large datasets and unknown noise level scenarios.
Two approaches in system identification and receding horizon control are
developed based on the derived optimal estimator. The first one identifies a
finite impulse response model. This approach improves the least-squares
estimator with less restrictive assumptions. The second one applies the signal
matrix model as the predictor in predictive control. The control performance is
shown to be better than existing data-driven predictive control algorithms,
especially under high noise levels. Both approaches demonstrate that the
derived estimator provides a promising framework to apply data-driven
algorithms to noisy data.
| no_new_dataset | 0.948965 |
2112.11454 | Omid Taheri | Omid Taheri, Vasileios Choutas, Michael J. Black, and Dimitrios
Tzionas | GOAL: Generating 4D Whole-Body Motion for Hand-Object Grasping | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generating digital humans that move realistically has many applications and
is widely studied, but existing methods focus on the major limbs of the body,
ignoring the hands and head. Hands have been separately studied, but the focus
has been on generating realistic static grasps of objects. To synthesize
virtual characters that interact with the world, we need to generate full-body
motions and realistic hand grasps simultaneously. Both sub-problems are
challenging on their own and, together, the state-space of poses is
significantly larger, the scales of hand and body motions differ, and the
whole-body posture and the hand grasp must agree, satisfy physical constraints,
and be plausible. Additionally, the head is involved because the avatar must
look at the object to interact with it. For the first time, we address the
problem of generating full-body, hand and head motions of an avatar grasping an
unknown object. As input, our method, called GOAL, takes a 3D object, its
position, and a starting 3D body pose and shape. GOAL outputs a sequence of
whole-body poses using two novel networks. First, GNet generates a goal
whole-body grasp with a realistic body, head, arm, and hand pose, as well as
hand-object contact. Second, MNet generates the motion between the starting and
goal pose. This is challenging, as it requires the avatar to walk towards the
object with foot-ground contact, orient the head towards it, reach out, and
grasp it with a realistic hand pose and hand-object contact. To achieve this,
the networks exploit a representation that combines SMPL-X body parameters and
3D vertex offsets. We train and evaluate GOAL, both qualitatively and
quantitatively, on the GRAB dataset. Results show that GOAL generalizes well to
unseen objects, outperforming baselines. GOAL takes a step towards synthesizing
realistic full-body object grasping.
| [
{
"version": "v1",
"created": "Tue, 21 Dec 2021 18:59:34 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 20:59:14 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Taheri",
"Omid",
""
],
[
"Choutas",
"Vasileios",
""
],
[
"Black",
"Michael J.",
""
],
[
"Tzionas",
"Dimitrios",
""
]
] | TITLE: GOAL: Generating 4D Whole-Body Motion for Hand-Object Grasping
ABSTRACT: Generating digital humans that move realistically has many applications and
is widely studied, but existing methods focus on the major limbs of the body,
ignoring the hands and head. Hands have been separately studied, but the focus
has been on generating realistic static grasps of objects. To synthesize
virtual characters that interact with the world, we need to generate full-body
motions and realistic hand grasps simultaneously. Both sub-problems are
challenging on their own and, together, the state-space of poses is
significantly larger, the scales of hand and body motions differ, and the
whole-body posture and the hand grasp must agree, satisfy physical constraints,
and be plausible. Additionally, the head is involved because the avatar must
look at the object to interact with it. For the first time, we address the
problem of generating full-body, hand and head motions of an avatar grasping an
unknown object. As input, our method, called GOAL, takes a 3D object, its
position, and a starting 3D body pose and shape. GOAL outputs a sequence of
whole-body poses using two novel networks. First, GNet generates a goal
whole-body grasp with a realistic body, head, arm, and hand pose, as well as
hand-object contact. Second, MNet generates the motion between the starting and
goal pose. This is challenging, as it requires the avatar to walk towards the
object with foot-ground contact, orient the head towards it, reach out, and
grasp it with a realistic hand pose and hand-object contact. To achieve this,
the networks exploit a representation that combines SMPL-X body parameters and
3D vertex offsets. We train and evaluate GOAL, both qualitatively and
quantitatively, on the GRAB dataset. Results show that GOAL generalizes well to
unseen objects, outperforming baselines. GOAL takes a step towards synthesizing
realistic full-body object grasping.
| no_new_dataset | 0.941331 |
2201.09391 | Ziqiao Ma | Jiaqi Ma, Ziqiao Ma, Joyce Chai, Qiaozhu Mei | Partition-Based Active Learning for Graph Neural Networks | Accepted to Transactions on Machine Learning Research (TMLR). Code
available at: https://github.com/Mars-tin/GraphPart | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of semi-supervised learning with Graph Neural Networks
(GNNs) in an active learning setup. We propose GraphPart, a novel
partition-based active learning approach for GNNs. GraphPart first splits the
graph into disjoint partitions and then selects representative nodes within
each partition to query. The proposed method is motivated by a novel analysis
of the classification error under realistic smoothness assumptions over the
graph and the node features. Extensive experiments on multiple benchmark
datasets demonstrate that the proposed method outperforms existing active
learning methods for GNNs under a wide range of annotation budget constraints.
In addition, the proposed method does not introduce additional hyperparameters,
which is crucial for model training, especially in the active learning setting
where a labeled validation set may not be available.
| [
{
"version": "v1",
"created": "Sun, 23 Jan 2022 22:51:14 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 05:57:03 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ma",
"Jiaqi",
""
],
[
"Ma",
"Ziqiao",
""
],
[
"Chai",
"Joyce",
""
],
[
"Mei",
"Qiaozhu",
""
]
] | TITLE: Partition-Based Active Learning for Graph Neural Networks
ABSTRACT: We study the problem of semi-supervised learning with Graph Neural Networks
(GNNs) in an active learning setup. We propose GraphPart, a novel
partition-based active learning approach for GNNs. GraphPart first splits the
graph into disjoint partitions and then selects representative nodes within
each partition to query. The proposed method is motivated by a novel analysis
of the classification error under realistic smoothness assumptions over the
graph and the node features. Extensive experiments on multiple benchmark
datasets demonstrate that the proposed method outperforms existing active
learning methods for GNNs under a wide range of annotation budget constraints.
In addition, the proposed method does not introduce additional hyperparameters,
which is crucial for model training, especially in the active learning setting
where a labeled validation set may not be available.
| no_new_dataset | 0.950134 |
2201.11316 | Moyuru Yamada | Moyuru Yamada, Vanessa D'Amario, Kentaro Takemoto, Xavier Boix, and
Tomotake Sasaki | Transformer Module Networks for Systematic Generalization in Visual
Question Answering | null | null | null | CBMM Memo No. 121, Center for Brains, Minds and Machines | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Transformers achieve great performance on Visual Question Answering (VQA).
However, their systematic generalization capabilities, i.e., handling novel
combinations of known concepts, is unclear. We reveal that Neural Module
Networks (NMNs), i.e., question-specific compositions of modules that tackle a
sub-task, achieve better or similar systematic generalization performance than
the conventional Transformers, even though NMNs' modules are CNN-based. In
order to address this shortcoming of Transformers with respect to NMNs, in this
paper we investigate whether and how modularity can bring benefits to
Transformers. Namely, we introduce Transformer Module Network (TMN), a novel
NMN based on compositions of Transformer modules. TMNs achieve state-of-the-art
systematic generalization performance in three VQA datasets, improving more
than 30% over standard Transformers for novel compositions of sub-tasks. We
show that not only the module composition but also the module specialization
for each sub-task are the key of such performance gain.
| [
{
"version": "v1",
"created": "Thu, 27 Jan 2022 04:22:25 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 11:10:31 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yamada",
"Moyuru",
""
],
[
"D'Amario",
"Vanessa",
""
],
[
"Takemoto",
"Kentaro",
""
],
[
"Boix",
"Xavier",
""
],
[
"Sasaki",
"Tomotake",
""
]
] | TITLE: Transformer Module Networks for Systematic Generalization in Visual
Question Answering
ABSTRACT: Transformers achieve great performance on Visual Question Answering (VQA).
However, their systematic generalization capabilities, i.e., handling novel
combinations of known concepts, is unclear. We reveal that Neural Module
Networks (NMNs), i.e., question-specific compositions of modules that tackle a
sub-task, achieve better or similar systematic generalization performance than
the conventional Transformers, even though NMNs' modules are CNN-based. In
order to address this shortcoming of Transformers with respect to NMNs, in this
paper we investigate whether and how modularity can bring benefits to
Transformers. Namely, we introduce Transformer Module Network (TMN), a novel
NMN based on compositions of Transformer modules. TMNs achieve state-of-the-art
systematic generalization performance in three VQA datasets, improving more
than 30% over standard Transformers for novel compositions of sub-tasks. We
show that not only the module composition but also the module specialization
for each sub-task are the key of such performance gain.
| no_new_dataset | 0.94428 |
2203.08213 | Zalan Fabian | Zalan Fabian, Berk Tinaz, Mahdi Soltanolkotabi | HUMUS-Net: Hybrid unrolled multi-scale network architecture for
accelerated MRI reconstruction | 18 pages, 11 figures, NeurIPS 2022 | null | null | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | In accelerated MRI reconstruction, the anatomy of a patient is recovered from
a set of under-sampled and noisy measurements. Deep learning approaches have
been proven to be successful in solving this ill-posed inverse problem and are
capable of producing very high quality reconstructions. However, current
architectures heavily rely on convolutions, that are content-independent and
have difficulties modeling long-range dependencies in images. Recently,
Transformers, the workhorse of contemporary natural language processing, have
emerged as powerful building blocks for a multitude of vision tasks. These
models split input images into non-overlapping patches, embed the patches into
lower-dimensional tokens and utilize a self-attention mechanism that does not
suffer from the aforementioned weaknesses of convolutional architectures.
However, Transformers incur extremely high compute and memory cost when 1) the
input image resolution is high and 2) when the image needs to be split into a
large number of patches to preserve fine detail information, both of which are
typical in low-level vision problems such as MRI reconstruction, having a
compounding effect. To tackle these challenges, we propose HUMUS-Net, a hybrid
architecture that combines the beneficial implicit bias and efficiency of
convolutions with the power of Transformer blocks in an unrolled and
multi-scale network. HUMUS-Net extracts high-resolution features via
convolutional blocks and refines low-resolution features via a novel
Transformer-based multi-scale feature extractor. Features from both levels are
then synthesized into a high-resolution output reconstruction. Our network
establishes new state of the art on the largest publicly available MRI dataset,
the fastMRI dataset. We further demonstrate the performance of HUMUS-Net on two
other popular MRI datasets and perform fine-grained ablation studies to
validate our design.
| [
{
"version": "v1",
"created": "Tue, 15 Mar 2022 19:26:29 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 00:00:54 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Fabian",
"Zalan",
""
],
[
"Tinaz",
"Berk",
""
],
[
"Soltanolkotabi",
"Mahdi",
""
]
] | TITLE: HUMUS-Net: Hybrid unrolled multi-scale network architecture for
accelerated MRI reconstruction
ABSTRACT: In accelerated MRI reconstruction, the anatomy of a patient is recovered from
a set of under-sampled and noisy measurements. Deep learning approaches have
been proven to be successful in solving this ill-posed inverse problem and are
capable of producing very high quality reconstructions. However, current
architectures heavily rely on convolutions, that are content-independent and
have difficulties modeling long-range dependencies in images. Recently,
Transformers, the workhorse of contemporary natural language processing, have
emerged as powerful building blocks for a multitude of vision tasks. These
models split input images into non-overlapping patches, embed the patches into
lower-dimensional tokens and utilize a self-attention mechanism that does not
suffer from the aforementioned weaknesses of convolutional architectures.
However, Transformers incur extremely high compute and memory cost when 1) the
input image resolution is high and 2) when the image needs to be split into a
large number of patches to preserve fine detail information, both of which are
typical in low-level vision problems such as MRI reconstruction, having a
compounding effect. To tackle these challenges, we propose HUMUS-Net, a hybrid
architecture that combines the beneficial implicit bias and efficiency of
convolutions with the power of Transformer blocks in an unrolled and
multi-scale network. HUMUS-Net extracts high-resolution features via
convolutional blocks and refines low-resolution features via a novel
Transformer-based multi-scale feature extractor. Features from both levels are
then synthesized into a high-resolution output reconstruction. Our network
establishes new state of the art on the largest publicly available MRI dataset,
the fastMRI dataset. We further demonstrate the performance of HUMUS-Net on two
other popular MRI datasets and perform fine-grained ablation studies to
validate our design.
| no_new_dataset | 0.949529 |
2203.11205 | Huy Hieu Pham | Hieu T. Nguyen, Ha Q. Nguyen, Hieu H. Pham, Khanh Lam, Linh T. Le,
Minh Dao, and Van Vu | VinDr-Mammo: A large-scale benchmark dataset for computer-aided
diagnosis in full-field digital mammography | The manuscript is accepted for publication by Scientific Data
(Nature) | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Mammography, or breast X-ray, is the most widely used imaging modality to
detect cancer and other breast diseases. Recent studies have shown that deep
learning-based computer-assisted detection and diagnosis (CADe or CADx) tools
have been developed to support physicians and improve the accuracy of
interpreting mammography. However, most published datasets of mammography are
either limited on sample size or digitalized from screen-film mammography
(SFM), hindering the development of CADe and CADx tools which are developed
based on full-field digital mammography (FFDM). To overcome this challenge, we
introduce VinDr-Mammo - a new benchmark dataset of FFDM for detecting and
diagnosing breast cancer and other diseases in mammography. The dataset
consists of 5,000 mammography exams, each of which has four standard views and
is double read with disagreement (if any) being resolved by arbitration. It is
created for the assessment of Breast Imaging Reporting and Data System
(BI-RADS) and density at the breast level. In addition, the dataset also
provides the category, location, and BI-RADS assessment of non-benign findings.
We make VinDr-Mammo publicly available on PhysioNet as a new imaging resource
to promote advances in developing CADe and CADx tools for breast cancer
screening.
| [
{
"version": "v1",
"created": "Sun, 20 Mar 2022 18:17:42 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 18:23:10 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Nguyen",
"Hieu T.",
""
],
[
"Nguyen",
"Ha Q.",
""
],
[
"Pham",
"Hieu H.",
""
],
[
"Lam",
"Khanh",
""
],
[
"Le",
"Linh T.",
""
],
[
"Dao",
"Minh",
""
],
[
"Vu",
"Van",
""
]
] | TITLE: VinDr-Mammo: A large-scale benchmark dataset for computer-aided
diagnosis in full-field digital mammography
ABSTRACT: Mammography, or breast X-ray, is the most widely used imaging modality to
detect cancer and other breast diseases. Recent studies have shown that deep
learning-based computer-assisted detection and diagnosis (CADe or CADx) tools
have been developed to support physicians and improve the accuracy of
interpreting mammography. However, most published datasets of mammography are
either limited on sample size or digitalized from screen-film mammography
(SFM), hindering the development of CADe and CADx tools which are developed
based on full-field digital mammography (FFDM). To overcome this challenge, we
introduce VinDr-Mammo - a new benchmark dataset of FFDM for detecting and
diagnosing breast cancer and other diseases in mammography. The dataset
consists of 5,000 mammography exams, each of which has four standard views and
is double read with disagreement (if any) being resolved by arbitration. It is
created for the assessment of Breast Imaging Reporting and Data System
(BI-RADS) and density at the breast level. In addition, the dataset also
provides the category, location, and BI-RADS assessment of non-benign findings.
We make VinDr-Mammo publicly available on PhysioNet as a new imaging resource
to promote advances in developing CADe and CADx tools for breast cancer
screening.
| new_dataset | 0.962708 |
2203.14731 | Mingzhou Yin | Mingzhou Yin, Mehmet Tolga Akan, Andrea Iannelli, Roy S. Smith | Infinite-Dimensional Sparse Learning in Linear System Identification | Accepted for presentation at IEEE Conference on Decision and Control
2022 | 2022 IEEE 61st Conference on Decision and Control (CDC), Cancun,
Mexico, 2022, pp. 850-855 | 10.1109/CDC51059.2022.9992728 | null | eess.SY cs.SY stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Regularized methods have been widely applied to system identification
problems without known model structures. This paper proposes an
infinite-dimensional sparse learning algorithm based on atomic norm
regularization. Atomic norm regularization decomposes the transfer function
into first-order atomic models and solves a group lasso problem that selects a
sparse set of poles and identifies the corresponding coefficients. The
difficulty in solving the problem lies in the fact that there are an infinite
number of possible atomic models. This work proposes a greedy algorithm that
generates new candidate atomic models maximizing the violation of the
optimality condition of the existing problem. This algorithm is able to solve
the infinite-dimensional group lasso problem with high precision. The algorithm
is further extended to reduce the bias and reject false positives in pole
location estimation by iteratively reweighted adaptive group lasso and
complementary pairs stability selection respectively. Numerical results
demonstrate that the proposed algorithm performs better than benchmark
parameterized and regularized methods in terms of both impulse response fitting
and pole location estimation.
| [
{
"version": "v1",
"created": "Mon, 28 Mar 2022 13:18:48 GMT"
},
{
"version": "v2",
"created": "Wed, 31 Aug 2022 12:48:29 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yin",
"Mingzhou",
""
],
[
"Akan",
"Mehmet Tolga",
""
],
[
"Iannelli",
"Andrea",
""
],
[
"Smith",
"Roy S.",
""
]
] | TITLE: Infinite-Dimensional Sparse Learning in Linear System Identification
ABSTRACT: Regularized methods have been widely applied to system identification
problems without known model structures. This paper proposes an
infinite-dimensional sparse learning algorithm based on atomic norm
regularization. Atomic norm regularization decomposes the transfer function
into first-order atomic models and solves a group lasso problem that selects a
sparse set of poles and identifies the corresponding coefficients. The
difficulty in solving the problem lies in the fact that there are an infinite
number of possible atomic models. This work proposes a greedy algorithm that
generates new candidate atomic models maximizing the violation of the
optimality condition of the existing problem. This algorithm is able to solve
the infinite-dimensional group lasso problem with high precision. The algorithm
is further extended to reduce the bias and reject false positives in pole
location estimation by iteratively reweighted adaptive group lasso and
complementary pairs stability selection respectively. Numerical results
demonstrate that the proposed algorithm performs better than benchmark
parameterized and regularized methods in terms of both impulse response fitting
and pole location estimation.
| no_new_dataset | 0.9463 |
2203.16306 | Mingzhou Yin | Defne E. Ozan, Mingzhou Yin, Andrea Iannelli, Roy S. Smith | Kernel-Based Identification of Local Limit Cycle Dynamics with Linear
Periodically Parameter-Varying Models | Accepted for presentation at IEEE Conference on Decision and Control
2022 | 2022 IEEE 61st Conference on Decision and Control (CDC), Cancun,
Mexico, 2022, pp. 221-226 | 10.1109/CDC51059.2022.9992791 | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Limit cycle oscillations are phenomena arising in nonlinear dynamical systems
and characterized by periodic, locally-stable, and self-sustained state
trajectories. Systems controlled in a closed loop along a periodic trajectory
can also be modelled as systems experiencing limit cycle behavior. The goal of
this work is to identify from data, the local dynamics around the limit cycle
using linear periodically parameter-varying models. Using a coordinate
transformation onto transversal surfaces, the dynamics are decomposed into two
parts: one along the limit cycle, and one on the transversal surfaces. Then,
the model is identified from trajectory data using kernel-based methods with a
periodic kernel design. The kernel-based model is extended to also account for
variations in system parameters associated with different operating conditions.
The performance of the proposed identification method is demonstrated on a
benchmark nonlinear system and on a simplified airborne wind energy model. The
method provides accurate model parameter estimation, compared to the analytical
linearization, and good prediction capability.
| [
{
"version": "v1",
"created": "Wed, 30 Mar 2022 13:51:24 GMT"
},
{
"version": "v2",
"created": "Wed, 31 Aug 2022 12:56:47 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ozan",
"Defne E.",
""
],
[
"Yin",
"Mingzhou",
""
],
[
"Iannelli",
"Andrea",
""
],
[
"Smith",
"Roy S.",
""
]
] | TITLE: Kernel-Based Identification of Local Limit Cycle Dynamics with Linear
Periodically Parameter-Varying Models
ABSTRACT: Limit cycle oscillations are phenomena arising in nonlinear dynamical systems
and characterized by periodic, locally-stable, and self-sustained state
trajectories. Systems controlled in a closed loop along a periodic trajectory
can also be modelled as systems experiencing limit cycle behavior. The goal of
this work is to identify from data, the local dynamics around the limit cycle
using linear periodically parameter-varying models. Using a coordinate
transformation onto transversal surfaces, the dynamics are decomposed into two
parts: one along the limit cycle, and one on the transversal surfaces. Then,
the model is identified from trajectory data using kernel-based methods with a
periodic kernel design. The kernel-based model is extended to also account for
variations in system parameters associated with different operating conditions.
The performance of the proposed identification method is demonstrated on a
benchmark nonlinear system and on a simplified airborne wind energy model. The
method provides accurate model parameter estimation, compared to the analytical
linearization, and good prediction capability.
| no_new_dataset | 0.951142 |
2204.03025 | Zichao Li | Zichao Li, Prakhar Sharma, Xing Han Lu, Jackie C.K. Cheung, Siva Reddy | Using Interactive Feedback to Improve the Accuracy and Explainability of
Question Answering Systems Post-Deployment | ACL 2022 Findings | Findings of the Association for Computational Linguistics: ACL
(2022) 926-937 | 10.18653/v1/2022.findings-acl.75 | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Most research on question answering focuses on the pre-deployment stage;
i.e., building an accurate model for deployment. In this paper, we ask the
question: Can we improve QA systems further \emph{post-}deployment based on
user interactions? We focus on two kinds of improvements: 1) improving the QA
system's performance itself, and 2) providing the model with the ability to
explain the correctness or incorrectness of an answer. We collect a
retrieval-based QA dataset, FeedbackQA, which contains interactive feedback
from users. We collect this dataset by deploying a base QA system to
crowdworkers who then engage with the system and provide feedback on the
quality of its answers. The feedback contains both structured ratings and
unstructured natural language explanations. We train a neural model with this
feedback data that can generate explanations and re-score answer candidates. We
show that feedback data not only improves the accuracy of the deployed QA
system but also other stronger non-deployed systems. The generated explanations
also help users make informed decisions about the correctness of answers.
Project page: https://mcgill-nlp.github.io/feedbackqa/
| [
{
"version": "v1",
"created": "Wed, 6 Apr 2022 18:17:09 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Li",
"Zichao",
""
],
[
"Sharma",
"Prakhar",
""
],
[
"Lu",
"Xing Han",
""
],
[
"Cheung",
"Jackie C. K.",
""
],
[
"Reddy",
"Siva",
""
]
] | TITLE: Using Interactive Feedback to Improve the Accuracy and Explainability of
Question Answering Systems Post-Deployment
ABSTRACT: Most research on question answering focuses on the pre-deployment stage;
i.e., building an accurate model for deployment. In this paper, we ask the
question: Can we improve QA systems further \emph{post-}deployment based on
user interactions? We focus on two kinds of improvements: 1) improving the QA
system's performance itself, and 2) providing the model with the ability to
explain the correctness or incorrectness of an answer. We collect a
retrieval-based QA dataset, FeedbackQA, which contains interactive feedback
from users. We collect this dataset by deploying a base QA system to
crowdworkers who then engage with the system and provide feedback on the
quality of its answers. The feedback contains both structured ratings and
unstructured natural language explanations. We train a neural model with this
feedback data that can generate explanations and re-score answer candidates. We
show that feedback data not only improves the accuracy of the deployed QA
system but also other stronger non-deployed systems. The generated explanations
also help users make informed decisions about the correctness of answers.
Project page: https://mcgill-nlp.github.io/feedbackqa/
| new_dataset | 0.959154 |
2204.05453 | Dong Huo | Dong Huo, Jian Wang, Yiming Qian, Yee-Hong Yang | Glass Segmentation with RGB-Thermal Image Pairs | IEEE Transactions on Image Processing (TIP), 2023 | null | 10.1109/TIP.2023.3256762 | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper proposes a new glass segmentation method utilizing paired RGB and
thermal images. Due to the large difference between the transmission property
of visible light and that of the thermal energy through the glass where most
glass is transparent to the visible light but opaque to thermal energy, glass
regions of a scene are made more distinguishable with a pair of RGB and thermal
images than solely with an RGB image. To exploit such a unique property, we
propose a neural network architecture that effectively combines an RGB-thermal
image pair with a new multi-modal fusion module based on attention, and
integrate CNN and transformer to extract local features and non-local
dependencies, respectively. As well, we have collected a new dataset containing
5551 RGB-thermal image pairs with ground-truth segmentation annotations. The
qualitative and quantitative evaluations demonstrate the effectiveness of the
proposed approach on fusing RGB and thermal data for glass segmentation. Our
code and data are available at
https://github.com/Dong-Huo/RGB-T-Glass-Segmentation.
| [
{
"version": "v1",
"created": "Tue, 12 Apr 2022 00:20:22 GMT"
},
{
"version": "v2",
"created": "Wed, 13 Apr 2022 06:18:12 GMT"
},
{
"version": "v3",
"created": "Mon, 23 May 2022 01:27:32 GMT"
},
{
"version": "v4",
"created": "Thu, 16 Mar 2023 19:21:26 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Huo",
"Dong",
""
],
[
"Wang",
"Jian",
""
],
[
"Qian",
"Yiming",
""
],
[
"Yang",
"Yee-Hong",
""
]
] | TITLE: Glass Segmentation with RGB-Thermal Image Pairs
ABSTRACT: This paper proposes a new glass segmentation method utilizing paired RGB and
thermal images. Due to the large difference between the transmission property
of visible light and that of the thermal energy through the glass where most
glass is transparent to the visible light but opaque to thermal energy, glass
regions of a scene are made more distinguishable with a pair of RGB and thermal
images than solely with an RGB image. To exploit such a unique property, we
propose a neural network architecture that effectively combines an RGB-thermal
image pair with a new multi-modal fusion module based on attention, and
integrate CNN and transformer to extract local features and non-local
dependencies, respectively. As well, we have collected a new dataset containing
5551 RGB-thermal image pairs with ground-truth segmentation annotations. The
qualitative and quantitative evaluations demonstrate the effectiveness of the
proposed approach on fusing RGB and thermal data for glass segmentation. Our
code and data are available at
https://github.com/Dong-Huo/RGB-T-Glass-Segmentation.
| new_dataset | 0.955899 |
2205.10685 | Jessica M\'egane | Jessica M\'egane and Nuno Louren\c{c}o and Penousal Machado | Probabilistic Structured Grammatical Evolution | arXiv admin note: text overlap with arXiv:2204.08985 | null | 10.1109/CEC55065.2022.9870397 | null | cs.NE | http://creativecommons.org/licenses/by/4.0/ | The grammars used in grammar-based Genetic Programming (GP) methods have a
significant impact on the quality of the solutions generated since they define
the search space by restricting the solutions to its syntax. In this work, we
propose Probabilistic Structured Grammatical Evolution (PSGE), a new approach
that combines the Structured Grammatical Evolution (SGE) and Probabilistic
Grammatical Evolution (PGE) representation variants and mapping mechanisms. The
genotype is a set of dynamic lists, one for each non-terminal in the grammar,
with each element of the list representing a probability used to select the
next Probabilistic Context-Free Grammar (PCFG) derivation rule. PSGE
statistically outperformed Grammatical Evolution (GE) on all six benchmark
problems studied. In comparison to PGE, PSGE outperformed 4 of the 6 problems
analyzed.
| [
{
"version": "v1",
"created": "Sat, 21 May 2022 22:29:10 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Mégane",
"Jessica",
""
],
[
"Lourenço",
"Nuno",
""
],
[
"Machado",
"Penousal",
""
]
] | TITLE: Probabilistic Structured Grammatical Evolution
ABSTRACT: The grammars used in grammar-based Genetic Programming (GP) methods have a
significant impact on the quality of the solutions generated since they define
the search space by restricting the solutions to its syntax. In this work, we
propose Probabilistic Structured Grammatical Evolution (PSGE), a new approach
that combines the Structured Grammatical Evolution (SGE) and Probabilistic
Grammatical Evolution (PGE) representation variants and mapping mechanisms. The
genotype is a set of dynamic lists, one for each non-terminal in the grammar,
with each element of the list representing a probability used to select the
next Probabilistic Context-Free Grammar (PCFG) derivation rule. PSGE
statistically outperformed Grammatical Evolution (GE) on all six benchmark
problems studied. In comparison to PGE, PSGE outperformed 4 of the 6 problems
analyzed.
| no_new_dataset | 0.952618 |
2206.04113 | Yu-Guan Hsieh | Yu-Guan Hsieh, Yassine Laguel, Franck Iutzeler, J\'er\^ome Malick | Push--Pull with Device Sampling | In IEEE Transactions on Automatic Control | null | null | null | math.OC cs.DC cs.LG cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider decentralized optimization problems in which a number of agents
collaborate to minimize the average of their local functions by exchanging over
an underlying communication graph. Specifically, we place ourselves in an
asynchronous model where only a random portion of nodes perform computation at
each iteration, while the information exchange can be conducted between all the
nodes and in an asymmetric fashion. For this setting, we propose an algorithm
that combines gradient tracking with a network-level variance reduction (in
contrast to variance reduction within each node). This enables each node to
track the average of the gradients of the objective functions. Our theoretical
analysis shows that the algorithm converges linearly, when the local objective
functions are strongly convex, under mild connectivity conditions on the
expected mixing matrices. In particular, our result does not require the mixing
matrices to be doubly stochastic. In the experiments, we investigate a
broadcast mechanism that transmits information from computing nodes to their
neighbors, and confirm the linear convergence of our method on both synthetic
and real-world datasets.
| [
{
"version": "v1",
"created": "Wed, 8 Jun 2022 18:18:18 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 16:38:10 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Hsieh",
"Yu-Guan",
""
],
[
"Laguel",
"Yassine",
""
],
[
"Iutzeler",
"Franck",
""
],
[
"Malick",
"Jérôme",
""
]
] | TITLE: Push--Pull with Device Sampling
ABSTRACT: We consider decentralized optimization problems in which a number of agents
collaborate to minimize the average of their local functions by exchanging over
an underlying communication graph. Specifically, we place ourselves in an
asynchronous model where only a random portion of nodes perform computation at
each iteration, while the information exchange can be conducted between all the
nodes and in an asymmetric fashion. For this setting, we propose an algorithm
that combines gradient tracking with a network-level variance reduction (in
contrast to variance reduction within each node). This enables each node to
track the average of the gradients of the objective functions. Our theoretical
analysis shows that the algorithm converges linearly, when the local objective
functions are strongly convex, under mild connectivity conditions on the
expected mixing matrices. In particular, our result does not require the mixing
matrices to be doubly stochastic. In the experiments, we investigate a
broadcast mechanism that transmits information from computing nodes to their
neighbors, and confirm the linear convergence of our method on both synthetic
and real-world datasets.
| no_new_dataset | 0.94887 |
2208.09015 | Lovish Madaan | Lovish Madaan, Srinadh Bhojanapalli, Himanshu Jain, Prateek Jain | Treeformer: Dense Gradient Trees for Efficient Attention Computation | ICLR 2023 | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Standard inference and training with transformer based architectures scale
quadratically with input sequence length. This is prohibitively large for a
variety of applications especially in web-page translation, query-answering
etc. Consequently, several approaches have been developed recently to speedup
attention computation by enforcing different attention structures such as
sparsity, low-rank, approximating attention using kernels. In this work, we
view attention computation as that of nearest neighbor retrieval, and use
decision tree based hierarchical navigation to reduce the retrieval cost per
query token from linear in sequence length to nearly logarithmic. Based on such
hierarchical navigation, we design Treeformer which can use one of two
efficient attention layers -- TF-Attention and TC-Attention. TF-Attention
computes the attention in a fine-grained style, while TC-Attention is a coarse
attention layer which also ensures that the gradients are "dense". To optimize
such challenging discrete layers, we propose a two-level bootstrapped training
method. Using extensive experiments on standard NLP benchmarks, especially for
long-sequences, we demonstrate that our Treeformer architecture can be almost
as accurate as baseline Transformer while using 30x lesser FLOPs in the
attention layer. Compared to Linformer, the accuracy can be as much as 12%
higher while using similar FLOPs in the attention layer.
| [
{
"version": "v1",
"created": "Thu, 18 Aug 2022 18:31:40 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 14:25:43 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Madaan",
"Lovish",
""
],
[
"Bhojanapalli",
"Srinadh",
""
],
[
"Jain",
"Himanshu",
""
],
[
"Jain",
"Prateek",
""
]
] | TITLE: Treeformer: Dense Gradient Trees for Efficient Attention Computation
ABSTRACT: Standard inference and training with transformer based architectures scale
quadratically with input sequence length. This is prohibitively large for a
variety of applications especially in web-page translation, query-answering
etc. Consequently, several approaches have been developed recently to speedup
attention computation by enforcing different attention structures such as
sparsity, low-rank, approximating attention using kernels. In this work, we
view attention computation as that of nearest neighbor retrieval, and use
decision tree based hierarchical navigation to reduce the retrieval cost per
query token from linear in sequence length to nearly logarithmic. Based on such
hierarchical navigation, we design Treeformer which can use one of two
efficient attention layers -- TF-Attention and TC-Attention. TF-Attention
computes the attention in a fine-grained style, while TC-Attention is a coarse
attention layer which also ensures that the gradients are "dense". To optimize
such challenging discrete layers, we propose a two-level bootstrapped training
method. Using extensive experiments on standard NLP benchmarks, especially for
long-sequences, we demonstrate that our Treeformer architecture can be almost
as accurate as baseline Transformer while using 30x lesser FLOPs in the
attention layer. Compared to Linformer, the accuracy can be as much as 12%
higher while using similar FLOPs in the attention layer.
| no_new_dataset | 0.949669 |
2208.13418 | Jiehui Zhou | Jiehui Zhou, Xumeng Wang, Jason K. Wong, Huanliang Wang, Zhongwei
Wang, Xiaoyu Yang, Xiaoran Yan, Haozhe Feng, Huamin Qu, Haochao Ying, and Wei
Chen | DPVisCreator: Incorporating Pattern Constraints to Privacy-preserving
Visualizations via Differential Privacy | 9 pages, 5 figures | IEEE Transactions on Visualization and Computer Graphics, vol. 29,
no. 1, pp. 809-819, Jan. 2023 | 10.1109/TVCG.2022.3209391 | null | cs.HC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Data privacy is an essential issue in publishing data visualizations.
However, it is challenging to represent multiple data patterns in
privacy-preserving visualizations. The prior approaches target specific chart
types or perform an anonymization model uniformly without considering the
importance of data patterns in visualizations. In this paper, we propose a
visual analytics approach that facilitates data custodians to generate multiple
private charts while maintaining user-preferred patterns. To this end, we
introduce pattern constraints to model users' preferences over data patterns in
the dataset and incorporate them into the proposed Bayesian network-based
Differential Privacy (DP) model PriVis. A prototype system, DPVisCreator, is
developed to assist data custodians in implementing our approach. The
effectiveness of our approach is demonstrated with quantitative evaluation of
pattern utility under the different levels of privacy protection, case studies,
and semi-structured expert interviews.
| [
{
"version": "v1",
"created": "Mon, 29 Aug 2022 08:23:44 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Zhou",
"Jiehui",
""
],
[
"Wang",
"Xumeng",
""
],
[
"Wong",
"Jason K.",
""
],
[
"Wang",
"Huanliang",
""
],
[
"Wang",
"Zhongwei",
""
],
[
"Yang",
"Xiaoyu",
""
],
[
"Yan",
"Xiaoran",
""
],
[
"Feng",
"Haozhe",
""
],
[
"Qu",
"Huamin",
""
],
[
"Ying",
"Haochao",
""
],
[
"Chen",
"Wei",
""
]
] | TITLE: DPVisCreator: Incorporating Pattern Constraints to Privacy-preserving
Visualizations via Differential Privacy
ABSTRACT: Data privacy is an essential issue in publishing data visualizations.
However, it is challenging to represent multiple data patterns in
privacy-preserving visualizations. The prior approaches target specific chart
types or perform an anonymization model uniformly without considering the
importance of data patterns in visualizations. In this paper, we propose a
visual analytics approach that facilitates data custodians to generate multiple
private charts while maintaining user-preferred patterns. To this end, we
introduce pattern constraints to model users' preferences over data patterns in
the dataset and incorporate them into the proposed Bayesian network-based
Differential Privacy (DP) model PriVis. A prototype system, DPVisCreator, is
developed to assist data custodians in implementing our approach. The
effectiveness of our approach is demonstrated with quantitative evaluation of
pattern utility under the different levels of privacy protection, case studies,
and semi-structured expert interviews.
| no_new_dataset | 0.95275 |
2209.01496 | Jingyuan Zhang | Jingyuan Zhang, Ao Wang, Xiaolong Ma, Benjamin Carver, Nicholas John
Newman, Ali Anwar, Lukas Rupprecht, Dimitrios Skourtis, Vasily Tarasov, Feng
Yan, Yue Cheng | InfiniStore: Elastic Serverless Cloud Storage | An extensive report of the paper accepted by VLDB 2023 | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cloud object storage such as AWS S3 is cost-effective and highly elastic but
relatively slow, while high-performance cloud storage such as AWS ElastiCache
is expensive and provides limited elasticity. We present a new cloud storage
service called ServerlessMemory, which stores data using the memory of
serverless functions. ServerlessMemory employs a sliding-window-based memory
management strategy inspired by the garbage collection mechanisms used in the
programming language to effectively segregate hot/cold data and provides
fine-grained elasticity, good performance, and a pay-per-access cost model with
extremely low cost.
We then design and implement InfiniStore, a persistent and elastic cloud
storage system, which seamlessly couples the function-based ServerlessMemory
layer with a persistent, inexpensive cloud object store layer. InfiniStore
enables durability despite function failures using a fast parallel recovery
scheme built on the autoscaling functionality of a FaaS (Function-as-a-Service)
platform. We evaluate InfiniStore extensively using both microbenchmarking and
two real-world applications. Results show that InfiniStore has more performance
benefits for objects larger than 10 MB compared to AWS ElastiCache and Anna,
and InfiniStore achieves 26.25% and 97.24% tenant-side cost reduction compared
to InfiniCache and ElastiCache, respectively.
| [
{
"version": "v1",
"created": "Sat, 3 Sep 2022 20:35:23 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Feb 2023 19:01:17 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 20:08:53 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Zhang",
"Jingyuan",
""
],
[
"Wang",
"Ao",
""
],
[
"Ma",
"Xiaolong",
""
],
[
"Carver",
"Benjamin",
""
],
[
"Newman",
"Nicholas John",
""
],
[
"Anwar",
"Ali",
""
],
[
"Rupprecht",
"Lukas",
""
],
[
"Skourtis",
"Dimitrios",
""
],
[
"Tarasov",
"Vasily",
""
],
[
"Yan",
"Feng",
""
],
[
"Cheng",
"Yue",
""
]
] | TITLE: InfiniStore: Elastic Serverless Cloud Storage
ABSTRACT: Cloud object storage such as AWS S3 is cost-effective and highly elastic but
relatively slow, while high-performance cloud storage such as AWS ElastiCache
is expensive and provides limited elasticity. We present a new cloud storage
service called ServerlessMemory, which stores data using the memory of
serverless functions. ServerlessMemory employs a sliding-window-based memory
management strategy inspired by the garbage collection mechanisms used in the
programming language to effectively segregate hot/cold data and provides
fine-grained elasticity, good performance, and a pay-per-access cost model with
extremely low cost.
We then design and implement InfiniStore, a persistent and elastic cloud
storage system, which seamlessly couples the function-based ServerlessMemory
layer with a persistent, inexpensive cloud object store layer. InfiniStore
enables durability despite function failures using a fast parallel recovery
scheme built on the autoscaling functionality of a FaaS (Function-as-a-Service)
platform. We evaluate InfiniStore extensively using both microbenchmarking and
two real-world applications. Results show that InfiniStore has more performance
benefits for objects larger than 10 MB compared to AWS ElastiCache and Anna,
and InfiniStore achieves 26.25% and 97.24% tenant-side cost reduction compared
to InfiniCache and ElastiCache, respectively.
| no_new_dataset | 0.945751 |
2209.08162 | Sanbao Su | Sanbao Su, Yiming Li, Sihong He, Songyang Han, Chen Feng, Caiwen Ding,
Fei Miao | Uncertainty Quantification of Collaborative Detection for Self-Driving | This paper has been accepted by the 2023 IEEE International
Conference on Robotics and Automation (ICRA 2023) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sharing information between connected and autonomous vehicles (CAVs)
fundamentally improves the performance of collaborative object detection for
self-driving. However, CAVs still have uncertainties on object detection due to
practical challenges, which will affect the later modules in self-driving such
as planning and control. Hence, uncertainty quantification is crucial for
safety-critical systems such as CAVs. Our work is the first to estimate the
uncertainty of collaborative object detection. We propose a novel uncertainty
quantification method, called Double-M Quantification, which tailors a moving
block bootstrap (MBB) algorithm with direct modeling of the multivariant
Gaussian distribution of each corner of the bounding box. Our method captures
both the epistemic uncertainty and aleatoric uncertainty with one inference
pass based on the offline Double-M training process. And it can be used with
different collaborative object detectors. Through experiments on the
comprehensive collaborative perception dataset, we show that our Double-M
method achieves more than 4X improvement on uncertainty score and more than 3%
accuracy improvement, compared with the state-of-the-art uncertainty
quantification methods. Our code is public on
https://coperception.github.io/double-m-quantification.
| [
{
"version": "v1",
"created": "Fri, 16 Sep 2022 20:30:45 GMT"
},
{
"version": "v2",
"created": "Fri, 20 Jan 2023 02:57:46 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 01:06:39 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Su",
"Sanbao",
""
],
[
"Li",
"Yiming",
""
],
[
"He",
"Sihong",
""
],
[
"Han",
"Songyang",
""
],
[
"Feng",
"Chen",
""
],
[
"Ding",
"Caiwen",
""
],
[
"Miao",
"Fei",
""
]
] | TITLE: Uncertainty Quantification of Collaborative Detection for Self-Driving
ABSTRACT: Sharing information between connected and autonomous vehicles (CAVs)
fundamentally improves the performance of collaborative object detection for
self-driving. However, CAVs still have uncertainties on object detection due to
practical challenges, which will affect the later modules in self-driving such
as planning and control. Hence, uncertainty quantification is crucial for
safety-critical systems such as CAVs. Our work is the first to estimate the
uncertainty of collaborative object detection. We propose a novel uncertainty
quantification method, called Double-M Quantification, which tailors a moving
block bootstrap (MBB) algorithm with direct modeling of the multivariant
Gaussian distribution of each corner of the bounding box. Our method captures
both the epistemic uncertainty and aleatoric uncertainty with one inference
pass based on the offline Double-M training process. And it can be used with
different collaborative object detectors. Through experiments on the
comprehensive collaborative perception dataset, we show that our Double-M
method achieves more than 4X improvement on uncertainty score and more than 3%
accuracy improvement, compared with the state-of-the-art uncertainty
quantification methods. Our code is public on
https://coperception.github.io/double-m-quantification.
| no_new_dataset | 0.942295 |
2210.05174 | Tianheng Cheng | Tianheng Cheng and Xinggang Wang and Shaoyu Chen and Qian Zhang and
Wenyu Liu | BoxTeacher: Exploring High-Quality Pseudo Labels for Weakly Supervised
Instance Segmentation | Accepted to CVPR 2023. Code and models:
https://github.com/hustvl/BoxTeacher | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Labeling objects with pixel-wise segmentation requires a huge amount of human
labor compared to bounding boxes. Most existing methods for weakly supervised
instance segmentation focus on designing heuristic losses with priors from
bounding boxes. While, we find that box-supervised methods can produce some
fine segmentation masks and we wonder whether the detectors could learn from
these fine masks while ignoring low-quality masks. To answer this question, we
present BoxTeacher, an efficient and end-to-end training framework for
high-performance weakly supervised instance segmentation, which leverages a
sophisticated teacher to generate high-quality masks as pseudo labels.
Considering the massive noisy masks hurt the training, we present a mask-aware
confidence score to estimate the quality of pseudo masks and propose the
noise-aware pixel loss and noise-reduced affinity loss to adaptively optimize
the student with pseudo masks. Extensive experiments can demonstrate the
effectiveness of the proposed BoxTeacher. Without bells and whistles,
BoxTeacher remarkably achieves 35.0 mask AP and 36.5 mask AP with ResNet-50 and
ResNet-101 respectively on the challenging COCO dataset, which outperforms the
previous state-of-the-art methods by a significant margin and bridges the gap
between box-supervised and mask-supervised methods. The code and models will be
available at https://github.com/hustvl/BoxTeacher.
| [
{
"version": "v1",
"created": "Tue, 11 Oct 2022 06:23:30 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 05:17:43 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Cheng",
"Tianheng",
""
],
[
"Wang",
"Xinggang",
""
],
[
"Chen",
"Shaoyu",
""
],
[
"Zhang",
"Qian",
""
],
[
"Liu",
"Wenyu",
""
]
] | TITLE: BoxTeacher: Exploring High-Quality Pseudo Labels for Weakly Supervised
Instance Segmentation
ABSTRACT: Labeling objects with pixel-wise segmentation requires a huge amount of human
labor compared to bounding boxes. Most existing methods for weakly supervised
instance segmentation focus on designing heuristic losses with priors from
bounding boxes. While, we find that box-supervised methods can produce some
fine segmentation masks and we wonder whether the detectors could learn from
these fine masks while ignoring low-quality masks. To answer this question, we
present BoxTeacher, an efficient and end-to-end training framework for
high-performance weakly supervised instance segmentation, which leverages a
sophisticated teacher to generate high-quality masks as pseudo labels.
Considering the massive noisy masks hurt the training, we present a mask-aware
confidence score to estimate the quality of pseudo masks and propose the
noise-aware pixel loss and noise-reduced affinity loss to adaptively optimize
the student with pseudo masks. Extensive experiments can demonstrate the
effectiveness of the proposed BoxTeacher. Without bells and whistles,
BoxTeacher remarkably achieves 35.0 mask AP and 36.5 mask AP with ResNet-50 and
ResNet-101 respectively on the challenging COCO dataset, which outperforms the
previous state-of-the-art methods by a significant margin and bridges the gap
between box-supervised and mask-supervised methods. The code and models will be
available at https://github.com/hustvl/BoxTeacher.
| no_new_dataset | 0.946349 |
2210.10723 | Stefan Hegselmann | Stefan Hegselmann, Alejandro Buendia, Hunter Lang, Monica Agrawal,
Xiaoyi Jiang, David Sontag | TabLLM: Few-shot Classification of Tabular Data with Large Language
Models | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the application of large language models to zero-shot and few-shot
classification of tabular data. We prompt the large language model with a
serialization of the tabular data to a natural-language string, together with a
short description of the classification problem. In the few-shot setting, we
fine-tune the large language model using some labeled examples. We evaluate
several serialization methods including templates, table-to-text models, and
large language models. Despite its simplicity, we find that this technique
outperforms prior deep-learning-based tabular classification methods on several
benchmark datasets. In most cases, even zero-shot classification obtains
non-trivial performance, illustrating the method's ability to exploit prior
knowledge encoded in large language models. Unlike many deep learning methods
for tabular datasets, this approach is also competitive with strong traditional
baselines like gradient-boosted trees, especially in the very-few-shot setting.
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2022 17:08:13 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 16:36:53 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Hegselmann",
"Stefan",
""
],
[
"Buendia",
"Alejandro",
""
],
[
"Lang",
"Hunter",
""
],
[
"Agrawal",
"Monica",
""
],
[
"Jiang",
"Xiaoyi",
""
],
[
"Sontag",
"David",
""
]
] | TITLE: TabLLM: Few-shot Classification of Tabular Data with Large Language
Models
ABSTRACT: We study the application of large language models to zero-shot and few-shot
classification of tabular data. We prompt the large language model with a
serialization of the tabular data to a natural-language string, together with a
short description of the classification problem. In the few-shot setting, we
fine-tune the large language model using some labeled examples. We evaluate
several serialization methods including templates, table-to-text models, and
large language models. Despite its simplicity, we find that this technique
outperforms prior deep-learning-based tabular classification methods on several
benchmark datasets. In most cases, even zero-shot classification obtains
non-trivial performance, illustrating the method's ability to exploit prior
knowledge encoded in large language models. Unlike many deep learning methods
for tabular datasets, this approach is also competitive with strong traditional
baselines like gradient-boosted trees, especially in the very-few-shot setting.
| no_new_dataset | 0.949856 |
2210.12353 | Joshua Robinson | Joshua Robinson, Christopher Michael Rytting, David Wingate | Leveraging Large Language Models for Multiple Choice Question Answering | Accepted for ICLR 2023 | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While large language models (LLMs) like GPT-3 have achieved impressive
results on multiple choice question answering (MCQA) tasks in the zero, one,
and few-shot settings, they generally lag behind the MCQA state of the art
(SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks.
An LLM is conditioned on a question (without the associated answer options) and
its chosen option is the one assigned the highest probability after
normalization (for length, etc.). A more natural prompting approach is to
present the question and answer options to the LLM jointly and have it output
the symbol (e.g., "A") associated with its chosen answer option. This approach
allows the model to explicitly compare answer options, reduces computational
costs, and mitigates the effects of tokenization scheme and answer option
representations on answer selection. For the natural approach to be effective,
the LLM it is used with must be able to associate answer options with the
symbols that represent them. The LLM needs what we term multiple choice symbol
binding (MCSB) ability. This ability varies greatly by model. We show that a
model with high MCSB ability performs much better with the natural approach
than with the traditional approach across 20 diverse datasets and largely
closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been
previously underestimated.
| [
{
"version": "v1",
"created": "Sat, 22 Oct 2022 05:04:54 GMT"
},
{
"version": "v2",
"created": "Wed, 16 Nov 2022 04:27:18 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 00:52:56 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Robinson",
"Joshua",
""
],
[
"Rytting",
"Christopher Michael",
""
],
[
"Wingate",
"David",
""
]
] | TITLE: Leveraging Large Language Models for Multiple Choice Question Answering
ABSTRACT: While large language models (LLMs) like GPT-3 have achieved impressive
results on multiple choice question answering (MCQA) tasks in the zero, one,
and few-shot settings, they generally lag behind the MCQA state of the art
(SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks.
An LLM is conditioned on a question (without the associated answer options) and
its chosen option is the one assigned the highest probability after
normalization (for length, etc.). A more natural prompting approach is to
present the question and answer options to the LLM jointly and have it output
the symbol (e.g., "A") associated with its chosen answer option. This approach
allows the model to explicitly compare answer options, reduces computational
costs, and mitigates the effects of tokenization scheme and answer option
representations on answer selection. For the natural approach to be effective,
the LLM it is used with must be able to associate answer options with the
symbols that represent them. The LLM needs what we term multiple choice symbol
binding (MCSB) ability. This ability varies greatly by model. We show that a
model with high MCSB ability performs much better with the natural approach
than with the traditional approach across 20 diverse datasets and largely
closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been
previously underestimated.
| no_new_dataset | 0.948537 |
2210.16114 | Chuqin Geng | Chuqin Geng, Nham Le, Xiaojie Xu, Zhaoyue Wang, Arie Gurfinkel, Xujie
Si | Towards Reliable Neural Specifications | 19 pages, 16 figures | null | null | null | cs.LG cs.AI cs.LO cs.SE | http://creativecommons.org/licenses/by/4.0/ | Having reliable specifications is an unavoidable challenge in achieving
verifiable correctness, robustness, and interpretability of AI systems.
Existing specifications for neural networks are in the paradigm of data as
specification. That is, the local neighborhood centering around a reference
input is considered to be correct (or robust). While existing specifications
contribute to verifying adversarial robustness, a significant problem in many
research domains, our empirical study shows that those verified regions are
somewhat tight, and thus fail to allow verification of test set inputs, making
them impractical for some real-world applications. To this end, we propose a
new family of specifications called neural representation as specification,
which uses the intrinsic information of neural networks - neural activation
patterns (NAPs), rather than input data to specify the correctness and/or
robustness of neural network predictions. We present a simple statistical
approach to mining neural activation patterns. To show the effectiveness of
discovered NAPs, we formally verify several important properties, such as
various types of misclassifications will never happen for a given NAP, and
there is no ambiguity between different NAPs. We show that by using NAP, we can
verify a significant region of the input space, while still recalling 84% of
the data on MNIST. Moreover, we can push the verifiable bound to 10 times
larger on the CIFAR10 benchmark. Thus, we argue that NAPs can potentially be
used as a more reliable and extensible specification for neural network
verification.
| [
{
"version": "v1",
"created": "Fri, 28 Oct 2022 13:21:28 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Nov 2022 02:50:39 GMT"
},
{
"version": "v3",
"created": "Fri, 10 Feb 2023 00:00:21 GMT"
},
{
"version": "v4",
"created": "Tue, 28 Feb 2023 16:00:34 GMT"
},
{
"version": "v5",
"created": "Fri, 17 Mar 2023 16:35:00 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Geng",
"Chuqin",
""
],
[
"Le",
"Nham",
""
],
[
"Xu",
"Xiaojie",
""
],
[
"Wang",
"Zhaoyue",
""
],
[
"Gurfinkel",
"Arie",
""
],
[
"Si",
"Xujie",
""
]
] | TITLE: Towards Reliable Neural Specifications
ABSTRACT: Having reliable specifications is an unavoidable challenge in achieving
verifiable correctness, robustness, and interpretability of AI systems.
Existing specifications for neural networks are in the paradigm of data as
specification. That is, the local neighborhood centering around a reference
input is considered to be correct (or robust). While existing specifications
contribute to verifying adversarial robustness, a significant problem in many
research domains, our empirical study shows that those verified regions are
somewhat tight, and thus fail to allow verification of test set inputs, making
them impractical for some real-world applications. To this end, we propose a
new family of specifications called neural representation as specification,
which uses the intrinsic information of neural networks - neural activation
patterns (NAPs), rather than input data to specify the correctness and/or
robustness of neural network predictions. We present a simple statistical
approach to mining neural activation patterns. To show the effectiveness of
discovered NAPs, we formally verify several important properties, such as
various types of misclassifications will never happen for a given NAP, and
there is no ambiguity between different NAPs. We show that by using NAP, we can
verify a significant region of the input space, while still recalling 84% of
the data on MNIST. Moreover, we can push the verifiable bound to 10 times
larger on the CIFAR10 benchmark. Thus, we argue that NAPs can potentially be
used as a more reliable and extensible specification for neural network
verification.
| no_new_dataset | 0.941331 |
2211.08542 | Tianxing Xu | Tian-Xing Xu, Yuan-Chen Guo, Yu-Kun Lai, Song-Hai Zhang | CXTrack: Improving 3D Point Cloud Tracking with Contextual Information | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 3D single object tracking plays an essential role in many applications, such
as autonomous driving. It remains a challenging problem due to the large
appearance variation and the sparsity of points caused by occlusion and limited
sensor capabilities. Therefore, contextual information across two consecutive
frames is crucial for effective object tracking. However, points containing
such useful information are often overlooked and cropped out in existing
methods, leading to insufficient use of important contextual knowledge. To
address this issue, we propose CXTrack, a novel transformer-based network for
3D object tracking, which exploits ConteXtual information to improve the
tracking results. Specifically, we design a target-centric transformer network
that directly takes point features from two consecutive frames and the previous
bounding box as input to explore contextual information and implicitly
propagate target cues. To achieve accurate localization for objects of all
sizes, we propose a transformer-based localization head with a novel center
embedding module to distinguish the target from distractors. Extensive
experiments on three large-scale datasets, KITTI, nuScenes and Waymo Open
Dataset, show that CXTrack achieves state-of-the-art tracking performance while
running at 34 FPS.
| [
{
"version": "v1",
"created": "Sat, 12 Nov 2022 11:29:01 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 02:34:48 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Xu",
"Tian-Xing",
""
],
[
"Guo",
"Yuan-Chen",
""
],
[
"Lai",
"Yu-Kun",
""
],
[
"Zhang",
"Song-Hai",
""
]
] | TITLE: CXTrack: Improving 3D Point Cloud Tracking with Contextual Information
ABSTRACT: 3D single object tracking plays an essential role in many applications, such
as autonomous driving. It remains a challenging problem due to the large
appearance variation and the sparsity of points caused by occlusion and limited
sensor capabilities. Therefore, contextual information across two consecutive
frames is crucial for effective object tracking. However, points containing
such useful information are often overlooked and cropped out in existing
methods, leading to insufficient use of important contextual knowledge. To
address this issue, we propose CXTrack, a novel transformer-based network for
3D object tracking, which exploits ConteXtual information to improve the
tracking results. Specifically, we design a target-centric transformer network
that directly takes point features from two consecutive frames and the previous
bounding box as input to explore contextual information and implicitly
propagate target cues. To achieve accurate localization for objects of all
sizes, we propose a transformer-based localization head with a novel center
embedding module to distinguish the target from distractors. Extensive
experiments on three large-scale datasets, KITTI, nuScenes and Waymo Open
Dataset, show that CXTrack achieves state-of-the-art tracking performance while
running at 34 FPS.
| no_new_dataset | 0.946941 |
2211.10156 | Jiahao Chang | Jiahao Chang, Shuo Wang, Haiming Xu, Zehui Chen, Chenhongyi Yang, Feng
Zhao | DETRDistill: A Universal Knowledge Distillation Framework for
DETR-families | 10 pages, 6 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transformer-based detectors (DETRs) are becoming popular for their simple
framework, but the large model size and heavy time consumption hinder their
deployment in the real world. While knowledge distillation (KD) can be an
appealing technique to compress giant detectors into small ones for comparable
detection performance and low inference cost. Since DETRs formulate object
detection as a set prediction problem, existing KD methods designed for classic
convolution-based detectors may not be directly applicable. In this paper, we
propose DETRDistill, a novel knowledge distillation method dedicated to
DETR-families. Specifically, we first design a Hungarian-matching logits
distillation to encourage the student model to have the exact predictions as
that of teacher DETRs. Next, we propose a target-aware feature distillation to
help the student model learn from the object-centric features of the teacher
model. Finally, in order to improve the convergence rate of the student DETR,
we introduce a query-prior assignment distillation to speed up the student
model learning from well-trained queries and stable assignment of the teacher
model. Extensive experimental results on the COCO dataset validate the
effectiveness of our approach. Notably, DETRDistill consistently improves
various DETRs by more than 2.0 mAP, even surpassing their teacher models.
| [
{
"version": "v1",
"created": "Thu, 17 Nov 2022 13:35:11 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Nov 2022 07:40:11 GMT"
},
{
"version": "v3",
"created": "Wed, 15 Mar 2023 13:30:47 GMT"
},
{
"version": "v4",
"created": "Fri, 17 Mar 2023 03:20:06 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Chang",
"Jiahao",
""
],
[
"Wang",
"Shuo",
""
],
[
"Xu",
"Haiming",
""
],
[
"Chen",
"Zehui",
""
],
[
"Yang",
"Chenhongyi",
""
],
[
"Zhao",
"Feng",
""
]
] | TITLE: DETRDistill: A Universal Knowledge Distillation Framework for
DETR-families
ABSTRACT: Transformer-based detectors (DETRs) are becoming popular for their simple
framework, but the large model size and heavy time consumption hinder their
deployment in the real world. While knowledge distillation (KD) can be an
appealing technique to compress giant detectors into small ones for comparable
detection performance and low inference cost. Since DETRs formulate object
detection as a set prediction problem, existing KD methods designed for classic
convolution-based detectors may not be directly applicable. In this paper, we
propose DETRDistill, a novel knowledge distillation method dedicated to
DETR-families. Specifically, we first design a Hungarian-matching logits
distillation to encourage the student model to have the exact predictions as
that of teacher DETRs. Next, we propose a target-aware feature distillation to
help the student model learn from the object-centric features of the teacher
model. Finally, in order to improve the convergence rate of the student DETR,
we introduce a query-prior assignment distillation to speed up the student
model learning from well-trained queries and stable assignment of the teacher
model. Extensive experimental results on the COCO dataset validate the
effectiveness of our approach. Notably, DETRDistill consistently improves
various DETRs by more than 2.0 mAP, even surpassing their teacher models.
| no_new_dataset | 0.945801 |
2211.14563 | Arushi Goel | Arushi Goel, Basura Fernando, Frank Keller and Hakan Bilen | Who are you referring to? Coreference resolution in image narrations | 15 pages | null | null | null | cs.CV cs.CL | http://creativecommons.org/licenses/by/4.0/ | Coreference resolution aims to identify words and phrases which refer to same
entity in a text, a core task in natural language processing. In this paper, we
extend this task to resolving coreferences in long-form narrations of visual
scenes. First we introduce a new dataset with annotated coreference chains and
their bounding boxes, as most existing image-text datasets only contain short
sentences without coreferring expressions or labeled chains. We propose a new
technique that learns to identify coreference chains using weak supervision,
only from image-text pairs and a regularization using prior linguistic
knowledge. Our model yields large performance gains over several strong
baselines in resolving coreferences. We also show that coreference resolution
helps improving grounding narratives in images.
| [
{
"version": "v1",
"created": "Sat, 26 Nov 2022 13:33:42 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 15:12:13 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Goel",
"Arushi",
""
],
[
"Fernando",
"Basura",
""
],
[
"Keller",
"Frank",
""
],
[
"Bilen",
"Hakan",
""
]
] | TITLE: Who are you referring to? Coreference resolution in image narrations
ABSTRACT: Coreference resolution aims to identify words and phrases which refer to same
entity in a text, a core task in natural language processing. In this paper, we
extend this task to resolving coreferences in long-form narrations of visual
scenes. First we introduce a new dataset with annotated coreference chains and
their bounding boxes, as most existing image-text datasets only contain short
sentences without coreferring expressions or labeled chains. We propose a new
technique that learns to identify coreference chains using weak supervision,
only from image-text pairs and a regularization using prior linguistic
knowledge. Our model yields large performance gains over several strong
baselines in resolving coreferences. We also show that coreference resolution
helps improving grounding narratives in images.
| new_dataset | 0.954435 |
2212.00648 | Sagi Eppel | Manuel S. Drehwald, Sagi Eppel, Jolina Li, Han Hao, Alan Aspuru-Guzik | One-shot recognition of any material anywhere using contrastive learning
with physics-based rendering | for associated code and dataset, see
https://zenodo.org/record/7390166#.Y5ku6mHMJH4 or
https://e1.pcloud.link/publink/show?code=kZIiSQZCYU5M4HOvnQykql9jxF4h0KiC5MX
and https://icedrive.net/s/A13FWzZ8V2aP9T4ufGQ1N3fBZxDF | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Visual recognition of materials and their states is essential for
understanding most aspects of the world, from determining whether food is
cooked, metal is rusted, or a chemical reaction has occurred. However, current
image recognition methods are limited to specific classes and properties and
can't handle the vast number of material states in the world. To address this,
we present MatSim: the first dataset and benchmark for computer vision-based
recognition of similarities and transitions between materials and textures,
focusing on identifying any material under any conditions using one or a few
examples. The dataset contains synthetic and natural images. The synthetic
images were rendered using giant collections of textures, objects, and
environments generated by computer graphics artists. We use mixtures and
gradual transitions between materials to allow the system to learn cases with
smooth transitions between states (like gradually cooked food). We also render
images with materials inside transparent containers to support beverage and
chemistry lab use cases. We use this dataset to train a siamese net that
identifies the same material in different objects, mixtures, and environments.
The descriptor generated by this net can be used to identify the states of
materials and their subclasses using a single image. We also present the first
few-shot material recognition benchmark with images from a wide range of
fields, including the state of foods and drinks, types of grounds, and many
other use cases. We show that a net trained on the MatSim synthetic dataset
outperforms state-of-the-art models like Clip on the benchmark and also
achieves good results on other unsupervised material classification tasks.
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2022 16:49:53 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Dec 2022 02:12:27 GMT"
},
{
"version": "v3",
"created": "Mon, 13 Mar 2023 04:06:24 GMT"
},
{
"version": "v4",
"created": "Fri, 17 Mar 2023 04:40:59 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Drehwald",
"Manuel S.",
""
],
[
"Eppel",
"Sagi",
""
],
[
"Li",
"Jolina",
""
],
[
"Hao",
"Han",
""
],
[
"Aspuru-Guzik",
"Alan",
""
]
] | TITLE: One-shot recognition of any material anywhere using contrastive learning
with physics-based rendering
ABSTRACT: Visual recognition of materials and their states is essential for
understanding most aspects of the world, from determining whether food is
cooked, metal is rusted, or a chemical reaction has occurred. However, current
image recognition methods are limited to specific classes and properties and
can't handle the vast number of material states in the world. To address this,
we present MatSim: the first dataset and benchmark for computer vision-based
recognition of similarities and transitions between materials and textures,
focusing on identifying any material under any conditions using one or a few
examples. The dataset contains synthetic and natural images. The synthetic
images were rendered using giant collections of textures, objects, and
environments generated by computer graphics artists. We use mixtures and
gradual transitions between materials to allow the system to learn cases with
smooth transitions between states (like gradually cooked food). We also render
images with materials inside transparent containers to support beverage and
chemistry lab use cases. We use this dataset to train a siamese net that
identifies the same material in different objects, mixtures, and environments.
The descriptor generated by this net can be used to identify the states of
materials and their subclasses using a single image. We also present the first
few-shot material recognition benchmark with images from a wide range of
fields, including the state of foods and drinks, types of grounds, and many
other use cases. We show that a net trained on the MatSim synthetic dataset
outperforms state-of-the-art models like Clip on the benchmark and also
achieves good results on other unsupervised material classification tasks.
| new_dataset | 0.97231 |
2212.03010 | Honghui Yang | Honghui Yang and Tong He and Jiaheng Liu and Hua Chen and Boxi Wu and
Binbin Lin and Xiaofei He and Wanli Ouyang | GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Clouds | CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the tremendous progress of Masked Autoencoders (MAE) in developing
vision tasks such as image and video, exploring MAE in large-scale 3D point
clouds remains challenging due to the inherent irregularity. In contrast to
previous 3D MAE frameworks, which either design a complex decoder to infer
masked information from maintained regions or adopt sophisticated masking
strategies, we instead propose a much simpler paradigm. The core idea is to
apply a \textbf{G}enerative \textbf{D}ecoder for MAE (GD-MAE) to automatically
merges the surrounding context to restore the masked geometric knowledge in a
hierarchical fusion manner. In doing so, our approach is free from introducing
the heuristic design of decoders and enjoys the flexibility of exploring
various masking strategies. The corresponding part costs less than
\textbf{12\%} latency compared with conventional methods, while achieving
better performance. We demonstrate the efficacy of the proposed method on
several large-scale benchmarks: Waymo, KITTI, and ONCE. Consistent improvement
on downstream detection tasks illustrates strong robustness and generalization
capability. Not only our method reveals state-of-the-art results, but
remarkably, we achieve comparable accuracy even with \textbf{20\%} of the
labeled data on the Waymo dataset. Code will be released at
https://github.com/Nightmare-n/GD-MAE.
| [
{
"version": "v1",
"created": "Tue, 6 Dec 2022 14:32:55 GMT"
},
{
"version": "v2",
"created": "Wed, 7 Dec 2022 13:18:55 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 03:12:33 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yang",
"Honghui",
""
],
[
"He",
"Tong",
""
],
[
"Liu",
"Jiaheng",
""
],
[
"Chen",
"Hua",
""
],
[
"Wu",
"Boxi",
""
],
[
"Lin",
"Binbin",
""
],
[
"He",
"Xiaofei",
""
],
[
"Ouyang",
"Wanli",
""
]
] | TITLE: GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Clouds
ABSTRACT: Despite the tremendous progress of Masked Autoencoders (MAE) in developing
vision tasks such as image and video, exploring MAE in large-scale 3D point
clouds remains challenging due to the inherent irregularity. In contrast to
previous 3D MAE frameworks, which either design a complex decoder to infer
masked information from maintained regions or adopt sophisticated masking
strategies, we instead propose a much simpler paradigm. The core idea is to
apply a \textbf{G}enerative \textbf{D}ecoder for MAE (GD-MAE) to automatically
merges the surrounding context to restore the masked geometric knowledge in a
hierarchical fusion manner. In doing so, our approach is free from introducing
the heuristic design of decoders and enjoys the flexibility of exploring
various masking strategies. The corresponding part costs less than
\textbf{12\%} latency compared with conventional methods, while achieving
better performance. We demonstrate the efficacy of the proposed method on
several large-scale benchmarks: Waymo, KITTI, and ONCE. Consistent improvement
on downstream detection tasks illustrates strong robustness and generalization
capability. Not only our method reveals state-of-the-art results, but
remarkably, we achieve comparable accuracy even with \textbf{20\%} of the
labeled data on the Waymo dataset. Code will be released at
https://github.com/Nightmare-n/GD-MAE.
| no_new_dataset | 0.946695 |
2212.05993 | Jiabao Lei | Jiabao Lei, Jiapeng Tang, Kui Jia | RGBD2: Generative Scene Synthesis via Incremental View Inpainting using
RGBD Diffusion Models | CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address the challenge of recovering an underlying scene geometry and
colors from a sparse set of RGBD view observations. In this work, we present a
new solution termed RGBD$^2$ that sequentially generates novel RGBD views along
a camera trajectory, and the scene geometry is simply the fusion result of
these views. More specifically, we maintain an intermediate surface mesh used
for rendering new RGBD views, which subsequently becomes complete by an
inpainting network; each rendered RGBD view is later back-projected as a
partial surface and is supplemented into the intermediate mesh. The use of
intermediate mesh and camera projection helps solve the tough problem of
multi-view inconsistency. We practically implement the RGBD inpainting network
as a versatile RGBD diffusion model, which is previously used for 2D generative
modeling; we make a modification to its reverse diffusion process to enable our
use. We evaluate our approach on the task of 3D scene synthesis from sparse
RGBD inputs; extensive experiments on the ScanNet dataset demonstrate the
superiority of our approach over existing ones. Project page:
https://jblei.site/proj/rgbd-diffusion.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 15:50:00 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 07:27:15 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Lei",
"Jiabao",
""
],
[
"Tang",
"Jiapeng",
""
],
[
"Jia",
"Kui",
""
]
] | TITLE: RGBD2: Generative Scene Synthesis via Incremental View Inpainting using
RGBD Diffusion Models
ABSTRACT: We address the challenge of recovering an underlying scene geometry and
colors from a sparse set of RGBD view observations. In this work, we present a
new solution termed RGBD$^2$ that sequentially generates novel RGBD views along
a camera trajectory, and the scene geometry is simply the fusion result of
these views. More specifically, we maintain an intermediate surface mesh used
for rendering new RGBD views, which subsequently becomes complete by an
inpainting network; each rendered RGBD view is later back-projected as a
partial surface and is supplemented into the intermediate mesh. The use of
intermediate mesh and camera projection helps solve the tough problem of
multi-view inconsistency. We practically implement the RGBD inpainting network
as a versatile RGBD diffusion model, which is previously used for 2D generative
modeling; we make a modification to its reverse diffusion process to enable our
use. We evaluate our approach on the task of 3D scene synthesis from sparse
RGBD inputs; extensive experiments on the ScanNet dataset demonstrate the
superiority of our approach over existing ones. Project page:
https://jblei.site/proj/rgbd-diffusion.
| no_new_dataset | 0.950319 |
2212.08756 | Zhenyuan Lu | Zhenyuan Lu | Multi-Scales Data Augmentation Approach In Natural Language Inference
For Artifacts Mitigation And Pre-Trained Model Optimization | null | null | null | null | cs.CL stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning models can reach high performance on benchmark natural
language processing (NLP) datasets but fail in more challenging settings. We
study this issue when a pre-trained model learns dataset artifacts in natural
language inference (NLI), the topic of studying the logical relationship
between a pair of text sequences. We provide a variety of techniques for
analyzing and locating dataset artifacts inside the crowdsourced Stanford
Natural Language Inference (SNLI) corpus. We study the stylistic pattern of
dataset artifacts in the SNLI. To mitigate dataset artifacts, we employ a
unique multi-scale data augmentation technique with two distinct frameworks: a
behavioral testing checklist at the sentence level and lexical synonym criteria
at the word level. Specifically, our combination method enhances our model's
resistance to perturbation testing, enabling it to continuously outperform the
pre-trained baseline.
| [
{
"version": "v1",
"created": "Fri, 16 Dec 2022 23:37:44 GMT"
},
{
"version": "v2",
"created": "Tue, 20 Dec 2022 20:42:24 GMT"
},
{
"version": "v3",
"created": "Thu, 5 Jan 2023 03:02:19 GMT"
},
{
"version": "v4",
"created": "Thu, 16 Mar 2023 21:31:25 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Lu",
"Zhenyuan",
""
]
] | TITLE: Multi-Scales Data Augmentation Approach In Natural Language Inference
For Artifacts Mitigation And Pre-Trained Model Optimization
ABSTRACT: Machine learning models can reach high performance on benchmark natural
language processing (NLP) datasets but fail in more challenging settings. We
study this issue when a pre-trained model learns dataset artifacts in natural
language inference (NLI), the topic of studying the logical relationship
between a pair of text sequences. We provide a variety of techniques for
analyzing and locating dataset artifacts inside the crowdsourced Stanford
Natural Language Inference (SNLI) corpus. We study the stylistic pattern of
dataset artifacts in the SNLI. To mitigate dataset artifacts, we employ a
unique multi-scale data augmentation technique with two distinct frameworks: a
behavioral testing checklist at the sentence level and lexical synonym criteria
at the word level. Specifically, our combination method enhances our model's
resistance to perturbation testing, enabling it to continuously outperform the
pre-trained baseline.
| no_new_dataset | 0.94699 |
2212.11375 | Jorge F. Lazo | Jorge F. Lazo, Benoit Rosa, Michele Catellani, Matteo Fontana,
Francesco A. Mistretta, Gennaro Musi, Ottavio de Cobelli, Michel de Mathelin
and Elena De Momi | Semi-supervised Bladder Tissue Classification in Multi-Domain Endoscopic
Images | Title and abstract updated. Typos corrected | null | null | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Objective: Accurate visual classification of bladder tissue during
Trans-Urethral Resection of Bladder Tumor (TURBT) procedures is essential to
improve early cancer diagnosis and treatment. During TURBT interventions, White
Light Imaging (WLI) and Narrow Band Imaging (NBI) techniques are used for
lesion detection. Each imaging technique provides diverse visual information
that allows clinicians to identify and classify cancerous lesions. Computer
vision methods that use both imaging techniques could improve endoscopic
diagnosis. We address the challenge of tissue classification when annotations
are available only in one domain, in our case WLI, and the endoscopic images
correspond to an unpaired dataset, i.e. there is no exact equivalent for every
image in both NBI and WLI domains. Method: We propose a semi-surprised
Generative Adversarial Network (GAN)-based method composed of three main
components: a teacher network trained on the labeled WLI data; a
cycle-consistency GAN to perform unpaired image-to-image translation, and a
multi-input student network. To ensure the quality of the synthetic images
generated by the proposed GAN we perform a detailed quantitative, and
qualitative analysis with the help of specialists. Conclusion: The overall
average classification accuracy, precision, and recall obtained with the
proposed method for tissue classification are 0.90, 0.88, and 0.89
respectively, while the same metrics obtained in the unlabeled domain (NBI) are
0.92, 0.64, and 0.94 respectively. The quality of the generated images is
reliable enough to deceive specialists. Significance: This study shows the
potential of using semi-supervised GAN-based bladder tissue classification when
annotations are limited in multi-domain data. The dataset is available at
https://zenodo.org/record/7741476#.ZBQUK7TMJ6k
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2022 21:32:36 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 07:43:42 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Lazo",
"Jorge F.",
""
],
[
"Rosa",
"Benoit",
""
],
[
"Catellani",
"Michele",
""
],
[
"Fontana",
"Matteo",
""
],
[
"Mistretta",
"Francesco A.",
""
],
[
"Musi",
"Gennaro",
""
],
[
"de Cobelli",
"Ottavio",
""
],
[
"de Mathelin",
"Michel",
""
],
[
"De Momi",
"Elena",
""
]
] | TITLE: Semi-supervised Bladder Tissue Classification in Multi-Domain Endoscopic
Images
ABSTRACT: Objective: Accurate visual classification of bladder tissue during
Trans-Urethral Resection of Bladder Tumor (TURBT) procedures is essential to
improve early cancer diagnosis and treatment. During TURBT interventions, White
Light Imaging (WLI) and Narrow Band Imaging (NBI) techniques are used for
lesion detection. Each imaging technique provides diverse visual information
that allows clinicians to identify and classify cancerous lesions. Computer
vision methods that use both imaging techniques could improve endoscopic
diagnosis. We address the challenge of tissue classification when annotations
are available only in one domain, in our case WLI, and the endoscopic images
correspond to an unpaired dataset, i.e. there is no exact equivalent for every
image in both NBI and WLI domains. Method: We propose a semi-surprised
Generative Adversarial Network (GAN)-based method composed of three main
components: a teacher network trained on the labeled WLI data; a
cycle-consistency GAN to perform unpaired image-to-image translation, and a
multi-input student network. To ensure the quality of the synthetic images
generated by the proposed GAN we perform a detailed quantitative, and
qualitative analysis with the help of specialists. Conclusion: The overall
average classification accuracy, precision, and recall obtained with the
proposed method for tissue classification are 0.90, 0.88, and 0.89
respectively, while the same metrics obtained in the unlabeled domain (NBI) are
0.92, 0.64, and 0.94 respectively. The quality of the generated images is
reliable enough to deceive specialists. Significance: This study shows the
potential of using semi-supervised GAN-based bladder tissue classification when
annotations are limited in multi-domain data. The dataset is available at
https://zenodo.org/record/7741476#.ZBQUK7TMJ6k
| no_new_dataset | 0.953837 |
2212.11565 | Jay Zhangjie Wu | Jay Zhangjie Wu, Yixiao Ge, Xintao Wang, Weixian Lei, Yuchao Gu, Yufei
Shi, Wynne Hsu, Ying Shan, Xiaohu Qie, Mike Zheng Shou | Tune-A-Video: One-Shot Tuning of Image Diffusion Models for
Text-to-Video Generation | Preprint | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To replicate the success of text-to-image (T2I) generation, recent works
employ large-scale video datasets to train a text-to-video (T2V) generator.
Despite their promising results, such paradigm is computationally expensive. In
this work, we propose a new T2V generation setting$\unicode{x2014}$One-Shot
Video Tuning, where only one text-video pair is presented. Our model is built
on state-of-the-art T2I diffusion models pre-trained on massive image data. We
make two key observations: 1) T2I models can generate still images that
represent verb terms; 2) extending T2I models to generate multiple images
concurrently exhibits surprisingly good content consistency. To further learn
continuous motion, we introduce Tune-A-Video, which involves a tailored
spatio-temporal attention mechanism and an efficient one-shot tuning strategy.
At inference, we employ DDIM inversion to provide structure guidance for
sampling. Extensive qualitative and numerical experiments demonstrate the
remarkable ability of our method across various applications.
| [
{
"version": "v1",
"created": "Thu, 22 Dec 2022 09:43:36 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 17:28:04 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Wu",
"Jay Zhangjie",
""
],
[
"Ge",
"Yixiao",
""
],
[
"Wang",
"Xintao",
""
],
[
"Lei",
"Weixian",
""
],
[
"Gu",
"Yuchao",
""
],
[
"Shi",
"Yufei",
""
],
[
"Hsu",
"Wynne",
""
],
[
"Shan",
"Ying",
""
],
[
"Qie",
"Xiaohu",
""
],
[
"Shou",
"Mike Zheng",
""
]
] | TITLE: Tune-A-Video: One-Shot Tuning of Image Diffusion Models for
Text-to-Video Generation
ABSTRACT: To replicate the success of text-to-image (T2I) generation, recent works
employ large-scale video datasets to train a text-to-video (T2V) generator.
Despite their promising results, such paradigm is computationally expensive. In
this work, we propose a new T2V generation setting$\unicode{x2014}$One-Shot
Video Tuning, where only one text-video pair is presented. Our model is built
on state-of-the-art T2I diffusion models pre-trained on massive image data. We
make two key observations: 1) T2I models can generate still images that
represent verb terms; 2) extending T2I models to generate multiple images
concurrently exhibits surprisingly good content consistency. To further learn
continuous motion, we introduce Tune-A-Video, which involves a tailored
spatio-temporal attention mechanism and an efficient one-shot tuning strategy.
At inference, we employ DDIM inversion to provide structure guidance for
sampling. Extensive qualitative and numerical experiments demonstrate the
remarkable ability of our method across various applications.
| no_new_dataset | 0.948537 |
2212.11596 | Oriol Barbany | Oriol Barbany, Adri\`a Colom\'e, Carme Torras | Deformable Surface Reconstruction via Riemannian Metric Preservation | This paper is under consideration at Computer Vision and Image
Understanding | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Estimating the pose of an object from a monocular image is an inverse problem
fundamental in computer vision. The ill-posed nature of this problem requires
incorporating deformation priors to solve it. In practice, many materials do
not perceptibly shrink or extend when manipulated, constituting a powerful and
well-known prior. Mathematically, this translates to the preservation of the
Riemannian metric. Neural networks offer the perfect playground to solve the
surface reconstruction problem as they can approximate surfaces with arbitrary
precision and allow the computation of differential geometry quantities. This
paper presents an approach to inferring continuous deformable surfaces from a
sequence of images, which is benchmarked against several techniques and obtains
state-of-the-art performance without the need for offline training.
| [
{
"version": "v1",
"created": "Thu, 22 Dec 2022 10:45:08 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 10:11:52 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Barbany",
"Oriol",
""
],
[
"Colomé",
"Adrià",
""
],
[
"Torras",
"Carme",
""
]
] | TITLE: Deformable Surface Reconstruction via Riemannian Metric Preservation
ABSTRACT: Estimating the pose of an object from a monocular image is an inverse problem
fundamental in computer vision. The ill-posed nature of this problem requires
incorporating deformation priors to solve it. In practice, many materials do
not perceptibly shrink or extend when manipulated, constituting a powerful and
well-known prior. Mathematically, this translates to the preservation of the
Riemannian metric. Neural networks offer the perfect playground to solve the
surface reconstruction problem as they can approximate surfaces with arbitrary
precision and allow the computation of differential geometry quantities. This
paper presents an approach to inferring continuous deformable surfaces from a
sequence of images, which is benchmarked against several techniques and obtains
state-of-the-art performance without the need for offline training.
| no_new_dataset | 0.95096 |
2212.12977 | Mengzhao Chen | Mengzhao Chen, Mingbao Lin, ZhiHang Lin, Yuxin Zhang, Fei Chao,
Rongrong Ji | SMMix: Self-Motivated Image Mixing for Vision Transformers | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | CutMix is a vital augmentation strategy that determines the performance and
generalization ability of vision transformers (ViTs). However, the
inconsistency between the mixed images and the corresponding labels harms its
efficacy. Existing CutMix variants tackle this problem by generating more
consistent mixed images or more precise mixed labels, but inevitably introduce
heavy training overhead or require extra information, undermining ease of use.
To this end, we propose an novel and effective Self-Motivated image Mixing
method (SMMix), which motivates both image and label enhancement by the model
under training itself. Specifically, we propose a max-min attention region
mixing approach that enriches the attention-focused objects in the mixed
images. Then, we introduce a fine-grained label assignment technique that
co-trains the output tokens of mixed images with fine-grained supervision.
Moreover, we devise a novel feature consistency constraint to align features
from mixed and unmixed images. Due to the subtle designs of the self-motivated
paradigm, our SMMix is significant in its smaller training overhead and better
performance than other CutMix variants. In particular, SMMix improves the
accuracy of DeiT-T/S/B, CaiT-XXS-24/36, and PVT-T/S/M/L by more than +1% on
ImageNet-1k. The generalization capability of our method is also demonstrated
on downstream tasks and out-of-distribution datasets. Our project is
anonymously available at https://github.com/ChenMnZ/SMMix.
| [
{
"version": "v1",
"created": "Mon, 26 Dec 2022 00:19:39 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 02:52:25 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Chen",
"Mengzhao",
""
],
[
"Lin",
"Mingbao",
""
],
[
"Lin",
"ZhiHang",
""
],
[
"Zhang",
"Yuxin",
""
],
[
"Chao",
"Fei",
""
],
[
"Ji",
"Rongrong",
""
]
] | TITLE: SMMix: Self-Motivated Image Mixing for Vision Transformers
ABSTRACT: CutMix is a vital augmentation strategy that determines the performance and
generalization ability of vision transformers (ViTs). However, the
inconsistency between the mixed images and the corresponding labels harms its
efficacy. Existing CutMix variants tackle this problem by generating more
consistent mixed images or more precise mixed labels, but inevitably introduce
heavy training overhead or require extra information, undermining ease of use.
To this end, we propose an novel and effective Self-Motivated image Mixing
method (SMMix), which motivates both image and label enhancement by the model
under training itself. Specifically, we propose a max-min attention region
mixing approach that enriches the attention-focused objects in the mixed
images. Then, we introduce a fine-grained label assignment technique that
co-trains the output tokens of mixed images with fine-grained supervision.
Moreover, we devise a novel feature consistency constraint to align features
from mixed and unmixed images. Due to the subtle designs of the self-motivated
paradigm, our SMMix is significant in its smaller training overhead and better
performance than other CutMix variants. In particular, SMMix improves the
accuracy of DeiT-T/S/B, CaiT-XXS-24/36, and PVT-T/S/M/L by more than +1% on
ImageNet-1k. The generalization capability of our method is also demonstrated
on downstream tasks and out-of-distribution datasets. Our project is
anonymously available at https://github.com/ChenMnZ/SMMix.
| no_new_dataset | 0.943971 |
2302.02614 | Kuan Xu | Kuan Xu, Kuo Yang, Hanyang Dong, Xinyan Wang, Jian Yu, Xuezhong Zhou | A Pre-training Framework for Knowledge Graph Completion | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge graph completion (KGC) is one of the effective methods to identify
new facts in knowledge graph. Except for a few methods based on graph network,
most of KGC methods trend to be trained based on independent triples, while are
difficult to take a full account of the information of global network
connection contained in knowledge network. To address these issues, in this
study, we propose a simple and effective Network-based Pre-training framework
for knowledge graph completion (termed NetPeace), which takes into account the
information of global network connection and local triple relationships in
knowledge graph. Experiments show that in NetPeace framework, multiple KGC
models yields consistent and significant improvements on benchmarks (e.g.,
36.45% Hits@1 and 27.40% MRR improvements for TuckER on FB15k-237), especially
dense knowledge graph. On the challenging low-resource task, NetPeace that
benefits from the global features of KG achieves higher performance (104.03%
MRR and 143.89% Hit@1 improvements at most) than original models.
| [
{
"version": "v1",
"created": "Mon, 6 Feb 2023 08:23:01 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 07:04:03 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Xu",
"Kuan",
""
],
[
"Yang",
"Kuo",
""
],
[
"Dong",
"Hanyang",
""
],
[
"Wang",
"Xinyan",
""
],
[
"Yu",
"Jian",
""
],
[
"Zhou",
"Xuezhong",
""
]
] | TITLE: A Pre-training Framework for Knowledge Graph Completion
ABSTRACT: Knowledge graph completion (KGC) is one of the effective methods to identify
new facts in knowledge graph. Except for a few methods based on graph network,
most of KGC methods trend to be trained based on independent triples, while are
difficult to take a full account of the information of global network
connection contained in knowledge network. To address these issues, in this
study, we propose a simple and effective Network-based Pre-training framework
for knowledge graph completion (termed NetPeace), which takes into account the
information of global network connection and local triple relationships in
knowledge graph. Experiments show that in NetPeace framework, multiple KGC
models yields consistent and significant improvements on benchmarks (e.g.,
36.45% Hits@1 and 27.40% MRR improvements for TuckER on FB15k-237), especially
dense knowledge graph. On the challenging low-resource task, NetPeace that
benefits from the global features of KG achieves higher performance (104.03%
MRR and 143.89% Hit@1 improvements at most) than original models.
| no_new_dataset | 0.944689 |
2302.06175 | Jannik Z\"urn | Martin B\"uchner, Jannik Z\"urn, Ion-George Todoran, Abhinav Valada,
Wolfram Burgard | Learning and Aggregating Lane Graphs for Urban Automated Driving | 22 pages, 17 figures | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lane graph estimation is an essential and highly challenging task in
automated driving and HD map learning. Existing methods using either onboard or
aerial imagery struggle with complex lane topologies, out-of-distribution
scenarios, or significant occlusions in the image space. Moreover, merging
overlapping lane graphs to obtain consistent large-scale graphs remains
difficult. To overcome these challenges, we propose a novel bottom-up approach
to lane graph estimation from aerial imagery that aggregates multiple
overlapping graphs into a single consistent graph. Due to its modular design,
our method allows us to address two complementary tasks: predicting
ego-respective successor lane graphs from arbitrary vehicle positions using a
graph neural network and aggregating these predictions into a consistent global
lane graph. Extensive experiments on a large-scale lane graph dataset
demonstrate that our approach yields highly accurate lane graphs, even in
regions with severe occlusions. The presented approach to graph aggregation
proves to eliminate inconsistent predictions while increasing the overall graph
quality. We make our large-scale urban lane graph dataset and code publicly
available at http://urbanlanegraph.cs.uni-freiburg.de.
| [
{
"version": "v1",
"created": "Mon, 13 Feb 2023 08:23:35 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 12:38:53 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Büchner",
"Martin",
""
],
[
"Zürn",
"Jannik",
""
],
[
"Todoran",
"Ion-George",
""
],
[
"Valada",
"Abhinav",
""
],
[
"Burgard",
"Wolfram",
""
]
] | TITLE: Learning and Aggregating Lane Graphs for Urban Automated Driving
ABSTRACT: Lane graph estimation is an essential and highly challenging task in
automated driving and HD map learning. Existing methods using either onboard or
aerial imagery struggle with complex lane topologies, out-of-distribution
scenarios, or significant occlusions in the image space. Moreover, merging
overlapping lane graphs to obtain consistent large-scale graphs remains
difficult. To overcome these challenges, we propose a novel bottom-up approach
to lane graph estimation from aerial imagery that aggregates multiple
overlapping graphs into a single consistent graph. Due to its modular design,
our method allows us to address two complementary tasks: predicting
ego-respective successor lane graphs from arbitrary vehicle positions using a
graph neural network and aggregating these predictions into a consistent global
lane graph. Extensive experiments on a large-scale lane graph dataset
demonstrate that our approach yields highly accurate lane graphs, even in
regions with severe occlusions. The presented approach to graph aggregation
proves to eliminate inconsistent predictions while increasing the overall graph
quality. We make our large-scale urban lane graph dataset and code publicly
available at http://urbanlanegraph.cs.uni-freiburg.de.
| new_dataset | 0.963712 |
2302.10803 | Steeven Janny | Steeven Janny, Aur\'elien B\'eneteau, Madiha Nadri, Julie Digne,
Nicolas Thome, Christian Wolf | Eagle: Large-Scale Learning of Turbulent Fluid Dynamics with Mesh
Transformers | Published as a conference paper at ICLR 2023 | International Conference on Learning Representation (ICLR) 2023 | null | null | cs.LG cs.AI physics.flu-dyn | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Estimating fluid dynamics is classically done through the simulation and
integration of numerical models solving the Navier-Stokes equations, which is
computationally complex and time-consuming even on high-end hardware. This is a
notoriously hard problem to solve, which has recently been addressed with
machine learning, in particular graph neural networks (GNN) and variants
trained and evaluated on datasets of static objects in static scenes with fixed
geometry. We attempt to go beyond existing work in complexity and introduce a
new model, method and benchmark. We propose EAGLE, a large-scale dataset of 1.1
million 2D meshes resulting from simulations of unsteady fluid dynamics caused
by a moving flow source interacting with nonlinear scene structure, comprised
of 600 different scenes of three different types. To perform future forecasting
of pressure and velocity on the challenging EAGLE dataset, we introduce a new
mesh transformer. It leverages node clustering, graph pooling and global
attention to learn long-range dependencies between spatially distant data
points without needing a large number of iterations, as existing GNN methods
do. We show that our transformer outperforms state-of-the-art performance on,
both, existing synthetic and real datasets and on EAGLE. Finally, we highlight
that our approach learns to attend to airflow, integrating complex information
in a single iteration.
| [
{
"version": "v1",
"created": "Thu, 16 Feb 2023 12:59:08 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 08:38:35 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Janny",
"Steeven",
""
],
[
"Béneteau",
"Aurélien",
""
],
[
"Nadri",
"Madiha",
""
],
[
"Digne",
"Julie",
""
],
[
"Thome",
"Nicolas",
""
],
[
"Wolf",
"Christian",
""
]
] | TITLE: Eagle: Large-Scale Learning of Turbulent Fluid Dynamics with Mesh
Transformers
ABSTRACT: Estimating fluid dynamics is classically done through the simulation and
integration of numerical models solving the Navier-Stokes equations, which is
computationally complex and time-consuming even on high-end hardware. This is a
notoriously hard problem to solve, which has recently been addressed with
machine learning, in particular graph neural networks (GNN) and variants
trained and evaluated on datasets of static objects in static scenes with fixed
geometry. We attempt to go beyond existing work in complexity and introduce a
new model, method and benchmark. We propose EAGLE, a large-scale dataset of 1.1
million 2D meshes resulting from simulations of unsteady fluid dynamics caused
by a moving flow source interacting with nonlinear scene structure, comprised
of 600 different scenes of three different types. To perform future forecasting
of pressure and velocity on the challenging EAGLE dataset, we introduce a new
mesh transformer. It leverages node clustering, graph pooling and global
attention to learn long-range dependencies between spatially distant data
points without needing a large number of iterations, as existing GNN methods
do. We show that our transformer outperforms state-of-the-art performance on,
both, existing synthetic and real datasets and on EAGLE. Finally, we highlight
that our approach learns to attend to airflow, integrating complex information
in a single iteration.
| no_new_dataset | 0.937669 |
2303.02735 | Yewon Lim | Dohyun Ham, Jaeyeop Jeong, June-Kyoo Park, Raehyeon Jeong, Seungmin
Jeon, Hyeongjun Jeon, and Yewon Lim | Scalable Object Detection on Embedded Devices Using Weight Pruning and
Singular Value Decomposition | 8 pages, 3 figures. A report of the project done as part of the
Yonsei-Roboin project for the 2nd semester, 2022 | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper presents a method for optimizing object detection models by
combining weight pruning and singular value decomposition (SVD). The proposed
method was evaluated on a custom dataset of street work images obtained from
https://universe.roboflow.com/roboflow-100/street-work. The dataset consists of
611 training images, 175 validation images, and 87 test images with 7 classes.
We compared the performance of the optimized models with the original
unoptimized model in terms of frame rate, mean average precision (mAP@50), and
weight size. The results show that the weight pruning + SVD model achieved a
0.724 mAP@50 with a frame rate of 1.48 FPS and a weight size of 12.1 MB,
outperforming the original model (0.717 mAP@50, 1.50 FPS, and 12.3 MB).
Precision-recall curves were also plotted for all models. Our work demonstrates
that the proposed method can effectively optimize object detection models while
balancing accuracy, speed, and model size.
| [
{
"version": "v1",
"created": "Sun, 5 Mar 2023 18:02:54 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 15:49:09 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ham",
"Dohyun",
""
],
[
"Jeong",
"Jaeyeop",
""
],
[
"Park",
"June-Kyoo",
""
],
[
"Jeong",
"Raehyeon",
""
],
[
"Jeon",
"Seungmin",
""
],
[
"Jeon",
"Hyeongjun",
""
],
[
"Lim",
"Yewon",
""
]
] | TITLE: Scalable Object Detection on Embedded Devices Using Weight Pruning and
Singular Value Decomposition
ABSTRACT: This paper presents a method for optimizing object detection models by
combining weight pruning and singular value decomposition (SVD). The proposed
method was evaluated on a custom dataset of street work images obtained from
https://universe.roboflow.com/roboflow-100/street-work. The dataset consists of
611 training images, 175 validation images, and 87 test images with 7 classes.
We compared the performance of the optimized models with the original
unoptimized model in terms of frame rate, mean average precision (mAP@50), and
weight size. The results show that the weight pruning + SVD model achieved a
0.724 mAP@50 with a frame rate of 1.48 FPS and a weight size of 12.1 MB,
outperforming the original model (0.717 mAP@50, 1.50 FPS, and 12.3 MB).
Precision-recall curves were also plotted for all models. Our work demonstrates
that the proposed method can effectively optimize object detection models while
balancing accuracy, speed, and model size.
| new_dataset | 0.958924 |
2303.03630 | Yingxiao Du | Yingxiao Du, Jianxin Wu | No One Left Behind: Improving the Worst Categories in Long-Tailed
Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unlike the case when using a balanced training dataset, the per-class recall
(i.e., accuracy) of neural networks trained with an imbalanced dataset are
known to vary a lot from category to category. The convention in long-tailed
recognition is to manually split all categories into three subsets and report
the average accuracy within each subset. We argue that under such an evaluation
setting, some categories are inevitably sacrificed. On one hand, focusing on
the average accuracy on a balanced test set incurs little penalty even if some
worst performing categories have zero accuracy. On the other hand, classes in
the "Few" subset do not necessarily perform worse than those in the "Many" or
"Medium" subsets. We therefore advocate to focus more on improving the lowest
recall among all categories and the harmonic mean of all recall values.
Specifically, we propose a simple plug-in method that is applicable to a wide
range of methods. By simply re-training the classifier of an existing
pre-trained model with our proposed loss function and using an optional
ensemble trick that combines the predictions of the two classifiers, we achieve
a more uniform distribution of recall values across categories, which leads to
a higher harmonic mean accuracy while the (arithmetic) average accuracy is
still high. The effectiveness of our method is justified on widely used
benchmark datasets.
| [
{
"version": "v1",
"created": "Tue, 7 Mar 2023 03:24:54 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 07:12:37 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Du",
"Yingxiao",
""
],
[
"Wu",
"Jianxin",
""
]
] | TITLE: No One Left Behind: Improving the Worst Categories in Long-Tailed
Learning
ABSTRACT: Unlike the case when using a balanced training dataset, the per-class recall
(i.e., accuracy) of neural networks trained with an imbalanced dataset are
known to vary a lot from category to category. The convention in long-tailed
recognition is to manually split all categories into three subsets and report
the average accuracy within each subset. We argue that under such an evaluation
setting, some categories are inevitably sacrificed. On one hand, focusing on
the average accuracy on a balanced test set incurs little penalty even if some
worst performing categories have zero accuracy. On the other hand, classes in
the "Few" subset do not necessarily perform worse than those in the "Many" or
"Medium" subsets. We therefore advocate to focus more on improving the lowest
recall among all categories and the harmonic mean of all recall values.
Specifically, we propose a simple plug-in method that is applicable to a wide
range of methods. By simply re-training the classifier of an existing
pre-trained model with our proposed loss function and using an optional
ensemble trick that combines the predictions of the two classifiers, we achieve
a more uniform distribution of recall values across categories, which leads to
a higher harmonic mean accuracy while the (arithmetic) average accuracy is
still high. The effectiveness of our method is justified on widely used
benchmark datasets.
| no_new_dataset | 0.949482 |
2303.03770 | Pietro Morerio | Mattia Litrico, Alessio Del Bue, Pietro Morerio | Guiding Pseudo-labels with Uncertainty Estimation for Source-free
Unsupervised Domain Adaptation | To be published in Proceedings of the IEEE/CVF Conference on Computer
Vision and Pattern Recognition (CVPR) 2023 | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Standard Unsupervised Domain Adaptation (UDA) methods assume the availability
of both source and target data during the adaptation. In this work, we
investigate Source-free Unsupervised Domain Adaptation (SF-UDA), a specific
case of UDA where a model is adapted to a target domain without access to
source data. We propose a novel approach for the SF-UDA setting based on a loss
reweighting strategy that brings robustness against the noise that inevitably
affects the pseudo-labels. The classification loss is reweighted based on the
reliability of the pseudo-labels that is measured by estimating their
uncertainty. Guided by such reweighting strategy, the pseudo-labels are
progressively refined by aggregating knowledge from neighbouring samples.
Furthermore, a self-supervised contrastive framework is leveraged as a target
space regulariser to enhance such knowledge aggregation. A novel negative pairs
exclusion strategy is proposed to identify and exclude negative pairs made of
samples sharing the same class, even in presence of some noise in the
pseudo-labels. Our method outperforms previous methods on three major
benchmarks by a large margin. We set the new SF-UDA state-of-the-art on VisDA-C
and DomainNet with a performance gain of +1.8% on both benchmarks and on PACS
with +12.3% in the single-source setting and +6.6% in multi-target adaptation.
Additional analyses demonstrate that the proposed approach is robust to the
noise, which results in significantly more accurate pseudo-labels compared to
state-of-the-art approaches.
| [
{
"version": "v1",
"created": "Tue, 7 Mar 2023 10:04:55 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 10:09:12 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 12:59:22 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Litrico",
"Mattia",
""
],
[
"Del Bue",
"Alessio",
""
],
[
"Morerio",
"Pietro",
""
]
] | TITLE: Guiding Pseudo-labels with Uncertainty Estimation for Source-free
Unsupervised Domain Adaptation
ABSTRACT: Standard Unsupervised Domain Adaptation (UDA) methods assume the availability
of both source and target data during the adaptation. In this work, we
investigate Source-free Unsupervised Domain Adaptation (SF-UDA), a specific
case of UDA where a model is adapted to a target domain without access to
source data. We propose a novel approach for the SF-UDA setting based on a loss
reweighting strategy that brings robustness against the noise that inevitably
affects the pseudo-labels. The classification loss is reweighted based on the
reliability of the pseudo-labels that is measured by estimating their
uncertainty. Guided by such reweighting strategy, the pseudo-labels are
progressively refined by aggregating knowledge from neighbouring samples.
Furthermore, a self-supervised contrastive framework is leveraged as a target
space regulariser to enhance such knowledge aggregation. A novel negative pairs
exclusion strategy is proposed to identify and exclude negative pairs made of
samples sharing the same class, even in presence of some noise in the
pseudo-labels. Our method outperforms previous methods on three major
benchmarks by a large margin. We set the new SF-UDA state-of-the-art on VisDA-C
and DomainNet with a performance gain of +1.8% on both benchmarks and on PACS
with +12.3% in the single-source setting and +6.6% in multi-target adaptation.
Additional analyses demonstrate that the proposed approach is robust to the
noise, which results in significantly more accurate pseudo-labels compared to
state-of-the-art approaches.
| no_new_dataset | 0.947817 |
2303.05012 | Danlei Hu | Danlei Hu, Lu Chen, Hanxi Fang, Ziquan Fang, Tianyi Li, Yunjun Gao | Spatio-Temporal Trajectory Similarity Measures: A Comprehensive Survey
and Quantitative Study | 20 pages, 21 figures | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spatio-temporal trajectory analytics is at the core of smart mobility
solutions, which offers unprecedented information for diversified applications
such as urban planning, infrastructure development, and vehicular networks.
Trajectory similarity measure, which aims to evaluate the distance between two
trajectories, is a fundamental functionality of trajectory analytics. In this
paper, we propose a comprehensive survey that investigates all the most common
and representative spatio-temporal trajectory measures. First, we provide an
overview of spatio-temporal trajectory measures in terms of three hierarchical
perspectives: Non-learning vs. Learning, Free Space vs. Road Network, and
Standalone vs. Distributed. Next, we present an evaluation benchmark by
designing five real-world transformation scenarios. Based on this benchmark,
extensive experiments are conducted to study the effectiveness,
robustness,nefficiency, and scalability of each measure, which offers
guidelines for trajectory measure selection among multiple techniques and
applications such as trajectory data mining, deep learning, and distributed
processing.
| [
{
"version": "v1",
"created": "Thu, 9 Mar 2023 03:28:22 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 06:12:17 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Hu",
"Danlei",
""
],
[
"Chen",
"Lu",
""
],
[
"Fang",
"Hanxi",
""
],
[
"Fang",
"Ziquan",
""
],
[
"Li",
"Tianyi",
""
],
[
"Gao",
"Yunjun",
""
]
] | TITLE: Spatio-Temporal Trajectory Similarity Measures: A Comprehensive Survey
and Quantitative Study
ABSTRACT: Spatio-temporal trajectory analytics is at the core of smart mobility
solutions, which offers unprecedented information for diversified applications
such as urban planning, infrastructure development, and vehicular networks.
Trajectory similarity measure, which aims to evaluate the distance between two
trajectories, is a fundamental functionality of trajectory analytics. In this
paper, we propose a comprehensive survey that investigates all the most common
and representative spatio-temporal trajectory measures. First, we provide an
overview of spatio-temporal trajectory measures in terms of three hierarchical
perspectives: Non-learning vs. Learning, Free Space vs. Road Network, and
Standalone vs. Distributed. Next, we present an evaluation benchmark by
designing five real-world transformation scenarios. Based on this benchmark,
extensive experiments are conducted to study the effectiveness,
robustness,nefficiency, and scalability of each measure, which offers
guidelines for trajectory measure selection among multiple techniques and
applications such as trajectory data mining, deep learning, and distributed
processing.
| no_new_dataset | 0.944944 |
2303.07130 | Anand Thyagachandran | Anand Thyagachandran, Hema A Murthy | Enhancing COVID-19 Severity Analysis through Ensemble Methods | null | null | null | null | eess.IV cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computed Tomography (CT) scans provide a detailed image of the lungs,
allowing clinicians to observe the extent of damage caused by COVID-19. The CT
severity score (CTSS) based scoring method is used to identify the extent of
lung involvement observed on a CT scan. This paper presents a domain
knowledge-based pipeline for extracting regions of infection in COVID-19
patients using a combination of image-processing algorithms and a pre-trained
UNET model. The severity of the infection is then classified into different
categories using an ensemble of three machine-learning models: Extreme Gradient
Boosting, Extremely Randomized Trees, and Support Vector Machine. The proposed
system was evaluated on a validation dataset in the AI-Enabled Medical Image
Analysis Workshop and COVID-19 Diagnosis Competition (AI-MIA-COV19D) and
achieved a macro F1 score of 64%. These results demonstrate the potential of
combining domain knowledge with machine learning techniques for accurate
COVID-19 diagnosis using CT scans. The implementation of the proposed system
for severity analysis is available at
\textit{https://github.com/aanandt/Enhancing-COVID-19-Severity-Analysis-through-Ensemble-Methods.git
}
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 13:59:47 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 08:59:24 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 13:58:52 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Thyagachandran",
"Anand",
""
],
[
"Murthy",
"Hema A",
""
]
] | TITLE: Enhancing COVID-19 Severity Analysis through Ensemble Methods
ABSTRACT: Computed Tomography (CT) scans provide a detailed image of the lungs,
allowing clinicians to observe the extent of damage caused by COVID-19. The CT
severity score (CTSS) based scoring method is used to identify the extent of
lung involvement observed on a CT scan. This paper presents a domain
knowledge-based pipeline for extracting regions of infection in COVID-19
patients using a combination of image-processing algorithms and a pre-trained
UNET model. The severity of the infection is then classified into different
categories using an ensemble of three machine-learning models: Extreme Gradient
Boosting, Extremely Randomized Trees, and Support Vector Machine. The proposed
system was evaluated on a validation dataset in the AI-Enabled Medical Image
Analysis Workshop and COVID-19 Diagnosis Competition (AI-MIA-COV19D) and
achieved a macro F1 score of 64%. These results demonstrate the potential of
combining domain knowledge with machine learning techniques for accurate
COVID-19 diagnosis using CT scans. The implementation of the proposed system
for severity analysis is available at
\textit{https://github.com/aanandt/Enhancing-COVID-19-Severity-Analysis-through-Ensemble-Methods.git
}
| no_new_dataset | 0.955277 |
2303.07269 | Zhuoran Yu | Zhuoran Yu, Yin Li, Yong Jae Lee | InPL: Pseudo-labeling the Inliers First for Imbalanced Semi-supervised
Learning | Accepted by ICLR 2023 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent state-of-the-art methods in imbalanced semi-supervised learning (SSL)
rely on confidence-based pseudo-labeling with consistency regularization. To
obtain high-quality pseudo-labels, a high confidence threshold is typically
adopted. However, it has been shown that softmax-based confidence scores in
deep networks can be arbitrarily high for samples far from the training data,
and thus, the pseudo-labels for even high-confidence unlabeled samples may
still be unreliable. In this work, we present a new perspective of
pseudo-labeling for imbalanced SSL. Without relying on model confidence, we
propose to measure whether an unlabeled sample is likely to be
``in-distribution''; i.e., close to the current training data. To decide
whether an unlabeled sample is ``in-distribution'' or ``out-of-distribution'',
we adopt the energy score from out-of-distribution detection literature. As
training progresses and more unlabeled samples become in-distribution and
contribute to training, the combined labeled and pseudo-labeled data can better
approximate the true class distribution to improve the model. Experiments
demonstrate that our energy-based pseudo-labeling method, \textbf{InPL}, albeit
conceptually simple, significantly outperforms confidence-based methods on
imbalanced SSL benchmarks. For example, it produces around 3\% absolute
accuracy improvement on CIFAR10-LT. When combined with state-of-the-art
long-tailed SSL methods, further improvements are attained. In particular, in
one of the most challenging scenarios, InPL achieves a 6.9\% accuracy
improvement over the best competitor.
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 16:45:41 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yu",
"Zhuoran",
""
],
[
"Li",
"Yin",
""
],
[
"Lee",
"Yong Jae",
""
]
] | TITLE: InPL: Pseudo-labeling the Inliers First for Imbalanced Semi-supervised
Learning
ABSTRACT: Recent state-of-the-art methods in imbalanced semi-supervised learning (SSL)
rely on confidence-based pseudo-labeling with consistency regularization. To
obtain high-quality pseudo-labels, a high confidence threshold is typically
adopted. However, it has been shown that softmax-based confidence scores in
deep networks can be arbitrarily high for samples far from the training data,
and thus, the pseudo-labels for even high-confidence unlabeled samples may
still be unreliable. In this work, we present a new perspective of
pseudo-labeling for imbalanced SSL. Without relying on model confidence, we
propose to measure whether an unlabeled sample is likely to be
``in-distribution''; i.e., close to the current training data. To decide
whether an unlabeled sample is ``in-distribution'' or ``out-of-distribution'',
we adopt the energy score from out-of-distribution detection literature. As
training progresses and more unlabeled samples become in-distribution and
contribute to training, the combined labeled and pseudo-labeled data can better
approximate the true class distribution to improve the model. Experiments
demonstrate that our energy-based pseudo-labeling method, \textbf{InPL}, albeit
conceptually simple, significantly outperforms confidence-based methods on
imbalanced SSL benchmarks. For example, it produces around 3\% absolute
accuracy improvement on CIFAR10-LT. When combined with state-of-the-art
long-tailed SSL methods, further improvements are attained. In particular, in
one of the most challenging scenarios, InPL achieves a 6.9\% accuracy
improvement over the best competitor.
| no_new_dataset | 0.947186 |
2303.08455 | Yu Yang | Yu Yang, Helin Gong, Qihong Yang, Yangtao Deng, Qiaolin He, Shiquan
Zhang | On the uncertainty analysis of the data-enabled physics-informed neural
network for solving neutron diffusion eigenvalue problem | The experiments in Figures 6 and 10 in the article have errors that
need to be corrected. Moreover, we intend to make massive changes to the
content of the article, and therefore need to withdraw the article | null | null | null | cs.LG physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In practical engineering experiments, the data obtained through detectors are
inevitably noisy. For the already proposed data-enabled physics-informed neural
network (DEPINN) \citep{DEPINN}, we investigate the performance of DEPINN in
calculating the neutron diffusion eigenvalue problem from several perspectives
when the prior data contain different scales of noise. Further, in order to
reduce the effect of noise and improve the utilization of the noisy prior data,
we propose innovative interval loss functions and give some rigorous
mathematical proofs. The robustness of DEPINN is examined on two typical
benchmark problems through a large number of numerical results, and the
effectiveness of the proposed interval loss function is demonstrated by
comparison. This paper confirms the feasibility of the improved DEPINN for
practical engineering applications in nuclear reactor physics.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 08:59:03 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 05:13:33 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Mar 2023 09:28:19 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yang",
"Yu",
""
],
[
"Gong",
"Helin",
""
],
[
"Yang",
"Qihong",
""
],
[
"Deng",
"Yangtao",
""
],
[
"He",
"Qiaolin",
""
],
[
"Zhang",
"Shiquan",
""
]
] | TITLE: On the uncertainty analysis of the data-enabled physics-informed neural
network for solving neutron diffusion eigenvalue problem
ABSTRACT: In practical engineering experiments, the data obtained through detectors are
inevitably noisy. For the already proposed data-enabled physics-informed neural
network (DEPINN) \citep{DEPINN}, we investigate the performance of DEPINN in
calculating the neutron diffusion eigenvalue problem from several perspectives
when the prior data contain different scales of noise. Further, in order to
reduce the effect of noise and improve the utilization of the noisy prior data,
we propose innovative interval loss functions and give some rigorous
mathematical proofs. The robustness of DEPINN is examined on two typical
benchmark problems through a large number of numerical results, and the
effectiveness of the proposed interval loss function is demonstrated by
comparison. This paper confirms the feasibility of the improved DEPINN for
practical engineering applications in nuclear reactor physics.
| no_new_dataset | 0.950134 |
2303.08954 | Aditya Gupta | Rahul Goel, Waleed Ammar, Aditya Gupta, Siddharth Vashishtha, Motoki
Sano, Faiz Surani, Max Chang, HyunJeong Choe, David Greene, Kyle He, Rattima
Nitisaroj, Anna Trukhina, Shachi Paul, Pararth Shah, Rushin Shah and Zhou Yu | PRESTO: A Multilingual Dataset for Parsing Realistic Task-Oriented
Dialogs | PRESTO v1 Release | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Research interest in task-oriented dialogs has increased as systems such as
Google Assistant, Alexa and Siri have become ubiquitous in everyday life.
However, the impact of academic research in this area has been limited by the
lack of datasets that realistically capture the wide array of user pain points.
To enable research on some of the more challenging aspects of parsing realistic
conversations, we introduce PRESTO, a public dataset of over 550K contextual
multilingual conversations between humans and virtual assistants. PRESTO
contains a diverse array of challenges that occur in real-world NLU tasks such
as disfluencies, code-switching, and revisions. It is the only large scale
human generated conversational parsing dataset that provides structured context
such as a user's contacts and lists for each example. Our mT5 model based
baselines demonstrate that the conversational phenomenon present in PRESTO are
challenging to model, which is further pronounced in a low-resource setup.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 21:51:13 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 02:26:52 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Goel",
"Rahul",
""
],
[
"Ammar",
"Waleed",
""
],
[
"Gupta",
"Aditya",
""
],
[
"Vashishtha",
"Siddharth",
""
],
[
"Sano",
"Motoki",
""
],
[
"Surani",
"Faiz",
""
],
[
"Chang",
"Max",
""
],
[
"Choe",
"HyunJeong",
""
],
[
"Greene",
"David",
""
],
[
"He",
"Kyle",
""
],
[
"Nitisaroj",
"Rattima",
""
],
[
"Trukhina",
"Anna",
""
],
[
"Paul",
"Shachi",
""
],
[
"Shah",
"Pararth",
""
],
[
"Shah",
"Rushin",
""
],
[
"Yu",
"Zhou",
""
]
] | TITLE: PRESTO: A Multilingual Dataset for Parsing Realistic Task-Oriented
Dialogs
ABSTRACT: Research interest in task-oriented dialogs has increased as systems such as
Google Assistant, Alexa and Siri have become ubiquitous in everyday life.
However, the impact of academic research in this area has been limited by the
lack of datasets that realistically capture the wide array of user pain points.
To enable research on some of the more challenging aspects of parsing realistic
conversations, we introduce PRESTO, a public dataset of over 550K contextual
multilingual conversations between humans and virtual assistants. PRESTO
contains a diverse array of challenges that occur in real-world NLU tasks such
as disfluencies, code-switching, and revisions. It is the only large scale
human generated conversational parsing dataset that provides structured context
such as a user's contacts and lists for each example. Our mT5 model based
baselines demonstrate that the conversational phenomenon present in PRESTO are
challenging to model, which is further pronounced in a low-resource setup.
| new_dataset | 0.962638 |
2303.09170 | Chen Yaosen | Yaosen Chen, Han Yang, Yuexin Yang, Yuegen Liu, Wei Wang, Xuming Wen,
Chaoping Xie | NLUT: Neural-based 3D Lookup Tables for Video Photorealistic Style
Transfer | null | null | null | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video photorealistic style transfer is desired to generate videos with a
similar photorealistic style to the style image while maintaining temporal
consistency. However, existing methods obtain stylized video sequences by
performing frame-by-frame photorealistic style transfer, which is inefficient
and does not ensure the temporal consistency of the stylized video. To address
this issue, we use neural network-based 3D Lookup Tables (LUTs) for the
photorealistic transfer of videos, achieving a balance between efficiency and
effectiveness. We first train a neural network for generating photorealistic
stylized 3D LUTs on a large-scale dataset; then, when performing photorealistic
style transfer for a specific video, we select a keyframe and style image in
the video as the data source and fine-turn the neural network; finally, we
query the 3D LUTs generated by the fine-tuned neural network for the colors in
the video, resulting in a super-fast photorealistic style transfer, even
processing 8K video takes less than 2 millisecond per frame. The experimental
results show that our method not only realizes the photorealistic style
transfer of arbitrary style images but also outperforms the existing methods in
terms of visual quality and consistency. Project
page:https://semchan.github.io/NLUT_Project.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 09:27:40 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 05:10:48 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Chen",
"Yaosen",
""
],
[
"Yang",
"Han",
""
],
[
"Yang",
"Yuexin",
""
],
[
"Liu",
"Yuegen",
""
],
[
"Wang",
"Wei",
""
],
[
"Wen",
"Xuming",
""
],
[
"Xie",
"Chaoping",
""
]
] | TITLE: NLUT: Neural-based 3D Lookup Tables for Video Photorealistic Style
Transfer
ABSTRACT: Video photorealistic style transfer is desired to generate videos with a
similar photorealistic style to the style image while maintaining temporal
consistency. However, existing methods obtain stylized video sequences by
performing frame-by-frame photorealistic style transfer, which is inefficient
and does not ensure the temporal consistency of the stylized video. To address
this issue, we use neural network-based 3D Lookup Tables (LUTs) for the
photorealistic transfer of videos, achieving a balance between efficiency and
effectiveness. We first train a neural network for generating photorealistic
stylized 3D LUTs on a large-scale dataset; then, when performing photorealistic
style transfer for a specific video, we select a keyframe and style image in
the video as the data source and fine-turn the neural network; finally, we
query the 3D LUTs generated by the fine-tuned neural network for the colors in
the video, resulting in a super-fast photorealistic style transfer, even
processing 8K video takes less than 2 millisecond per frame. The experimental
results show that our method not only realizes the photorealistic style
transfer of arbitrary style images but also outperforms the existing methods in
terms of visual quality and consistency. Project
page:https://semchan.github.io/NLUT_Project.
| no_new_dataset | 0.956431 |
2303.09189 | Hiba Arnaout | Hiba Arnaout, Simon Razniewski, Jeff Z. Pan | Wiki-based Communities of Interest: Demographics and Outliers | Accepted to ICWSM 2023. For demo, see
https://wikiknowledge.onrender.com/demographics/ and for dataset see
https://doi.org/10.5281/zenodo.7410436 | null | null | null | cs.SI cs.CY | http://creativecommons.org/licenses/by/4.0/ | In this paper, we release data about demographic information and outliers of
communities of interest. Identified from Wiki-based sources, mainly Wikidata,
the data covers 7.5k communities, such as members of the White House
Coronavirus Task Force, and 345k subjects, e.g., Deborah Birx. We describe the
statistical inference methodology adopted to mine such data. We release
subject-centric and group-centric datasets in JSON format, as well as a
browsing interface. Finally, we forsee three areas this research can have an
impact on: in social sciences research, it provides a resource for demographic
analyses; in web-scale collaborative encyclopedias, it serves as an edit
recommender to fill knowledge gaps; and in web search, it offers lists of
salient statements about queried subjects for higher user engagement.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 09:58:11 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 08:34:30 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Arnaout",
"Hiba",
""
],
[
"Razniewski",
"Simon",
""
],
[
"Pan",
"Jeff Z.",
""
]
] | TITLE: Wiki-based Communities of Interest: Demographics and Outliers
ABSTRACT: In this paper, we release data about demographic information and outliers of
communities of interest. Identified from Wiki-based sources, mainly Wikidata,
the data covers 7.5k communities, such as members of the White House
Coronavirus Task Force, and 345k subjects, e.g., Deborah Birx. We describe the
statistical inference methodology adopted to mine such data. We release
subject-centric and group-centric datasets in JSON format, as well as a
browsing interface. Finally, we forsee three areas this research can have an
impact on: in social sciences research, it provides a resource for demographic
analyses; in web-scale collaborative encyclopedias, it serves as an edit
recommender to fill knowledge gaps; and in web search, it offers lists of
salient statements about queried subjects for higher user engagement.
| no_new_dataset | 0.741206 |
2303.09306 | Md Zarif Ul Alam | HAZ Sameen Shahgir, Ramisa Alam, Md. Zarif Ul Alam | BanglaCoNER: Towards Robust Bangla Complex Named Entity Recognition | Winning Solution for the Bangla Complex Named Entity Recognition
Challenge | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Named Entity Recognition (NER) is a fundamental task in natural language
processing that involves identifying and classifying named entities in text.
But much work hasn't been done for complex named entity recognition in Bangla,
despite being the seventh most spoken language globally. CNER is a more
challenging task than traditional NER as it involves identifying and
classifying complex and compound entities, which are not common in Bangla
language. In this paper, we present the winning solution of Bangla Complex
Named Entity Recognition Challenge - addressing the CNER task on BanglaCoNER
dataset using two different approaches, namely Conditional Random Fields (CRF)
and finetuning transformer based Deep Learning models such as BanglaBERT.
The dataset consisted of 15300 sentences for training and 800 sentences for
validation, in the .conll format. Exploratory Data Analysis (EDA) on the
dataset revealed that the dataset had 7 different NER tags, with notable
presence of English words, suggesting that the dataset is synthetic and likely
a product of translation.
We experimented with a variety of feature combinations including Part of
Speech (POS) tags, word suffixes, Gazetteers, and cluster information from
embeddings, while also finetuning the BanglaBERT (large) model for NER. We
found that not all linguistic patterns are immediately apparent or even
intuitive to humans, which is why Deep Learning based models has proved to be
the more effective model in NLP, including CNER task. Our fine tuned BanglaBERT
(large) model achieves an F1 Score of 0.79 on the validation set. Overall, our
study highlights the importance of Bangla Complex Named Entity Recognition,
particularly in the context of synthetic datasets. Our findings also
demonstrate the efficacy of Deep Learning models such as BanglaBERT for NER in
Bangla language.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 13:31:31 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 15:13:01 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Shahgir",
"HAZ Sameen",
""
],
[
"Alam",
"Ramisa",
""
],
[
"Alam",
"Md. Zarif Ul",
""
]
] | TITLE: BanglaCoNER: Towards Robust Bangla Complex Named Entity Recognition
ABSTRACT: Named Entity Recognition (NER) is a fundamental task in natural language
processing that involves identifying and classifying named entities in text.
But much work hasn't been done for complex named entity recognition in Bangla,
despite being the seventh most spoken language globally. CNER is a more
challenging task than traditional NER as it involves identifying and
classifying complex and compound entities, which are not common in Bangla
language. In this paper, we present the winning solution of Bangla Complex
Named Entity Recognition Challenge - addressing the CNER task on BanglaCoNER
dataset using two different approaches, namely Conditional Random Fields (CRF)
and finetuning transformer based Deep Learning models such as BanglaBERT.
The dataset consisted of 15300 sentences for training and 800 sentences for
validation, in the .conll format. Exploratory Data Analysis (EDA) on the
dataset revealed that the dataset had 7 different NER tags, with notable
presence of English words, suggesting that the dataset is synthetic and likely
a product of translation.
We experimented with a variety of feature combinations including Part of
Speech (POS) tags, word suffixes, Gazetteers, and cluster information from
embeddings, while also finetuning the BanglaBERT (large) model for NER. We
found that not all linguistic patterns are immediately apparent or even
intuitive to humans, which is why Deep Learning based models has proved to be
the more effective model in NLP, including CNER task. Our fine tuned BanglaBERT
(large) model achieves an F1 Score of 0.79 on the validation set. Overall, our
study highlights the importance of Bangla Complex Named Entity Recognition,
particularly in the context of synthetic datasets. Our findings also
demonstrate the efficacy of Deep Learning models such as BanglaBERT for NER in
Bangla language.
| no_new_dataset | 0.921711 |
2303.09558 | Pedro Neto | Laura Duarte, Pedro Neto | Classification of Primitive Manufacturing Tasks from Filtered Event Data | null | Journal of Manufacturing Systems, Volume 68, 2023 | 10.1016/j.jmsy.2023.03.001 | null | cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | Collaborative robots are increasingly present in industry to support human
activities. However, to make the human-robot collaborative process more
effective, there are several challenges to be addressed. Collaborative robotic
systems need to be aware of the human activities to (1) anticipate
collaborative/assistive actions, (2) learn by demonstration, and (3) activate
safety procedures in shared workspace. This study proposes an action
classification system to recognize primitive assembly tasks from human motion
events data captured by a Dynamic and Active-pixel Vision Sensor (DAVIS).
Several filters are compared and combined to remove event data noise. Task
patterns are classified from a continuous stream of event data using advanced
deep learning and recurrent networks to classify spatial and temporal features.
Experiments were conducted on a novel dataset, the dataset of manufacturing
tasks (DMT22), featuring 5 classes of representative manufacturing primitives
(PickUp, Place, Screw, Hold, Idle) from 5 participants. Results show that the
proposed filters remove about 65\% of all events (noise) per recording,
conducting to a classification accuracy up to 99,37\% for subjects that trained
the system and 97.08\% for new subjects. Data from a left-handed subject were
successfully classified using only right-handed training data. These results
are object independent.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 19:59:38 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Duarte",
"Laura",
""
],
[
"Neto",
"Pedro",
""
]
] | TITLE: Classification of Primitive Manufacturing Tasks from Filtered Event Data
ABSTRACT: Collaborative robots are increasingly present in industry to support human
activities. However, to make the human-robot collaborative process more
effective, there are several challenges to be addressed. Collaborative robotic
systems need to be aware of the human activities to (1) anticipate
collaborative/assistive actions, (2) learn by demonstration, and (3) activate
safety procedures in shared workspace. This study proposes an action
classification system to recognize primitive assembly tasks from human motion
events data captured by a Dynamic and Active-pixel Vision Sensor (DAVIS).
Several filters are compared and combined to remove event data noise. Task
patterns are classified from a continuous stream of event data using advanced
deep learning and recurrent networks to classify spatial and temporal features.
Experiments were conducted on a novel dataset, the dataset of manufacturing
tasks (DMT22), featuring 5 classes of representative manufacturing primitives
(PickUp, Place, Screw, Hold, Idle) from 5 participants. Results show that the
proposed filters remove about 65\% of all events (noise) per recording,
conducting to a classification accuracy up to 99,37\% for subjects that trained
the system and 97.08\% for new subjects. Data from a left-handed subject were
successfully classified using only right-handed training data. These results
are object independent.
| new_dataset | 0.963127 |
2303.09611 | Jiang Hu | Jiang Hu, Kangkang Deng, Na Li, Quanzheng Li | Decentralized Riemannian natural gradient methods with Kronecker-product
approximations | 17 pages | null | null | null | math.OC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With a computationally efficient approximation of the second-order
information, natural gradient methods have been successful in solving
large-scale structured optimization problems. We study the natural gradient
methods for the large-scale decentralized optimization problems on Riemannian
manifolds, where the local objective function defined by the local dataset is
of a log-probability type. By utilizing the structure of the Riemannian Fisher
information matrix (RFIM), we present an efficient decentralized Riemannian
natural gradient descent (DRNGD) method. To overcome the communication issue of
the high-dimension RFIM, we consider a class of structured problems for which
the RFIM can be approximated by a Kronecker product of two low-dimension
matrices. By performing the communications over the Kronecker factors, a
high-quality approximation of the RFIM can be obtained in a low cost. We prove
that DRNGD converges to a stationary point with the best-known rate of
$\mathcal{O}(1/K)$. Numerical experiments demonstrate the efficiency of our
proposed method compared with the state-of-the-art ones. To the best of our
knowledge, this is the first Riemannian second-order method for solving
decentralized manifold optimization problems.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 19:36:31 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Hu",
"Jiang",
""
],
[
"Deng",
"Kangkang",
""
],
[
"Li",
"Na",
""
],
[
"Li",
"Quanzheng",
""
]
] | TITLE: Decentralized Riemannian natural gradient methods with Kronecker-product
approximations
ABSTRACT: With a computationally efficient approximation of the second-order
information, natural gradient methods have been successful in solving
large-scale structured optimization problems. We study the natural gradient
methods for the large-scale decentralized optimization problems on Riemannian
manifolds, where the local objective function defined by the local dataset is
of a log-probability type. By utilizing the structure of the Riemannian Fisher
information matrix (RFIM), we present an efficient decentralized Riemannian
natural gradient descent (DRNGD) method. To overcome the communication issue of
the high-dimension RFIM, we consider a class of structured problems for which
the RFIM can be approximated by a Kronecker product of two low-dimension
matrices. By performing the communications over the Kronecker factors, a
high-quality approximation of the RFIM can be obtained in a low cost. We prove
that DRNGD converges to a stationary point with the best-known rate of
$\mathcal{O}(1/K)$. Numerical experiments demonstrate the efficiency of our
proposed method compared with the state-of-the-art ones. To the best of our
knowledge, this is the first Riemannian second-order method for solving
decentralized manifold optimization problems.
| no_new_dataset | 0.943191 |
2303.09617 | Paul Mvula | William Aiken, Paul K. Mvula, Paula Branco, Guy-Vincent Jourdan,
Mehrdad Sabetzadeh, Herna Viktor | Measuring Improvement of F$_1$-Scores in Detection of Self-Admitted
Technical Debt | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Artificial Intelligence and Machine Learning have witnessed rapid,
significant improvements in Natural Language Processing (NLP) tasks. Utilizing
Deep Learning, researchers have taken advantage of repository comments in
Software Engineering to produce accurate methods for detecting Self-Admitted
Technical Debt (SATD) from 20 open-source Java projects' code. In this work, we
improve SATD detection with a novel approach that leverages the Bidirectional
Encoder Representations from Transformers (BERT) architecture. For comparison,
we re-evaluated previous deep learning methods and applied stratified 10-fold
cross-validation to report reliable F$_1$-scores. We examine our model in both
cross-project and intra-project contexts. For each context, we use re-sampling
and duplication as augmentation strategies to account for data imbalance. We
find that our trained BERT model improves over the best performance of all
previous methods in 19 of the 20 projects in cross-project scenarios. However,
the data augmentation techniques were not sufficient to overcome the lack of
data present in the intra-project scenarios, and existing methods still perform
better. Future research will look into ways to diversify SATD datasets in order
to maximize the latent power in large BERT models.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 19:47:38 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Aiken",
"William",
""
],
[
"Mvula",
"Paul K.",
""
],
[
"Branco",
"Paula",
""
],
[
"Jourdan",
"Guy-Vincent",
""
],
[
"Sabetzadeh",
"Mehrdad",
""
],
[
"Viktor",
"Herna",
""
]
] | TITLE: Measuring Improvement of F$_1$-Scores in Detection of Self-Admitted
Technical Debt
ABSTRACT: Artificial Intelligence and Machine Learning have witnessed rapid,
significant improvements in Natural Language Processing (NLP) tasks. Utilizing
Deep Learning, researchers have taken advantage of repository comments in
Software Engineering to produce accurate methods for detecting Self-Admitted
Technical Debt (SATD) from 20 open-source Java projects' code. In this work, we
improve SATD detection with a novel approach that leverages the Bidirectional
Encoder Representations from Transformers (BERT) architecture. For comparison,
we re-evaluated previous deep learning methods and applied stratified 10-fold
cross-validation to report reliable F$_1$-scores. We examine our model in both
cross-project and intra-project contexts. For each context, we use re-sampling
and duplication as augmentation strategies to account for data imbalance. We
find that our trained BERT model improves over the best performance of all
previous methods in 19 of the 20 projects in cross-project scenarios. However,
the data augmentation techniques were not sufficient to overcome the lack of
data present in the intra-project scenarios, and existing methods still perform
better. Future research will look into ways to diversify SATD datasets in order
to maximize the latent power in large BERT models.
| no_new_dataset | 0.939248 |
2303.09634 | Fearghal O'Donncha | Abigail Langbridge, Fearghal O'Donncha, Amadou Ba, Fabio Lorenzi,
Christopher Lohse, Joern Ploennigs | Causal Temporal Graph Convolutional Neural Networks (CTGCN) | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Many large-scale applications can be elegantly represented using graph
structures. Their scalability, however, is often limited by the domain
knowledge required to apply them. To address this problem, we propose a novel
Causal Temporal Graph Convolutional Neural Network (CTGCN). Our CTGCN
architecture is based on a causal discovery mechanism, and is capable of
discovering the underlying causal processes. The major advantages of our
approach stem from its ability to overcome computational scalability problems
with a divide and conquer technique, and from the greater explainability of
predictions made using a causal model. We evaluate the scalability of our CTGCN
on two datasets to demonstrate that our method is applicable to large scale
problems, and show that the integration of causality into the TGCN architecture
improves prediction performance up to 40% over typical TGCN approach. Our
results are obtained without requiring additional domain knowledge, making our
approach adaptable to various domains, specifically when little contextual
knowledge is available.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 20:28:36 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Langbridge",
"Abigail",
""
],
[
"O'Donncha",
"Fearghal",
""
],
[
"Ba",
"Amadou",
""
],
[
"Lorenzi",
"Fabio",
""
],
[
"Lohse",
"Christopher",
""
],
[
"Ploennigs",
"Joern",
""
]
] | TITLE: Causal Temporal Graph Convolutional Neural Networks (CTGCN)
ABSTRACT: Many large-scale applications can be elegantly represented using graph
structures. Their scalability, however, is often limited by the domain
knowledge required to apply them. To address this problem, we propose a novel
Causal Temporal Graph Convolutional Neural Network (CTGCN). Our CTGCN
architecture is based on a causal discovery mechanism, and is capable of
discovering the underlying causal processes. The major advantages of our
approach stem from its ability to overcome computational scalability problems
with a divide and conquer technique, and from the greater explainability of
predictions made using a causal model. We evaluate the scalability of our CTGCN
on two datasets to demonstrate that our method is applicable to large scale
problems, and show that the integration of causality into the TGCN architecture
improves prediction performance up to 40% over typical TGCN approach. Our
results are obtained without requiring additional domain knowledge, making our
approach adaptable to various domains, specifically when little contextual
knowledge is available.
| no_new_dataset | 0.946646 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.