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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2303.09648 | Jinfan Zhou | Jinfan Zhou, William Muirhead, Simon C. Williams, Danail Stoyanov,
Hani J. Marcus, and Evangelos B. Mazomenos | Shifted-Windows Transformers for the Detection of Cerebral Aneurysms in
Microsurgery | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Purpose: Microsurgical Aneurysm Clipping Surgery (MACS) carries a high risk
for intraoperative aneurysm rupture. Automated recognition of instances when
the aneurysm is exposed in the surgical video would be a valuable reference
point for neuronavigation, indicating phase transitioning and more importantly
designating moments of high risk for rupture. This article introduces the MACS
dataset containing 16 surgical videos with frame-level expert annotations and
proposes a learning methodology for surgical scene understanding identifying
video frames with the aneurysm present in the operating microscope's
field-of-view. Methods: Despite the dataset imbalance (80% no presence, 20%
presence) and developed without explicit annotations, we demonstrate the
applicability of Transformer-based deep learning architectures (MACSSwin-T,
vidMACSSwin-T) to detect the aneurysm and classify MACS frames accordingly. We
evaluate the proposed models in multiple-fold cross-validation experiments with
independent sets and in an unseen set of 15 images against 10 human experts
(neurosurgeons). Results: Average (across folds) accuracy of 80.8% (range
78.5%-82.4%) and 87.1% (range 85.1%-91.3%) is obtained for the image- and
video-level approach respectively, demonstrating that the models effectively
learn the classification task. Qualitative evaluation of the models' class
activation maps show these to be localized on the aneurysm's actual location.
Depending on the decision threshold, MACSWin-T achieves 66.7% to 86.7% accuracy
in the unseen images, compared to 82% of human raters, with moderate to strong
correlation.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 20:58:48 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Zhou",
"Jinfan",
""
],
[
"Muirhead",
"William",
""
],
[
"Williams",
"Simon C.",
""
],
[
"Stoyanov",
"Danail",
""
],
[
"Marcus",
"Hani J.",
""
],
[
"Mazomenos",
"Evangelos B.",
""
]
] | TITLE: Shifted-Windows Transformers for the Detection of Cerebral Aneurysms in
Microsurgery
ABSTRACT: Purpose: Microsurgical Aneurysm Clipping Surgery (MACS) carries a high risk
for intraoperative aneurysm rupture. Automated recognition of instances when
the aneurysm is exposed in the surgical video would be a valuable reference
point for neuronavigation, indicating phase transitioning and more importantly
designating moments of high risk for rupture. This article introduces the MACS
dataset containing 16 surgical videos with frame-level expert annotations and
proposes a learning methodology for surgical scene understanding identifying
video frames with the aneurysm present in the operating microscope's
field-of-view. Methods: Despite the dataset imbalance (80% no presence, 20%
presence) and developed without explicit annotations, we demonstrate the
applicability of Transformer-based deep learning architectures (MACSSwin-T,
vidMACSSwin-T) to detect the aneurysm and classify MACS frames accordingly. We
evaluate the proposed models in multiple-fold cross-validation experiments with
independent sets and in an unseen set of 15 images against 10 human experts
(neurosurgeons). Results: Average (across folds) accuracy of 80.8% (range
78.5%-82.4%) and 87.1% (range 85.1%-91.3%) is obtained for the image- and
video-level approach respectively, demonstrating that the models effectively
learn the classification task. Qualitative evaluation of the models' class
activation maps show these to be localized on the aneurysm's actual location.
Depending on the decision threshold, MACSWin-T achieves 66.7% to 86.7% accuracy
in the unseen images, compared to 82% of human raters, with moderate to strong
correlation.
| no_new_dataset | 0.951729 |
2303.09660 | Chia-Yu Hsu | Chia-Yu Hsu and Wenwen Li | Explainable GeoAI: Can saliency maps help interpret artificial
intelligence's learning process? An empirical study on natural feature
detection | 24 Pages, 7 Figures, Accepted to International Journal of
Geographical Information Science, 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Improving the interpretability of geospatial artificial intelligence (GeoAI)
models has become critically important to open the "black box" of complex AI
models, such as deep learning. This paper compares popular saliency map
generation techniques and their strengths and weaknesses in interpreting GeoAI
and deep learning models' reasoning behaviors, particularly when applied to
geospatial analysis and image processing tasks. We surveyed two broad classes
of model explanation methods: perturbation-based and gradient-based methods.
The former identifies important image areas, which help machines make
predictions by modifying a localized area of the input image. The latter
evaluates the contribution of every single pixel of the input image to the
model's prediction results through gradient backpropagation. In this study,
three algorithms-the occlusion method, the integrated gradients method, and the
class activation map method-are examined for a natural feature detection task
using deep learning. The algorithms' strengths and weaknesses are discussed,
and the consistency between model-learned and human-understandable concepts for
object recognition is also compared. The experiments used two GeoAI-ready
datasets to demonstrate the generalizability of the research findings.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 21:37:29 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Hsu",
"Chia-Yu",
""
],
[
"Li",
"Wenwen",
""
]
] | TITLE: Explainable GeoAI: Can saliency maps help interpret artificial
intelligence's learning process? An empirical study on natural feature
detection
ABSTRACT: Improving the interpretability of geospatial artificial intelligence (GeoAI)
models has become critically important to open the "black box" of complex AI
models, such as deep learning. This paper compares popular saliency map
generation techniques and their strengths and weaknesses in interpreting GeoAI
and deep learning models' reasoning behaviors, particularly when applied to
geospatial analysis and image processing tasks. We surveyed two broad classes
of model explanation methods: perturbation-based and gradient-based methods.
The former identifies important image areas, which help machines make
predictions by modifying a localized area of the input image. The latter
evaluates the contribution of every single pixel of the input image to the
model's prediction results through gradient backpropagation. In this study,
three algorithms-the occlusion method, the integrated gradients method, and the
class activation map method-are examined for a natural feature detection task
using deep learning. The algorithms' strengths and weaknesses are discussed,
and the consistency between model-learned and human-understandable concepts for
object recognition is also compared. The experiments used two GeoAI-ready
datasets to demonstrate the generalizability of the research findings.
| no_new_dataset | 0.94801 |
2303.09674 | Jiawei Ma | Jiawei Ma, Yulei Niu, Jincheng Xu, Shiyuan Huang, Guangxing Han,
Shih-Fu Chang | DiGeo: Discriminative Geometry-Aware Learning for Generalized Few-Shot
Object Detection | CVPR 2023 Camera Ready (Supp Attached). Code Link:
https://github.com/Phoenix-V/DiGeo | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generalized few-shot object detection aims to achieve precise detection on
both base classes with abundant annotations and novel classes with limited
training data. Existing approaches enhance few-shot generalization with the
sacrifice of base-class performance, or maintain high precision in base-class
detection with limited improvement in novel-class adaptation. In this paper, we
point out the reason is insufficient Discriminative feature learning for all of
the classes. As such, we propose a new training framework, DiGeo, to learn
Geometry-aware features of inter-class separation and intra-class compactness.
To guide the separation of feature clusters, we derive an offline simplex
equiangular tight frame (ETF) classifier whose weights serve as class centers
and are maximally and equally separated. To tighten the cluster for each class,
we include adaptive class-specific margins into the classification loss and
encourage the features close to the class centers. Experimental studies on two
few-shot benchmark datasets (VOC, COCO) and one long-tail dataset (LVIS)
demonstrate that, with a single model, our method can effectively improve
generalization on novel classes without hurting the detection of base classes.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 22:37:09 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ma",
"Jiawei",
""
],
[
"Niu",
"Yulei",
""
],
[
"Xu",
"Jincheng",
""
],
[
"Huang",
"Shiyuan",
""
],
[
"Han",
"Guangxing",
""
],
[
"Chang",
"Shih-Fu",
""
]
] | TITLE: DiGeo: Discriminative Geometry-Aware Learning for Generalized Few-Shot
Object Detection
ABSTRACT: Generalized few-shot object detection aims to achieve precise detection on
both base classes with abundant annotations and novel classes with limited
training data. Existing approaches enhance few-shot generalization with the
sacrifice of base-class performance, or maintain high precision in base-class
detection with limited improvement in novel-class adaptation. In this paper, we
point out the reason is insufficient Discriminative feature learning for all of
the classes. As such, we propose a new training framework, DiGeo, to learn
Geometry-aware features of inter-class separation and intra-class compactness.
To guide the separation of feature clusters, we derive an offline simplex
equiangular tight frame (ETF) classifier whose weights serve as class centers
and are maximally and equally separated. To tighten the cluster for each class,
we include adaptive class-specific margins into the classification loss and
encourage the features close to the class centers. Experimental studies on two
few-shot benchmark datasets (VOC, COCO) and one long-tail dataset (LVIS)
demonstrate that, with a single model, our method can effectively improve
generalization on novel classes without hurting the detection of base classes.
| no_new_dataset | 0.950778 |
2303.09677 | Pietro Astolfi | Pietro Astolfi, Arantxa Casanova, Jakob Verbeek, Pascal Vincent,
Adriana Romero-Soriano, Michal Drozdzal | Instance-Conditioned GAN Data Augmentation for Representation Learning | TMLR reviews at
https://openreview.net/forum?id=1n7q9mxG3T&referrer=%5BTMLR%5D(%2Fgroup%3Fid%3DTMLR) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data augmentation has become a crucial component to train state-of-the-art
visual representation models. However, handcrafting combinations of
transformations that lead to improved performances is a laborious task, which
can result in visually unrealistic samples. To overcome these limitations,
recent works have explored the use of generative models as learnable data
augmentation tools, showing promising results in narrow application domains,
e.g., few-shot learning and low-data medical imaging. In this paper, we
introduce a data augmentation module, called DA_IC-GAN, which leverages
instance-conditioned GAN generations and can be used off-the-shelf in
conjunction with most state-of-the-art training recipes. We showcase the
benefits of DA_IC-GAN by plugging it out-of-the-box into the supervised
training of ResNets and DeiT models on the ImageNet dataset, and achieving
accuracy boosts up to between 1%p and 2%p with the highest capacity models.
Moreover, the learnt representations are shown to be more robust than the
baselines when transferred to a handful of out-of-distribution datasets, and
exhibit increased invariance to variations of instance and viewpoints. We
additionally couple DA_IC-GAN with a self-supervised training recipe and show
that we can also achieve an improvement of 1%p in accuracy in some settings.
With this work, we strengthen the evidence on the potential of learnable data
augmentations to improve visual representation learning, paving the road
towards non-handcrafted augmentations in model training.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 22:45:43 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Astolfi",
"Pietro",
""
],
[
"Casanova",
"Arantxa",
""
],
[
"Verbeek",
"Jakob",
""
],
[
"Vincent",
"Pascal",
""
],
[
"Romero-Soriano",
"Adriana",
""
],
[
"Drozdzal",
"Michal",
""
]
] | TITLE: Instance-Conditioned GAN Data Augmentation for Representation Learning
ABSTRACT: Data augmentation has become a crucial component to train state-of-the-art
visual representation models. However, handcrafting combinations of
transformations that lead to improved performances is a laborious task, which
can result in visually unrealistic samples. To overcome these limitations,
recent works have explored the use of generative models as learnable data
augmentation tools, showing promising results in narrow application domains,
e.g., few-shot learning and low-data medical imaging. In this paper, we
introduce a data augmentation module, called DA_IC-GAN, which leverages
instance-conditioned GAN generations and can be used off-the-shelf in
conjunction with most state-of-the-art training recipes. We showcase the
benefits of DA_IC-GAN by plugging it out-of-the-box into the supervised
training of ResNets and DeiT models on the ImageNet dataset, and achieving
accuracy boosts up to between 1%p and 2%p with the highest capacity models.
Moreover, the learnt representations are shown to be more robust than the
baselines when transferred to a handful of out-of-distribution datasets, and
exhibit increased invariance to variations of instance and viewpoints. We
additionally couple DA_IC-GAN with a self-supervised training recipe and show
that we can also achieve an improvement of 1%p in accuracy in some settings.
With this work, we strengthen the evidence on the potential of learnable data
augmentations to improve visual representation learning, paving the road
towards non-handcrafted augmentations in model training.
| no_new_dataset | 0.941654 |
2303.09679 | Yongmin Li | Ifeyinwa Linda Anene and Yongmin Li | Segmentation of Retinal Blood Vessels Using Deep Learning | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The morphology of retinal blood vessels can indicate various diseases in the
human body, and researchers have been working on automatic scanning and
segmentation of retinal images to aid diagnosis. This project compares the
performance of four neural network architectures in segmenting retinal images,
using a combined dataset from different databases, namely the UNet, DR-VNet,
UNet-ResNet and UNet-VGG.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 22:47:48 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Anene",
"Ifeyinwa Linda",
""
],
[
"Li",
"Yongmin",
""
]
] | TITLE: Segmentation of Retinal Blood Vessels Using Deep Learning
ABSTRACT: The morphology of retinal blood vessels can indicate various diseases in the
human body, and researchers have been working on automatic scanning and
segmentation of retinal images to aid diagnosis. This project compares the
performance of four neural network architectures in segmenting retinal images,
using a combined dataset from different databases, namely the UNet, DR-VNet,
UNet-ResNet and UNet-VGG.
| no_new_dataset | 0.945298 |
2303.09694 | Balsam Alkouz | Shilong Guo, Balsam Alkouz, Babar Shahzaad, Abdallah Lakhdari, Athman
Bouguettaya | Drone Formation for Efficient Swarm Energy Consumption | 3 pages, 7 figures. This is an accepted demo paper and it will appear
in The 21st International Conference on Pervasive Computing and
Communications (PerCom 2023) | null | null | null | cs.RO eess.SP | http://creativecommons.org/licenses/by/4.0/ | We demonstrate formation flying for drone swarm services. A set of drones fly
in four different swarm formations. A dataset is collected to study the effect
of formation flying on energy consumption. We conduct a set of experiments to
study the effect of wind on formation flying. We examine the forces the drones
exert on each other when flying in a formation. We finally identify and
classify the formations that conserve most energy under varying wind
conditions. The collected dataset aims at providing researchers data to conduct
further research in swarm-based drone service delivery. Demo:
https://youtu.be/NnucUWhUwLs
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 23:52:52 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Guo",
"Shilong",
""
],
[
"Alkouz",
"Balsam",
""
],
[
"Shahzaad",
"Babar",
""
],
[
"Lakhdari",
"Abdallah",
""
],
[
"Bouguettaya",
"Athman",
""
]
] | TITLE: Drone Formation for Efficient Swarm Energy Consumption
ABSTRACT: We demonstrate formation flying for drone swarm services. A set of drones fly
in four different swarm formations. A dataset is collected to study the effect
of formation flying on energy consumption. We conduct a set of experiments to
study the effect of wind on formation flying. We examine the forces the drones
exert on each other when flying in a formation. We finally identify and
classify the formations that conserve most energy under varying wind
conditions. The collected dataset aims at providing researchers data to conduct
further research in swarm-based drone service delivery. Demo:
https://youtu.be/NnucUWhUwLs
| new_dataset | 0.95096 |
2303.09697 | Wenceslao Arroyo-Machado | Pilar Valderrama-Baca, Wenceslao Arroyo-Machado, Daniel Torres-Salinas | Altmetrics can capture research evidence: a study across types of
studies in COVID-19 literature | null | Profesional De La informaci\'on 32 (2023) | 10.3145/epi.2023.mar.13 | null | cs.DL physics.soc-ph | http://creativecommons.org/licenses/by/4.0/ | There has been a proliferation of descriptive for COVID-19 papers using
altmetrics. The main objective of this study is to analyse whether the
altmetric mentions of COVID-19 medical studies are associated with the type of
study and its level of evidence. Data were collected from PubMed and
Altmetric.com databases. A total of 16,672 study types (e.g., Case reports or
Clinical trials) published in the year 2021 and with at least one altmetric
mention were retrieved. The altmetric indicators considered were Altmetric
Attention Score (AAS), News mentions, Twitter mentions, and Mendeley readers.
Once the dataset had been created, the first step was to carry out a
descriptive study. Then a normality hypothesis was contrasted by means of the
Kolmogorov-Smirnov test, and since it was significant in all cases, the overall
comparison of groups was performed using the non-parametric Kruskal-Wallis
test. When this test rejected the null hypothesis, pair-by-pair comparisons
were performed with the Mann-Whitney U test, and the intensity of the possible
association was measured using Cramers V coefficient. The results suggest that
the data do not fit a normal distribution. The Mann-Whitney U test revealed
coincidences in five groups of study types, the altmetric indicator with most
coincidences being news mentions and the study types with the most coincidences
were the systematic reviews together with the meta-analyses, which coincided
with four altmetric indicators. Likewise, between the study types and the
altmetric indicators, a weak but significant association was observed through
the chi-square and Cramers V. It is concluded that the positive association
between altmetrics and study types in medicine could reflect the level of the
pyramid of scientific evidence.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 00:05:44 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Valderrama-Baca",
"Pilar",
""
],
[
"Arroyo-Machado",
"Wenceslao",
""
],
[
"Torres-Salinas",
"Daniel",
""
]
] | TITLE: Altmetrics can capture research evidence: a study across types of
studies in COVID-19 literature
ABSTRACT: There has been a proliferation of descriptive for COVID-19 papers using
altmetrics. The main objective of this study is to analyse whether the
altmetric mentions of COVID-19 medical studies are associated with the type of
study and its level of evidence. Data were collected from PubMed and
Altmetric.com databases. A total of 16,672 study types (e.g., Case reports or
Clinical trials) published in the year 2021 and with at least one altmetric
mention were retrieved. The altmetric indicators considered were Altmetric
Attention Score (AAS), News mentions, Twitter mentions, and Mendeley readers.
Once the dataset had been created, the first step was to carry out a
descriptive study. Then a normality hypothesis was contrasted by means of the
Kolmogorov-Smirnov test, and since it was significant in all cases, the overall
comparison of groups was performed using the non-parametric Kruskal-Wallis
test. When this test rejected the null hypothesis, pair-by-pair comparisons
were performed with the Mann-Whitney U test, and the intensity of the possible
association was measured using Cramers V coefficient. The results suggest that
the data do not fit a normal distribution. The Mann-Whitney U test revealed
coincidences in five groups of study types, the altmetric indicator with most
coincidences being news mentions and the study types with the most coincidences
were the systematic reviews together with the meta-analyses, which coincided
with four altmetric indicators. Likewise, between the study types and the
altmetric indicators, a weak but significant association was observed through
the chi-square and Cramers V. It is concluded that the positive association
between altmetrics and study types in medicine could reflect the level of the
pyramid of scientific evidence.
| no_new_dataset | 0.941975 |
2303.09703 | Ahmed Shoyeb Raihan | Ahmed Shoyeb Raihan and Imtiaz Ahmed | A Bi-LSTM Autoencoder Framework for Anomaly Detection -- A Case Study of
a Wind Power Dataset | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Anomalies refer to data points or events that deviate from normal and
homogeneous events, which can include fraudulent activities, network
infiltrations, equipment malfunctions, process changes, or other significant
but infrequent events. Prompt detection of such events can prevent potential
losses in terms of finances, information, and human resources. With the
advancement of computational capabilities and the availability of large
datasets, anomaly detection has become a major area of research. Among these,
anomaly detection in time series has gained more attention recently due to the
added complexity imposed by the time dimension. This study presents a novel
framework for time series anomaly detection using a combination of
Bidirectional Long Short Term Memory (Bi-LSTM) architecture and Autoencoder.
The Bi-LSTM network, which comprises two unidirectional LSTM networks, can
analyze the time series data from both directions and thus effectively discover
the long-term dependencies hidden in the sequential data. Meanwhile, the
Autoencoder mechanism helps to establish the optimal threshold beyond which an
event can be classified as an anomaly. To demonstrate the effectiveness of the
proposed framework, it is applied to a real-world multivariate time series
dataset collected from a wind farm. The Bi-LSTM Autoencoder model achieved a
classification accuracy of 96.79% and outperformed more commonly used LSTM
Autoencoder models.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 00:24:28 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Raihan",
"Ahmed Shoyeb",
""
],
[
"Ahmed",
"Imtiaz",
""
]
] | TITLE: A Bi-LSTM Autoencoder Framework for Anomaly Detection -- A Case Study of
a Wind Power Dataset
ABSTRACT: Anomalies refer to data points or events that deviate from normal and
homogeneous events, which can include fraudulent activities, network
infiltrations, equipment malfunctions, process changes, or other significant
but infrequent events. Prompt detection of such events can prevent potential
losses in terms of finances, information, and human resources. With the
advancement of computational capabilities and the availability of large
datasets, anomaly detection has become a major area of research. Among these,
anomaly detection in time series has gained more attention recently due to the
added complexity imposed by the time dimension. This study presents a novel
framework for time series anomaly detection using a combination of
Bidirectional Long Short Term Memory (Bi-LSTM) architecture and Autoencoder.
The Bi-LSTM network, which comprises two unidirectional LSTM networks, can
analyze the time series data from both directions and thus effectively discover
the long-term dependencies hidden in the sequential data. Meanwhile, the
Autoencoder mechanism helps to establish the optimal threshold beyond which an
event can be classified as an anomaly. To demonstrate the effectiveness of the
proposed framework, it is applied to a real-world multivariate time series
dataset collected from a wind farm. The Bi-LSTM Autoencoder model achieved a
classification accuracy of 96.79% and outperformed more commonly used LSTM
Autoencoder models.
| no_new_dataset | 0.943243 |
2303.09719 | Xiuying Chen | Xiuying Chen, Mingzhe Li, Jiayi Zhang, Xiaoqiang Xia, Chen Wei,
Jianwei Cui, Xin Gao, Xiangliang Zhang, Rui Yan | Learning towards Selective Data Augmentation for Dialogue Generation | 9 pages, 4 figures, AAAI 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | As it is cumbersome and expensive to acquire a huge amount of data for
training neural dialog models, data augmentation is proposed to effectively
utilize existing training samples. However, current data augmentation
techniques on the dialog generation task mostly augment all cases in the
training dataset without considering the intrinsic attributes between different
cases. We argue that not all cases are beneficial for augmentation task, and
the cases suitable for augmentation should obey the following two attributes:
(1) low-quality (the dialog model cannot generate a high-quality response for
the case), (2) representative (the case should represent the property of the
whole dataset). Herein, we explore this idea by proposing a Selective Data
Augmentation framework (SDA) for the response generation task. SDA employs a
dual adversarial network to select the lowest quality and most representative
data points for augmentation in one stage. Extensive experiments conducted on
two publicly available datasets, i.e., DailyDialog and OpenSubtitles, show that
our framework can improve the response generation performance with respect to
various metrics.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 01:26:39 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Chen",
"Xiuying",
""
],
[
"Li",
"Mingzhe",
""
],
[
"Zhang",
"Jiayi",
""
],
[
"Xia",
"Xiaoqiang",
""
],
[
"Wei",
"Chen",
""
],
[
"Cui",
"Jianwei",
""
],
[
"Gao",
"Xin",
""
],
[
"Zhang",
"Xiangliang",
""
],
[
"Yan",
"Rui",
""
]
] | TITLE: Learning towards Selective Data Augmentation for Dialogue Generation
ABSTRACT: As it is cumbersome and expensive to acquire a huge amount of data for
training neural dialog models, data augmentation is proposed to effectively
utilize existing training samples. However, current data augmentation
techniques on the dialog generation task mostly augment all cases in the
training dataset without considering the intrinsic attributes between different
cases. We argue that not all cases are beneficial for augmentation task, and
the cases suitable for augmentation should obey the following two attributes:
(1) low-quality (the dialog model cannot generate a high-quality response for
the case), (2) representative (the case should represent the property of the
whole dataset). Herein, we explore this idea by proposing a Selective Data
Augmentation framework (SDA) for the response generation task. SDA employs a
dual adversarial network to select the lowest quality and most representative
data points for augmentation in one stage. Extensive experiments conducted on
two publicly available datasets, i.e., DailyDialog and OpenSubtitles, show that
our framework can improve the response generation performance with respect to
various metrics.
| no_new_dataset | 0.952042 |
2303.09733 | Zhengyi Liu | Zhengyi Liu, Xiaoshen Huang, Guanghui Zhang, Xianyong Fang, Linbo
Wang, Bin Tang | Scribble-Supervised RGB-T Salient Object Detection | ICME2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Salient object detection segments attractive objects in scenes. RGB and
thermal modalities provide complementary information and scribble annotations
alleviate large amounts of human labor. Based on the above facts, we propose a
scribble-supervised RGB-T salient object detection model. By a four-step
solution (expansion, prediction, aggregation, and supervision), label-sparse
challenge of scribble-supervised method is solved. To expand scribble
annotations, we collect the superpixels that foreground scribbles pass through
in RGB and thermal images, respectively. The expanded multi-modal labels
provide the coarse object boundary. To further polish the expanded labels, we
propose a prediction module to alleviate the sharpness of boundary. To play the
complementary roles of two modalities, we combine the two into aggregated
pseudo labels. Supervised by scribble annotations and pseudo labels, our model
achieves the state-of-the-art performance on the relabeled RGBT-S dataset.
Furthermore, the model is applied to RGB-D and video scribble-supervised
applications, achieving consistently excellent performance.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 02:27:59 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Liu",
"Zhengyi",
""
],
[
"Huang",
"Xiaoshen",
""
],
[
"Zhang",
"Guanghui",
""
],
[
"Fang",
"Xianyong",
""
],
[
"Wang",
"Linbo",
""
],
[
"Tang",
"Bin",
""
]
] | TITLE: Scribble-Supervised RGB-T Salient Object Detection
ABSTRACT: Salient object detection segments attractive objects in scenes. RGB and
thermal modalities provide complementary information and scribble annotations
alleviate large amounts of human labor. Based on the above facts, we propose a
scribble-supervised RGB-T salient object detection model. By a four-step
solution (expansion, prediction, aggregation, and supervision), label-sparse
challenge of scribble-supervised method is solved. To expand scribble
annotations, we collect the superpixels that foreground scribbles pass through
in RGB and thermal images, respectively. The expanded multi-modal labels
provide the coarse object boundary. To further polish the expanded labels, we
propose a prediction module to alleviate the sharpness of boundary. To play the
complementary roles of two modalities, we combine the two into aggregated
pseudo labels. Supervised by scribble annotations and pseudo labels, our model
achieves the state-of-the-art performance on the relabeled RGBT-S dataset.
Furthermore, the model is applied to RGB-D and video scribble-supervised
applications, achieving consistently excellent performance.
| no_new_dataset | 0.95096 |
2303.09736 | Jun-Hyung Park | Jun-Hyung Park, Yeachan Kim, Junho Kim, Joon-Young Choi, SangKeun Lee | Dynamic Structure Pruning for Compressing CNNs | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Structure pruning is an effective method to compress and accelerate neural
networks. While filter and channel pruning are preferable to other structure
pruning methods in terms of realistic acceleration and hardware compatibility,
pruning methods with a finer granularity, such as intra-channel pruning, are
expected to be capable of yielding more compact and computationally efficient
networks. Typical intra-channel pruning methods utilize a static and
hand-crafted pruning granularity due to a large search space, which leaves room
for improvement in their pruning performance. In this work, we introduce a
novel structure pruning method, termed as dynamic structure pruning, to
identify optimal pruning granularities for intra-channel pruning. In contrast
to existing intra-channel pruning methods, the proposed method automatically
optimizes dynamic pruning granularities in each layer while training deep
neural networks. To achieve this, we propose a differentiable group learning
method designed to efficiently learn a pruning granularity based on
gradient-based learning of filter groups. The experimental results show that
dynamic structure pruning achieves state-of-the-art pruning performance and
better realistic acceleration on a GPU compared with channel pruning. In
particular, it reduces the FLOPs of ResNet50 by 71.85% without accuracy
degradation on the ImageNet dataset. Our code is available at
https://github.com/irishev/DSP.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 02:38:53 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Park",
"Jun-Hyung",
""
],
[
"Kim",
"Yeachan",
""
],
[
"Kim",
"Junho",
""
],
[
"Choi",
"Joon-Young",
""
],
[
"Lee",
"SangKeun",
""
]
] | TITLE: Dynamic Structure Pruning for Compressing CNNs
ABSTRACT: Structure pruning is an effective method to compress and accelerate neural
networks. While filter and channel pruning are preferable to other structure
pruning methods in terms of realistic acceleration and hardware compatibility,
pruning methods with a finer granularity, such as intra-channel pruning, are
expected to be capable of yielding more compact and computationally efficient
networks. Typical intra-channel pruning methods utilize a static and
hand-crafted pruning granularity due to a large search space, which leaves room
for improvement in their pruning performance. In this work, we introduce a
novel structure pruning method, termed as dynamic structure pruning, to
identify optimal pruning granularities for intra-channel pruning. In contrast
to existing intra-channel pruning methods, the proposed method automatically
optimizes dynamic pruning granularities in each layer while training deep
neural networks. To achieve this, we propose a differentiable group learning
method designed to efficiently learn a pruning granularity based on
gradient-based learning of filter groups. The experimental results show that
dynamic structure pruning achieves state-of-the-art pruning performance and
better realistic acceleration on a GPU compared with channel pruning. In
particular, it reduces the FLOPs of ResNet50 by 71.85% without accuracy
degradation on the ImageNet dataset. Our code is available at
https://github.com/irishev/DSP.
| no_new_dataset | 0.947088 |
2303.09743 | Tzu-Sheng Kuo | Tzu-Sheng Kuo, Hong Shen, Jisoo Geum, Nev Jones, Jason I. Hong, Haiyi
Zhu, Kenneth Holstein | Understanding Frontline Workers' and Unhoused Individuals' Perspectives
on AI Used in Homeless Services | null | Proceedings of the 2023 CHI Conference on Human Factors in
Computing Systems (CHI '23) | 10.1145/3544548.3580882 | null | cs.HC cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent years have seen growing adoption of AI-based decision-support systems
(ADS) in homeless services, yet we know little about stakeholder desires and
concerns surrounding their use. In this work, we aim to understand impacted
stakeholders' perspectives on a deployed ADS that prioritizes scarce housing
resources. We employed AI lifecycle comicboarding, an adapted version of the
comicboarding method, to elicit stakeholder feedback and design ideas across
various components of an AI system's design. We elicited feedback from county
workers who operate the ADS daily, service providers whose work is directly
impacted by the ADS, and unhoused individuals in the region. Our participants
shared concerns and design suggestions around the AI system's overall
objective, specific model design choices, dataset selection, and use in
deployment. Our findings demonstrate that stakeholders, even without AI
knowledge, can provide specific and critical feedback on an AI system's design
and deployment, if empowered to do so.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 02:46:45 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Kuo",
"Tzu-Sheng",
""
],
[
"Shen",
"Hong",
""
],
[
"Geum",
"Jisoo",
""
],
[
"Jones",
"Nev",
""
],
[
"Hong",
"Jason I.",
""
],
[
"Zhu",
"Haiyi",
""
],
[
"Holstein",
"Kenneth",
""
]
] | TITLE: Understanding Frontline Workers' and Unhoused Individuals' Perspectives
on AI Used in Homeless Services
ABSTRACT: Recent years have seen growing adoption of AI-based decision-support systems
(ADS) in homeless services, yet we know little about stakeholder desires and
concerns surrounding their use. In this work, we aim to understand impacted
stakeholders' perspectives on a deployed ADS that prioritizes scarce housing
resources. We employed AI lifecycle comicboarding, an adapted version of the
comicboarding method, to elicit stakeholder feedback and design ideas across
various components of an AI system's design. We elicited feedback from county
workers who operate the ADS daily, service providers whose work is directly
impacted by the ADS, and unhoused individuals in the region. Our participants
shared concerns and design suggestions around the AI system's overall
objective, specific model design choices, dataset selection, and use in
deployment. Our findings demonstrate that stakeholders, even without AI
knowledge, can provide specific and critical feedback on an AI system's design
and deployment, if empowered to do so.
| no_new_dataset | 0.951369 |
2303.09746 | Jinggang Chen | Jinggang Chen, Xiaoyang Qu, Junjie Li, Jianzong Wang, Jiguang Wan,
Jing Xiao | Detecting Out-of-distribution Examples via Class-conditional Impressions
Reappearing | Accepted by ICASSP 2023 | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Out-of-distribution (OOD) detection aims at enhancing standard deep neural
networks to distinguish anomalous inputs from original training data. Previous
progress has introduced various approaches where the in-distribution training
data and even several OOD examples are prerequisites. However, due to privacy
and security, auxiliary data tends to be impractical in a real-world scenario.
In this paper, we propose a data-free method without training on natural data,
called Class-Conditional Impressions Reappearing (C2IR), which utilizes image
impressions from the fixed model to recover class-conditional feature
statistics. Based on that, we introduce Integral Probability Metrics to
estimate layer-wise class-conditional deviations and obtain layer weights by
Measuring Gradient-based Importance (MGI). The experiments verify the
effectiveness of our method and indicate that C2IR outperforms other post-hoc
methods and reaches comparable performance to the full access (ID and OOD)
detection method, especially in the far-OOD dataset (SVHN).
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 02:55:08 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Chen",
"Jinggang",
""
],
[
"Qu",
"Xiaoyang",
""
],
[
"Li",
"Junjie",
""
],
[
"Wang",
"Jianzong",
""
],
[
"Wan",
"Jiguang",
""
],
[
"Xiao",
"Jing",
""
]
] | TITLE: Detecting Out-of-distribution Examples via Class-conditional Impressions
Reappearing
ABSTRACT: Out-of-distribution (OOD) detection aims at enhancing standard deep neural
networks to distinguish anomalous inputs from original training data. Previous
progress has introduced various approaches where the in-distribution training
data and even several OOD examples are prerequisites. However, due to privacy
and security, auxiliary data tends to be impractical in a real-world scenario.
In this paper, we propose a data-free method without training on natural data,
called Class-Conditional Impressions Reappearing (C2IR), which utilizes image
impressions from the fixed model to recover class-conditional feature
statistics. Based on that, we introduce Integral Probability Metrics to
estimate layer-wise class-conditional deviations and obtain layer weights by
Measuring Gradient-based Importance (MGI). The experiments verify the
effectiveness of our method and indicate that C2IR outperforms other post-hoc
methods and reaches comparable performance to the full access (ID and OOD)
detection method, especially in the far-OOD dataset (SVHN).
| no_new_dataset | 0.948965 |
2303.09755 | Amjed Tahir | Shawn Rasheed, Jens Dietrich, Amjed Tahir | On the Effect of Instrumentation on Test Flakiness | To appear: In Proceedings of the 2023 ACM/IEEE International
Conference on Automation of Software Test (AST) | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Test flakiness is a problem that affects testing and processes that rely on
it. Several factors cause or influence the flakiness of test outcomes. Test
execution order, randomness and concurrency are some of the more common and
well-studied causes. Some studies mention code instrumentation as a factor that
causes or affects test flakiness. However, evidence for this issue is scarce.
In this study, we attempt to systematically collect evidence for the effects of
instrumentation on test flakiness. We experiment with common types of
instrumentation for Java programs - namely, application performance monitoring,
coverage and profiling instrumentation. We then study the effects of
instrumentation on a set of nine programs obtained from an existing dataset
used to study test flakiness, consisting of popular GitHub projects written in
Java. We observe cases where real-world instrumentation causes flakiness in a
program. However, this effect is rare. We also discuss a related issue - how
instrumentation may interfere with flakiness detection and prevention.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 03:39:32 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Rasheed",
"Shawn",
""
],
[
"Dietrich",
"Jens",
""
],
[
"Tahir",
"Amjed",
""
]
] | TITLE: On the Effect of Instrumentation on Test Flakiness
ABSTRACT: Test flakiness is a problem that affects testing and processes that rely on
it. Several factors cause or influence the flakiness of test outcomes. Test
execution order, randomness and concurrency are some of the more common and
well-studied causes. Some studies mention code instrumentation as a factor that
causes or affects test flakiness. However, evidence for this issue is scarce.
In this study, we attempt to systematically collect evidence for the effects of
instrumentation on test flakiness. We experiment with common types of
instrumentation for Java programs - namely, application performance monitoring,
coverage and profiling instrumentation. We then study the effects of
instrumentation on a set of nine programs obtained from an existing dataset
used to study test flakiness, consisting of popular GitHub projects written in
Java. We observe cases where real-world instrumentation causes flakiness in a
program. However, this effect is rare. We also discuss a related issue - how
instrumentation may interfere with flakiness detection and prevention.
| new_dataset | 0.877319 |
2303.09757 | Jiaqi Xu | Jiaqi Xu, Xiaowei Hu, Lei Zhu, Qi Dou, Jifeng Dai, Yu Qiao, Pheng-Ann
Heng | Video Dehazing via a Multi-Range Temporal Alignment Network with
Physical Prior | CVPR 2023; Code: https://github.com/jiaqixuac/MAP-Net | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video dehazing aims to recover haze-free frames with high visibility and
contrast. This paper presents a novel framework to effectively explore the
physical haze priors and aggregate temporal information. Specifically, we
design a memory-based physical prior guidance module to encode the
prior-related features into long-range memory. Besides, we formulate a
multi-range scene radiance recovery module to capture space-time dependencies
in multiple space-time ranges, which helps to effectively aggregate temporal
information from adjacent frames. Moreover, we construct the first large-scale
outdoor video dehazing benchmark dataset, which contains videos in various
real-world scenarios. Experimental results on both synthetic and real
conditions show the superiority of our proposed method.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 03:44:17 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Xu",
"Jiaqi",
""
],
[
"Hu",
"Xiaowei",
""
],
[
"Zhu",
"Lei",
""
],
[
"Dou",
"Qi",
""
],
[
"Dai",
"Jifeng",
""
],
[
"Qiao",
"Yu",
""
],
[
"Heng",
"Pheng-Ann",
""
]
] | TITLE: Video Dehazing via a Multi-Range Temporal Alignment Network with
Physical Prior
ABSTRACT: Video dehazing aims to recover haze-free frames with high visibility and
contrast. This paper presents a novel framework to effectively explore the
physical haze priors and aggregate temporal information. Specifically, we
design a memory-based physical prior guidance module to encode the
prior-related features into long-range memory. Besides, we formulate a
multi-range scene radiance recovery module to capture space-time dependencies
in multiple space-time ranges, which helps to effectively aggregate temporal
information from adjacent frames. Moreover, we construct the first large-scale
outdoor video dehazing benchmark dataset, which contains videos in various
real-world scenarios. Experimental results on both synthetic and real
conditions show the superiority of our proposed method.
| new_dataset | 0.949995 |
2303.09772 | Koichiro Yawata | Koichiro Yawata, Yoshihiro Osakabe, Takuya Okuyama, Akinori Asahara | QUBO Decision Tree: Annealing Machine Extends Decision Tree Splitting | 2022 IEEE International Conference on Knowledge Graph (ICKG) | null | 10.1109/ICKG55886.2022.00052 | null | cs.LG physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes an extension of regression trees by quadratic
unconstrained binary optimization (QUBO). Regression trees are very popular
prediction models that are trainable with tabular datasets, but their accuracy
is insufficient because the decision rules are too simple. The proposed method
extends the decision rules in decision trees to multi-dimensional boundaries.
Such an extension is generally unimplementable because of computational
limitations, however, the proposed method transforms the training process to
QUBO, which enables an annealing machine to solve this problem.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 04:40:07 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yawata",
"Koichiro",
""
],
[
"Osakabe",
"Yoshihiro",
""
],
[
"Okuyama",
"Takuya",
""
],
[
"Asahara",
"Akinori",
""
]
] | TITLE: QUBO Decision Tree: Annealing Machine Extends Decision Tree Splitting
ABSTRACT: This paper proposes an extension of regression trees by quadratic
unconstrained binary optimization (QUBO). Regression trees are very popular
prediction models that are trainable with tabular datasets, but their accuracy
is insufficient because the decision rules are too simple. The proposed method
extends the decision rules in decision trees to multi-dimensional boundaries.
Such an extension is generally unimplementable because of computational
limitations, however, the proposed method transforms the training process to
QUBO, which enables an annealing machine to solve this problem.
| no_new_dataset | 0.95222 |
2303.09774 | Zhaoheng Li | Zhaoheng Li, Xinyu Pi, Yongjoo Park | S/C: Speeding up Data Materialization with Bounded Memory | Accepted to ICDE 2023 | null | null | null | cs.DB | http://creativecommons.org/licenses/by/4.0/ | With data pipeline tools and the expressiveness of SQL, managing
interdependent materialized views (MVs) are becoming increasingly easy. These
MVs are updated repeatedly upon new data ingestion (e.g., daily), from which
database admins can observe performance metrics (e.g., refresh time of each MV,
size on disk) in a consistent way for different types of updates (full vs.
incremental) and for different systems (single node, distributed,
cloud-hosted). One missed opportunity is that existing data systems treat those
MV updates as independent SQL statements without fully exploiting their
dependency information and performance metrics. However, if we know that the
result of a SQL statement will be consumed immediately after for subsequent
operations, those subsequent operations do not have to wait until the early
results are fully materialized on storage because the results are already
readily available in memory. Of course, this may come at a cost because keeping
results in memory (even temporarily) will reduce the amount of available
memory; thus, our decision should be careful. In this paper, we introduce a new
system, called S/C, which tackles this problem through efficient creation and
update of a set of MVs with acyclic dependencies among them. S/C judiciously
uses bounded memory to reduce end-to-end MV refresh time by short-circuiting
expensive reads and writes; S/C's objective function accurately estimates time
savings from keeping intermediate data in memory for particular periods. Our
solution jointly optimizes an MV refresh order, what data to keep in memory,
and when to release data from memory. At a high level, S/C still materializes
all data exactly as defined in MV definitions; thus, it doesn't impact any
service-level agreements. In our experiments with TPC-DS datasets (up to 1TB),
we show S/C's optimization can speedup end-to-end runtime by 1.04x-5.08x with
1.6GB memory.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 04:50:03 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Li",
"Zhaoheng",
""
],
[
"Pi",
"Xinyu",
""
],
[
"Park",
"Yongjoo",
""
]
] | TITLE: S/C: Speeding up Data Materialization with Bounded Memory
ABSTRACT: With data pipeline tools and the expressiveness of SQL, managing
interdependent materialized views (MVs) are becoming increasingly easy. These
MVs are updated repeatedly upon new data ingestion (e.g., daily), from which
database admins can observe performance metrics (e.g., refresh time of each MV,
size on disk) in a consistent way for different types of updates (full vs.
incremental) and for different systems (single node, distributed,
cloud-hosted). One missed opportunity is that existing data systems treat those
MV updates as independent SQL statements without fully exploiting their
dependency information and performance metrics. However, if we know that the
result of a SQL statement will be consumed immediately after for subsequent
operations, those subsequent operations do not have to wait until the early
results are fully materialized on storage because the results are already
readily available in memory. Of course, this may come at a cost because keeping
results in memory (even temporarily) will reduce the amount of available
memory; thus, our decision should be careful. In this paper, we introduce a new
system, called S/C, which tackles this problem through efficient creation and
update of a set of MVs with acyclic dependencies among them. S/C judiciously
uses bounded memory to reduce end-to-end MV refresh time by short-circuiting
expensive reads and writes; S/C's objective function accurately estimates time
savings from keeping intermediate data in memory for particular periods. Our
solution jointly optimizes an MV refresh order, what data to keep in memory,
and when to release data from memory. At a high level, S/C still materializes
all data exactly as defined in MV definitions; thus, it doesn't impact any
service-level agreements. In our experiments with TPC-DS datasets (up to 1TB),
we show S/C's optimization can speedup end-to-end runtime by 1.04x-5.08x with
1.6GB memory.
| no_new_dataset | 0.942876 |
2303.09779 | Daehan Kim | Daehan Kim, Minseok Seo, Kwanyong Park, Inkyu Shin, Sanghyun Woo,
In-So Kweon, Dong-Geol Choi | Bidirectional Domain Mixup for Domain Adaptive Semantic Segmentation | 10 pages, 3 figures, Accepted on AAAI 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Mixup provides interpolated training samples and allows the model to obtain
smoother decision boundaries for better generalization. The idea can be
naturally applied to the domain adaptation task, where we can mix the source
and target samples to obtain domain-mixed samples for better adaptation.
However, the extension of the idea from classification to segmentation (i.e.,
structured output) is nontrivial. This paper systematically studies the impact
of mixup under the domain adaptaive semantic segmentation task and presents a
simple yet effective mixup strategy called Bidirectional Domain Mixup (BDM). In
specific, we achieve domain mixup in two-step: cut and paste. Given the warm-up
model trained from any adaptation techniques, we forward the source and target
samples and perform a simple threshold-based cut out of the unconfident regions
(cut). After then, we fill-in the dropped regions with the other domain region
patches (paste). In doing so, we jointly consider class distribution, spatial
structure, and pseudo label confidence. Based on our analysis, we found that
BDM leaves domain transferable regions by cutting, balances the dataset-level
class distribution while preserving natural scene context by pasting. We
coupled our proposal with various state-of-the-art adaptation models and
observe significant improvement consistently. We also provide extensive
ablation experiments to empirically verify our main components of the
framework. Visit our project page with the code at
https://sites.google.com/view/bidirectional-domain-mixup
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 05:22:44 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Kim",
"Daehan",
""
],
[
"Seo",
"Minseok",
""
],
[
"Park",
"Kwanyong",
""
],
[
"Shin",
"Inkyu",
""
],
[
"Woo",
"Sanghyun",
""
],
[
"Kweon",
"In-So",
""
],
[
"Choi",
"Dong-Geol",
""
]
] | TITLE: Bidirectional Domain Mixup for Domain Adaptive Semantic Segmentation
ABSTRACT: Mixup provides interpolated training samples and allows the model to obtain
smoother decision boundaries for better generalization. The idea can be
naturally applied to the domain adaptation task, where we can mix the source
and target samples to obtain domain-mixed samples for better adaptation.
However, the extension of the idea from classification to segmentation (i.e.,
structured output) is nontrivial. This paper systematically studies the impact
of mixup under the domain adaptaive semantic segmentation task and presents a
simple yet effective mixup strategy called Bidirectional Domain Mixup (BDM). In
specific, we achieve domain mixup in two-step: cut and paste. Given the warm-up
model trained from any adaptation techniques, we forward the source and target
samples and perform a simple threshold-based cut out of the unconfident regions
(cut). After then, we fill-in the dropped regions with the other domain region
patches (paste). In doing so, we jointly consider class distribution, spatial
structure, and pseudo label confidence. Based on our analysis, we found that
BDM leaves domain transferable regions by cutting, balances the dataset-level
class distribution while preserving natural scene context by pasting. We
coupled our proposal with various state-of-the-art adaptation models and
observe significant improvement consistently. We also provide extensive
ablation experiments to empirically verify our main components of the
framework. Visit our project page with the code at
https://sites.google.com/view/bidirectional-domain-mixup
| no_new_dataset | 0.951639 |
2303.09782 | Huy Hieu Pham | Anh Duy Nguyen, Huy Hieu Pham, Huynh Thanh Trung, Quoc Viet Hung
Nguyen, Thao Nguyen Truong, Phi Le Nguyen | High Accurate and Explainable Multi-Pill Detection Framework with Graph
Neural Network-Assisted Multimodal Data Fusion | Under review by Plos ONE journal | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Due to the significant resemblance in visual appearance, pill misuse is
prevalent and has become a critical issue, responsible for one-third of all
deaths worldwide. Pill identification, thus, is a crucial concern needed to be
investigated thoroughly. Recently, several attempts have been made to exploit
deep learning to tackle the pill identification problem. However, most
published works consider only single-pill identification and fail to
distinguish hard samples with identical appearances. Also, most existing pill
image datasets only feature single pill images captured in carefully controlled
environments under ideal lighting conditions and clean backgrounds. In this
work, we are the first to tackle the multi-pill detection problem in real-world
settings, aiming at localizing and identifying pills captured by users in a
pill intake. Moreover, we also introduce a multi-pill image dataset taken in
unconstrained conditions. To handle hard samples, we propose a novel method for
constructing heterogeneous a priori graphs incorporating three forms of
inter-pill relationships, including co-occurrence likelihood, relative size,
and visual semantic correlation. We then offer a framework for integrating a
priori with pills' visual features to enhance detection accuracy. Our
experimental results have proved the robustness, reliability, and
explainability of the proposed framework. Experimentally, it outperforms all
detection benchmarks in terms of all evaluation metrics. Specifically, our
proposed framework improves COCO mAP metrics by 9.4% over Faster R-CNN and
12.0% compared to vanilla YOLOv5. Our study opens up new opportunities for
protecting patients from medication errors using an AI-based pill
identification solution.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 05:48:26 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Nguyen",
"Anh Duy",
""
],
[
"Pham",
"Huy Hieu",
""
],
[
"Trung",
"Huynh Thanh",
""
],
[
"Nguyen",
"Quoc Viet Hung",
""
],
[
"Truong",
"Thao Nguyen",
""
],
[
"Nguyen",
"Phi Le",
""
]
] | TITLE: High Accurate and Explainable Multi-Pill Detection Framework with Graph
Neural Network-Assisted Multimodal Data Fusion
ABSTRACT: Due to the significant resemblance in visual appearance, pill misuse is
prevalent and has become a critical issue, responsible for one-third of all
deaths worldwide. Pill identification, thus, is a crucial concern needed to be
investigated thoroughly. Recently, several attempts have been made to exploit
deep learning to tackle the pill identification problem. However, most
published works consider only single-pill identification and fail to
distinguish hard samples with identical appearances. Also, most existing pill
image datasets only feature single pill images captured in carefully controlled
environments under ideal lighting conditions and clean backgrounds. In this
work, we are the first to tackle the multi-pill detection problem in real-world
settings, aiming at localizing and identifying pills captured by users in a
pill intake. Moreover, we also introduce a multi-pill image dataset taken in
unconstrained conditions. To handle hard samples, we propose a novel method for
constructing heterogeneous a priori graphs incorporating three forms of
inter-pill relationships, including co-occurrence likelihood, relative size,
and visual semantic correlation. We then offer a framework for integrating a
priori with pills' visual features to enhance detection accuracy. Our
experimental results have proved the robustness, reliability, and
explainability of the proposed framework. Experimentally, it outperforms all
detection benchmarks in terms of all evaluation metrics. Specifically, our
proposed framework improves COCO mAP metrics by 9.4% over Faster R-CNN and
12.0% compared to vanilla YOLOv5. Our study opens up new opportunities for
protecting patients from medication errors using an AI-based pill
identification solution.
| no_new_dataset | 0.920146 |
2303.09794 | Yuhao Lin | Yuhao Lin, Haiming Xu, Lingqiao Liu, Jinan Zou, Javen Qinfeng Shi | Revisiting Image Reconstruction for Semi-supervised Semantic
Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Autoencoding, which aims to reconstruct the input images through a bottleneck
latent representation, is one of the classic feature representation learning
strategies. It has been shown effective as an auxiliary task for
semi-supervised learning but has become less popular as more sophisticated
methods have been proposed in recent years. In this paper, we revisit the idea
of using image reconstruction as the auxiliary task and incorporate it with a
modern semi-supervised semantic segmentation framework. Surprisingly, we
discover that such an old idea in semi-supervised learning can produce results
competitive with state-of-the-art semantic segmentation algorithms. By
visualizing the intermediate layer activations of the image reconstruction
module, we show that the feature map channel could correlate well with the
semantic concept, which explains why joint training with the reconstruction
task is helpful for the segmentation task. Motivated by our observation, we
further proposed a modification to the image reconstruction task, aiming to
further disentangle the object clue from the background patterns. From
experiment evaluation on various datasets, we show that using reconstruction as
auxiliary loss can lead to consistent improvements in various datasets and
methods. The proposed method can further lead to significant improvement in
object-centric segmentation tasks.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 06:31:06 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Lin",
"Yuhao",
""
],
[
"Xu",
"Haiming",
""
],
[
"Liu",
"Lingqiao",
""
],
[
"Zou",
"Jinan",
""
],
[
"Shi",
"Javen Qinfeng",
""
]
] | TITLE: Revisiting Image Reconstruction for Semi-supervised Semantic
Segmentation
ABSTRACT: Autoencoding, which aims to reconstruct the input images through a bottleneck
latent representation, is one of the classic feature representation learning
strategies. It has been shown effective as an auxiliary task for
semi-supervised learning but has become less popular as more sophisticated
methods have been proposed in recent years. In this paper, we revisit the idea
of using image reconstruction as the auxiliary task and incorporate it with a
modern semi-supervised semantic segmentation framework. Surprisingly, we
discover that such an old idea in semi-supervised learning can produce results
competitive with state-of-the-art semantic segmentation algorithms. By
visualizing the intermediate layer activations of the image reconstruction
module, we show that the feature map channel could correlate well with the
semantic concept, which explains why joint training with the reconstruction
task is helpful for the segmentation task. Motivated by our observation, we
further proposed a modification to the image reconstruction task, aiming to
further disentangle the object clue from the background patterns. From
experiment evaluation on various datasets, we show that using reconstruction as
auxiliary loss can lead to consistent improvements in various datasets and
methods. The proposed method can further lead to significant improvement in
object-centric segmentation tasks.
| no_new_dataset | 0.946794 |
2303.09795 | Zhou Yang | Zhou Yang, Chenyu Wang, Jieke Shi, Thong Hoang, Pavneet Kochhar,
Qinghua Lu, Zhenchang Xing, David Lo | What Do Users Ask in Open-Source AI Repositories? An Empirical Study of
GitHub Issues | Accepted by MSR 2023, 13 Pages | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artificial Intelligence systems, which benefit from the availability of
large-scale datasets and increasing computational power, have become effective
solutions to various critical tasks, such as natural language understanding,
speech recognition, and image processing. The advancement of these AI systems
is inseparable from open-source software (OSS). This paper presents an
empirical study that investigates the issues in the repositories of open-source
AI repositories to assist developers in understanding problems during the
process of employing AI systems. We collect 576 repositories from the
PapersWithCode platform. Among these repositories, we find 24,953 issues by
utilizing GitHub REST APIs. Our empirical study includes three phases. First,
we manually analyze these issues to categorize the problems that developers are
likely to encounter in open-source AI repositories. Specifically, we provide a
taxonomy of 13 categories related to AI systems. The two most common issues are
runtime errors (23.18%) and unclear instructions (19.53%). Second, we see that
67.5% of issues are closed. We also find that half of these issues resolve
within four days. Moreover, issue management features, e.g., label and assign,
are not widely adopted in open-source AI repositories. In particular, only
7.81% and 5.9% of repositories label issues and assign these issues to
assignees, respectively. Finally, we empirically show that employing GitHub
issue management features and writing issues with detailed descriptions
facilitate the resolution of issues. Based on our findings, we make
recommendations for developers to help better manage the issues of open-source
AI repositories and improve their quality.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 06:31:13 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yang",
"Zhou",
""
],
[
"Wang",
"Chenyu",
""
],
[
"Shi",
"Jieke",
""
],
[
"Hoang",
"Thong",
""
],
[
"Kochhar",
"Pavneet",
""
],
[
"Lu",
"Qinghua",
""
],
[
"Xing",
"Zhenchang",
""
],
[
"Lo",
"David",
""
]
] | TITLE: What Do Users Ask in Open-Source AI Repositories? An Empirical Study of
GitHub Issues
ABSTRACT: Artificial Intelligence systems, which benefit from the availability of
large-scale datasets and increasing computational power, have become effective
solutions to various critical tasks, such as natural language understanding,
speech recognition, and image processing. The advancement of these AI systems
is inseparable from open-source software (OSS). This paper presents an
empirical study that investigates the issues in the repositories of open-source
AI repositories to assist developers in understanding problems during the
process of employing AI systems. We collect 576 repositories from the
PapersWithCode platform. Among these repositories, we find 24,953 issues by
utilizing GitHub REST APIs. Our empirical study includes three phases. First,
we manually analyze these issues to categorize the problems that developers are
likely to encounter in open-source AI repositories. Specifically, we provide a
taxonomy of 13 categories related to AI systems. The two most common issues are
runtime errors (23.18%) and unclear instructions (19.53%). Second, we see that
67.5% of issues are closed. We also find that half of these issues resolve
within four days. Moreover, issue management features, e.g., label and assign,
are not widely adopted in open-source AI repositories. In particular, only
7.81% and 5.9% of repositories label issues and assign these issues to
assignees, respectively. Finally, we empirically show that employing GitHub
issue management features and writing issues with detailed descriptions
facilitate the resolution of issues. Based on our findings, we make
recommendations for developers to help better manage the issues of open-source
AI repositories and improve their quality.
| no_new_dataset | 0.941761 |
2303.09808 | Mojtaba Shahin | Mojtaba Shahin, Mansooreh Zahedi, Hourieh Khalajzadeh, Ali Rezaei
Nasab | A Study of Gender Discussions in Mobile Apps | Preprint- Accepted for publication in 2023 IEEE/ACM 20th
International Conference on Mining Software Repositories (MSR) | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mobile software apps ("apps") are one of the prevailing digital technologies
that our modern life heavily depends on. A key issue in the development of apps
is how to design gender-inclusive apps. Apps that do not consider gender
inclusion, diversity, and equality in their design can create barriers (e.g.,
excluding some of the users because of their gender) for their diverse users.
While there have been some efforts to develop gender-inclusive apps, a lack of
deep understanding regarding user perspectives on gender may prevent app
developers and owners from identifying issues related to gender and proposing
solutions for improvement. Users express many different opinions about apps in
their reviews, from sharing their experiences, and reporting bugs, to
requesting new features. In this study, we aim at unpacking gender discussions
about apps from the user perspective by analysing app reviews. We first develop
and evaluate several Machine Learning (ML) and Deep Learning (DL) classifiers
that automatically detect gender reviews (i.e., reviews that contain
discussions about gender). We apply our ML and DL classifiers on a manually
constructed dataset of 1,440 app reviews from the Google App Store, composing
620 gender reviews and 820 non-gender reviews. Our best classifier achieves an
F1-score of 90.77%. Second, our qualitative analysis of a randomly selected 388
out of 620 gender reviews shows that gender discussions in app reviews revolve
around six topics: App Features, Appearance, Content, Company Policy and
Censorship, Advertisement, and Community. Finally, we provide some practical
implications and recommendations for developing gender-inclusive apps.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 07:32:38 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Shahin",
"Mojtaba",
""
],
[
"Zahedi",
"Mansooreh",
""
],
[
"Khalajzadeh",
"Hourieh",
""
],
[
"Nasab",
"Ali Rezaei",
""
]
] | TITLE: A Study of Gender Discussions in Mobile Apps
ABSTRACT: Mobile software apps ("apps") are one of the prevailing digital technologies
that our modern life heavily depends on. A key issue in the development of apps
is how to design gender-inclusive apps. Apps that do not consider gender
inclusion, diversity, and equality in their design can create barriers (e.g.,
excluding some of the users because of their gender) for their diverse users.
While there have been some efforts to develop gender-inclusive apps, a lack of
deep understanding regarding user perspectives on gender may prevent app
developers and owners from identifying issues related to gender and proposing
solutions for improvement. Users express many different opinions about apps in
their reviews, from sharing their experiences, and reporting bugs, to
requesting new features. In this study, we aim at unpacking gender discussions
about apps from the user perspective by analysing app reviews. We first develop
and evaluate several Machine Learning (ML) and Deep Learning (DL) classifiers
that automatically detect gender reviews (i.e., reviews that contain
discussions about gender). We apply our ML and DL classifiers on a manually
constructed dataset of 1,440 app reviews from the Google App Store, composing
620 gender reviews and 820 non-gender reviews. Our best classifier achieves an
F1-score of 90.77%. Second, our qualitative analysis of a randomly selected 388
out of 620 gender reviews shows that gender discussions in app reviews revolve
around six topics: App Features, Appearance, Content, Company Policy and
Censorship, Advertisement, and Community. Finally, we provide some practical
implications and recommendations for developing gender-inclusive apps.
| new_dataset | 0.968291 |
2303.09825 | Jianhao Jiao | Jianhao Jiao, Feiyi Chen, Hexiang Wei, Jin Wu, Ming Liu | LCE-Calib: Automatic LiDAR-Frame/Event Camera Extrinsic Calibration With
A Globally Optimal Solution | 11 pages, 13 figures, accepted to IEEE/ASME Transactions on
Mechatronics | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The combination of LiDARs and cameras enables a mobile robot to perceive
environments with multi-modal data, becoming a key factor in achieving robust
perception. Traditional frame cameras are sensitive to changing illumination
conditions, motivating us to introduce novel event cameras to make LiDAR-camera
fusion more complete and robust. However, to jointly exploit these sensors, the
challenging extrinsic calibration problem should be addressed. This paper
proposes an automatic checkerboard-based approach to calibrate extrinsics
between a LiDAR and a frame/event camera, where four contributions are
presented. Firstly, we present an automatic feature extraction and checkerboard
tracking method from LiDAR's point clouds. Secondly, we reconstruct realistic
frame images from event streams, applying traditional corner detectors to event
cameras. Thirdly, we propose an initialization-refinement procedure to estimate
extrinsics using point-to-plane and point-to-line constraints in a
coarse-to-fine manner. Fourthly, we introduce a unified and globally optimal
solution to address two optimization problems in calibration. Our approach has
been validated with extensive experiments on 19 simulated and real-world
datasets and outperforms the state-of-the-art.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 08:07:56 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Jiao",
"Jianhao",
""
],
[
"Chen",
"Feiyi",
""
],
[
"Wei",
"Hexiang",
""
],
[
"Wu",
"Jin",
""
],
[
"Liu",
"Ming",
""
]
] | TITLE: LCE-Calib: Automatic LiDAR-Frame/Event Camera Extrinsic Calibration With
A Globally Optimal Solution
ABSTRACT: The combination of LiDARs and cameras enables a mobile robot to perceive
environments with multi-modal data, becoming a key factor in achieving robust
perception. Traditional frame cameras are sensitive to changing illumination
conditions, motivating us to introduce novel event cameras to make LiDAR-camera
fusion more complete and robust. However, to jointly exploit these sensors, the
challenging extrinsic calibration problem should be addressed. This paper
proposes an automatic checkerboard-based approach to calibrate extrinsics
between a LiDAR and a frame/event camera, where four contributions are
presented. Firstly, we present an automatic feature extraction and checkerboard
tracking method from LiDAR's point clouds. Secondly, we reconstruct realistic
frame images from event streams, applying traditional corner detectors to event
cameras. Thirdly, we propose an initialization-refinement procedure to estimate
extrinsics using point-to-plane and point-to-line constraints in a
coarse-to-fine manner. Fourthly, we introduce a unified and globally optimal
solution to address two optimization problems in calibration. Our approach has
been validated with extensive experiments on 19 simulated and real-world
datasets and outperforms the state-of-the-art.
| no_new_dataset | 0.948202 |
2303.09827 | Jihyun Lee | Jihyun Lee, Seungyeon Seo, Yunsu Kim, Gary Geunbae Lee | DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through
Dependency Parsing | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present our work on Track 2 in the Dialog System Technology Challenges 11
(DSTC11). DSTC11-Track2 aims to provide a benchmark for zero-shot,
cross-domain, intent-set induction. In the absence of in-domain training
dataset, robust utterance representation that can be used across domains is
necessary to induce users' intentions. To achieve this, we leveraged a
multi-domain dialogue dataset to fine-tune the language model and proposed
extracting Verb-Object pairs to remove the artifacts of unnecessary
information. Furthermore, we devised the method that generates each cluster's
name for the explainability of clustered results. Our approach achieved 3rd
place in the precision score and showed superior accuracy and normalized mutual
information (NMI) score than the baseline model on various domain datasets.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 08:12:36 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Lee",
"Jihyun",
""
],
[
"Seo",
"Seungyeon",
""
],
[
"Kim",
"Yunsu",
""
],
[
"Lee",
"Gary Geunbae",
""
]
] | TITLE: DORIC : Domain Robust Fine-Tuning for Open Intent Clustering through
Dependency Parsing
ABSTRACT: We present our work on Track 2 in the Dialog System Technology Challenges 11
(DSTC11). DSTC11-Track2 aims to provide a benchmark for zero-shot,
cross-domain, intent-set induction. In the absence of in-domain training
dataset, robust utterance representation that can be used across domains is
necessary to induce users' intentions. To achieve this, we leveraged a
multi-domain dialogue dataset to fine-tune the language model and proposed
extracting Verb-Object pairs to remove the artifacts of unnecessary
information. Furthermore, we devised the method that generates each cluster's
name for the explainability of clustered results. Our approach achieved 3rd
place in the precision score and showed superior accuracy and normalized mutual
information (NMI) score than the baseline model on various domain datasets.
| no_new_dataset | 0.949435 |
2303.09831 | Yuhe Ding | Yuhe Ding, Jian Liang, Jie Cao, Aihua Zheng, Ran He | MODIFY: Model-driven Face Stylization without Style Images | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing face stylization methods always acquire the presence of the target
(style) domain during the translation process, which violates privacy
regulations and limits their applicability in real-world systems. To address
this issue, we propose a new method called MODel-drIven Face stYlization
(MODIFY), which relies on the generative model to bypass the dependence of the
target images. Briefly, MODIFY first trains a generative model in the target
domain and then translates a source input to the target domain via the provided
style model. To preserve the multimodal style information, MODIFY further
introduces an additional remapping network, mapping a known continuous
distribution into the encoder's embedding space. During translation in the
source domain, MODIFY fine-tunes the encoder module within the target
style-persevering model to capture the content of the source input as precisely
as possible. Our method is extremely simple and satisfies versatile training
modes for face stylization. Experimental results on several different datasets
validate the effectiveness of MODIFY for unsupervised face stylization.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 08:35:17 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Ding",
"Yuhe",
""
],
[
"Liang",
"Jian",
""
],
[
"Cao",
"Jie",
""
],
[
"Zheng",
"Aihua",
""
],
[
"He",
"Ran",
""
]
] | TITLE: MODIFY: Model-driven Face Stylization without Style Images
ABSTRACT: Existing face stylization methods always acquire the presence of the target
(style) domain during the translation process, which violates privacy
regulations and limits their applicability in real-world systems. To address
this issue, we propose a new method called MODel-drIven Face stYlization
(MODIFY), which relies on the generative model to bypass the dependence of the
target images. Briefly, MODIFY first trains a generative model in the target
domain and then translates a source input to the target domain via the provided
style model. To preserve the multimodal style information, MODIFY further
introduces an additional remapping network, mapping a known continuous
distribution into the encoder's embedding space. During translation in the
source domain, MODIFY fine-tunes the encoder module within the target
style-persevering model to capture the content of the source input as precisely
as possible. Our method is extremely simple and satisfies versatile training
modes for face stylization. Experimental results on several different datasets
validate the effectiveness of MODIFY for unsupervised face stylization.
| no_new_dataset | 0.95297 |
2303.09849 | Zhengbo Wang | Zhengbo Wang, Jian Liang, Zilei Wang, Tieniu Tan | Exploiting Semantic Attributes for Transductive Zero-Shot Learning | ICASSP 2023 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Zero-shot learning (ZSL) aims to recognize unseen classes by generalizing the
relation between visual features and semantic attributes learned from the seen
classes. A recent paradigm called transductive zero-shot learning further
leverages unlabeled unseen data during training and has obtained impressive
results. These methods always synthesize unseen features from attributes
through a generative adversarial network to mitigate the bias towards seen
classes. However, they neglect the semantic information in the unlabeled unseen
data and thus fail to generate high-fidelity attribute-consistent unseen
features. To address this issue, we present a novel transductive ZSL method
that produces semantic attributes of the unseen data and imposes them on the
generative process. In particular, we first train an attribute decoder that
learns the mapping from visual features to semantic attributes. Then, from the
attribute decoder, we obtain pseudo-attributes of unlabeled data and integrate
them into the generative model, which helps capture the detailed differences
within unseen classes so as to synthesize more discriminative features.
Experiments on five standard benchmarks show that our method yields
state-of-the-art results for zero-shot learning.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 09:09:48 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Wang",
"Zhengbo",
""
],
[
"Liang",
"Jian",
""
],
[
"Wang",
"Zilei",
""
],
[
"Tan",
"Tieniu",
""
]
] | TITLE: Exploiting Semantic Attributes for Transductive Zero-Shot Learning
ABSTRACT: Zero-shot learning (ZSL) aims to recognize unseen classes by generalizing the
relation between visual features and semantic attributes learned from the seen
classes. A recent paradigm called transductive zero-shot learning further
leverages unlabeled unseen data during training and has obtained impressive
results. These methods always synthesize unseen features from attributes
through a generative adversarial network to mitigate the bias towards seen
classes. However, they neglect the semantic information in the unlabeled unseen
data and thus fail to generate high-fidelity attribute-consistent unseen
features. To address this issue, we present a novel transductive ZSL method
that produces semantic attributes of the unseen data and imposes them on the
generative process. In particular, we first train an attribute decoder that
learns the mapping from visual features to semantic attributes. Then, from the
attribute decoder, we obtain pseudo-attributes of unlabeled data and integrate
them into the generative model, which helps capture the detailed differences
within unseen classes so as to synthesize more discriminative features.
Experiments on five standard benchmarks show that our method yields
state-of-the-art results for zero-shot learning.
| no_new_dataset | 0.945248 |
2303.09857 | Jungin Park | Jungin Park, Jiyoung Lee, Kwanghoon Sohn | Dual-path Adaptation from Image to Video Transformers | CVPR 2023. Code is available at
https://github.com/park-jungin/DualPath | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | In this paper, we efficiently transfer the surpassing representation power of
the vision foundation models, such as ViT and Swin, for video understanding
with only a few trainable parameters. Previous adaptation methods have
simultaneously considered spatial and temporal modeling with a unified
learnable module but still suffered from fully leveraging the representative
capabilities of image transformers. We argue that the popular dual-path
(two-stream) architecture in video models can mitigate this problem. We propose
a novel DualPath adaptation separated into spatial and temporal adaptation
paths, where a lightweight bottleneck adapter is employed in each transformer
block. Especially for temporal dynamic modeling, we incorporate consecutive
frames into a grid-like frameset to precisely imitate vision transformers'
capability that extrapolates relationships between tokens. In addition, we
extensively investigate the multiple baselines from a unified perspective in
video understanding and compare them with DualPath. Experimental results on
four action recognition benchmarks prove that pretrained image transformers
with DualPath can be effectively generalized beyond the data domain.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 09:37:07 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Park",
"Jungin",
""
],
[
"Lee",
"Jiyoung",
""
],
[
"Sohn",
"Kwanghoon",
""
]
] | TITLE: Dual-path Adaptation from Image to Video Transformers
ABSTRACT: In this paper, we efficiently transfer the surpassing representation power of
the vision foundation models, such as ViT and Swin, for video understanding
with only a few trainable parameters. Previous adaptation methods have
simultaneously considered spatial and temporal modeling with a unified
learnable module but still suffered from fully leveraging the representative
capabilities of image transformers. We argue that the popular dual-path
(two-stream) architecture in video models can mitigate this problem. We propose
a novel DualPath adaptation separated into spatial and temporal adaptation
paths, where a lightweight bottleneck adapter is employed in each transformer
block. Especially for temporal dynamic modeling, we incorporate consecutive
frames into a grid-like frameset to precisely imitate vision transformers'
capability that extrapolates relationships between tokens. In addition, we
extensively investigate the multiple baselines from a unified perspective in
video understanding and compare them with DualPath. Experimental results on
four action recognition benchmarks prove that pretrained image transformers
with DualPath can be effectively generalized beyond the data domain.
| no_new_dataset | 0.944791 |
2303.09870 | Behzad Bozorgtabar | Devavrat Tomar, Guillaume Vray, Behzad Bozorgtabar, Jean-Philippe
Thiran | TeSLA: Test-Time Self-Learning With Automatic Adversarial Augmentation | CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Most recent test-time adaptation methods focus on only classification tasks,
use specialized network architectures, destroy model calibration or rely on
lightweight information from the source domain. To tackle these issues, this
paper proposes a novel Test-time Self-Learning method with automatic
Adversarial augmentation dubbed TeSLA for adapting a pre-trained source model
to the unlabeled streaming test data. In contrast to conventional self-learning
methods based on cross-entropy, we introduce a new test-time loss function
through an implicitly tight connection with the mutual information and online
knowledge distillation. Furthermore, we propose a learnable efficient
adversarial augmentation module that further enhances online knowledge
distillation by simulating high entropy augmented images. Our method achieves
state-of-the-art classification and segmentation results on several benchmarks
and types of domain shifts, particularly on challenging measurement shifts of
medical images. TeSLA also benefits from several desirable properties compared
to competing methods in terms of calibration, uncertainty metrics,
insensitivity to model architectures, and source training strategies, all
supported by extensive ablations. Our code and models are available on GitHub.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 10:15:13 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Tomar",
"Devavrat",
""
],
[
"Vray",
"Guillaume",
""
],
[
"Bozorgtabar",
"Behzad",
""
],
[
"Thiran",
"Jean-Philippe",
""
]
] | TITLE: TeSLA: Test-Time Self-Learning With Automatic Adversarial Augmentation
ABSTRACT: Most recent test-time adaptation methods focus on only classification tasks,
use specialized network architectures, destroy model calibration or rely on
lightweight information from the source domain. To tackle these issues, this
paper proposes a novel Test-time Self-Learning method with automatic
Adversarial augmentation dubbed TeSLA for adapting a pre-trained source model
to the unlabeled streaming test data. In contrast to conventional self-learning
methods based on cross-entropy, we introduce a new test-time loss function
through an implicitly tight connection with the mutual information and online
knowledge distillation. Furthermore, we propose a learnable efficient
adversarial augmentation module that further enhances online knowledge
distillation by simulating high entropy augmented images. Our method achieves
state-of-the-art classification and segmentation results on several benchmarks
and types of domain shifts, particularly on challenging measurement shifts of
medical images. TeSLA also benefits from several desirable properties compared
to competing methods in terms of calibration, uncertainty metrics,
insensitivity to model architectures, and source training strategies, all
supported by extensive ablations. Our code and models are available on GitHub.
| no_new_dataset | 0.9462 |
2303.09877 | Daniel J. Trosten | Daniel J. Trosten, Sigurd L{\o}kse, Robert Jenssen, Michael C.
Kampffmeyer | On the Effects of Self-supervision and Contrastive Alignment in Deep
Multi-view Clustering | CVPR 2023. Code available at https://github.com/DanielTrosten/DeepMVC | null | null | null | stat.ML cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Self-supervised learning is a central component in recent approaches to deep
multi-view clustering (MVC). However, we find large variations in the
development of self-supervision-based methods for deep MVC, potentially slowing
the progress of the field. To address this, we present DeepMVC, a unified
framework for deep MVC that includes many recent methods as instances. We
leverage our framework to make key observations about the effect of
self-supervision, and in particular, drawbacks of aligning representations with
contrastive learning. Further, we prove that contrastive alignment can
negatively influence cluster separability, and that this effect becomes worse
when the number of views increases. Motivated by our findings, we develop
several new DeepMVC instances with new forms of self-supervision. We conduct
extensive experiments and find that (i) in line with our theoretical findings,
contrastive alignments decreases performance on datasets with many views; (ii)
all methods benefit from some form of self-supervision; and (iii) our new
instances outperform previous methods on several datasets. Based on our
results, we suggest several promising directions for future research. To
enhance the openness of the field, we provide an open-source implementation of
DeepMVC, including recent models and our new instances. Our implementation
includes a consistent evaluation protocol, facilitating fair and accurate
evaluation of methods and components.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 10:51:38 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Trosten",
"Daniel J.",
""
],
[
"Løkse",
"Sigurd",
""
],
[
"Jenssen",
"Robert",
""
],
[
"Kampffmeyer",
"Michael C.",
""
]
] | TITLE: On the Effects of Self-supervision and Contrastive Alignment in Deep
Multi-view Clustering
ABSTRACT: Self-supervised learning is a central component in recent approaches to deep
multi-view clustering (MVC). However, we find large variations in the
development of self-supervision-based methods for deep MVC, potentially slowing
the progress of the field. To address this, we present DeepMVC, a unified
framework for deep MVC that includes many recent methods as instances. We
leverage our framework to make key observations about the effect of
self-supervision, and in particular, drawbacks of aligning representations with
contrastive learning. Further, we prove that contrastive alignment can
negatively influence cluster separability, and that this effect becomes worse
when the number of views increases. Motivated by our findings, we develop
several new DeepMVC instances with new forms of self-supervision. We conduct
extensive experiments and find that (i) in line with our theoretical findings,
contrastive alignments decreases performance on datasets with many views; (ii)
all methods benefit from some form of self-supervision; and (iii) our new
instances outperform previous methods on several datasets. Based on our
results, we suggest several promising directions for future research. To
enhance the openness of the field, we provide an open-source implementation of
DeepMVC, including recent models and our new instances. Our implementation
includes a consistent evaluation protocol, facilitating fair and accurate
evaluation of methods and components.
| no_new_dataset | 0.945851 |
2303.09905 | Alexandru Coca | A. Coca, B.H. Tseng, W. Lin, B. Byrne | More Robust Schema-Guided Dialogue State Tracking via Tree-Based
Paraphrase Ranking | Accepted at EACL (Findings) 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The schema-guided paradigm overcomes scalability issues inherent in building
task-oriented dialogue (TOD) agents with static ontologies. Instead of
operating on dialogue context alone, agents have access to hierarchical schemas
containing task-relevant natural language descriptions. Fine-tuned language
models excel at schema-guided dialogue state tracking (DST) but are sensitive
to the writing style of the schemas. We explore methods for improving the
robustness of DST models. We propose a framework for generating synthetic
schemas which uses tree-based ranking to jointly optimise lexical diversity and
semantic faithfulness. The generalisation of strong baselines is improved when
augmenting their training data with prompts generated by our framework, as
demonstrated by marked improvements in average joint goal accuracy (JGA) and
schema sensitivity (SS) on the SGD-X benchmark.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 11:43:08 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Coca",
"A.",
""
],
[
"Tseng",
"B. H.",
""
],
[
"Lin",
"W.",
""
],
[
"Byrne",
"B.",
""
]
] | TITLE: More Robust Schema-Guided Dialogue State Tracking via Tree-Based
Paraphrase Ranking
ABSTRACT: The schema-guided paradigm overcomes scalability issues inherent in building
task-oriented dialogue (TOD) agents with static ontologies. Instead of
operating on dialogue context alone, agents have access to hierarchical schemas
containing task-relevant natural language descriptions. Fine-tuned language
models excel at schema-guided dialogue state tracking (DST) but are sensitive
to the writing style of the schemas. We explore methods for improving the
robustness of DST models. We propose a framework for generating synthetic
schemas which uses tree-based ranking to jointly optimise lexical diversity and
semantic faithfulness. The generalisation of strong baselines is improved when
augmenting their training data with prompts generated by our framework, as
demonstrated by marked improvements in average joint goal accuracy (JGA) and
schema sensitivity (SS) on the SGD-X benchmark.
| no_new_dataset | 0.948202 |
2303.09913 | Xugui Zhou | Chloe Smith, Maxfield Kouzel, Xugui Zhou, Homa Alemzadeh | Short: Basal-Adjust: Trend Prediction Alerts and Adjusted Basal Rates
for Hyperglycemia Prevention | 5 pages, 4 figures, 4 tables, to appear in the IEEE/ACM international
conference on Connected Health: Applications, Systems and Engineering
Technologies (CHASE), 2023 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Significant advancements in type 1 diabetes treatment have been made in the
development of state-of-the-art Artificial Pancreas Systems (APS). However,
lapses currently exist in the timely treatment of unsafe blood glucose (BG)
levels, especially in the case of rebound hyperglycemia. We propose a machine
learning (ML) method for predictive BG scenario categorization that outputs
messages alerting the patient to upcoming BG trends to allow for earlier,
educated treatment. In addition to standard notifications of predicted
hypoglycemia and hyperglycemia, we introduce BG scenario-specific alert
messages and the preliminary steps toward precise basal suggestions for the
prevention of rebound hyperglycemia. Experimental evaluation on the DCLP3
clinical dataset achieves >98% accuracy and >79% precision for predicting
rebound high events for patient alerts.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 17:04:05 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Smith",
"Chloe",
""
],
[
"Kouzel",
"Maxfield",
""
],
[
"Zhou",
"Xugui",
""
],
[
"Alemzadeh",
"Homa",
""
]
] | TITLE: Short: Basal-Adjust: Trend Prediction Alerts and Adjusted Basal Rates
for Hyperglycemia Prevention
ABSTRACT: Significant advancements in type 1 diabetes treatment have been made in the
development of state-of-the-art Artificial Pancreas Systems (APS). However,
lapses currently exist in the timely treatment of unsafe blood glucose (BG)
levels, especially in the case of rebound hyperglycemia. We propose a machine
learning (ML) method for predictive BG scenario categorization that outputs
messages alerting the patient to upcoming BG trends to allow for earlier,
educated treatment. In addition to standard notifications of predicted
hypoglycemia and hyperglycemia, we introduce BG scenario-specific alert
messages and the preliminary steps toward precise basal suggestions for the
prevention of rebound hyperglycemia. Experimental evaluation on the DCLP3
clinical dataset achieves >98% accuracy and >79% precision for predicting
rebound high events for patient alerts.
| no_new_dataset | 0.943971 |
2303.09916 | Hong Zhang | YuPeng Huang, Hong Zhang, Siyuan Jiang, Dajiong Yue, Xiaohan Lin, Jun
Zhang, Yi Qin Gao | DSDP: A Blind Docking Strategy Accelerated by GPUs | null | null | null | null | physics.chem-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Virtual screening, including molecular docking, plays an essential role in
drug discovery. Many traditional and machine-learning based methods are
available to fulfil the docking task. The traditional docking methods are
normally extensively time-consuming, and their performance in blind docking
remains to be improved. Although the runtime of docking based on machine
learning is significantly decreased, their accuracy is still limited. In this
study, we take the advantage of both traditional and machine-learning based
methods, and present a method Deep Site and Docking Pose (DSDP) to improve the
performance of blind docking. For the traditional blind docking, the entire
protein is covered by a cube, and the initial positions of ligands are randomly
generated in the cube. In contract, DSDP can predict the binding site of
proteins and provide an accurate searching space and initial positions for the
further conformational sampling. The docking task of DSDP makes use of the
score function and a similar but modified searching strategy of AutoDock Vina,
accelerated by implementation in GPUs. We systematically compare its
performance with the state-of-the-art methods, including Autodock Vina, GNINA,
QuickVina, SMINA, and DiffDock. DSDP reaches a 29.8% top-1 success rate (RMSD <
2 {\AA}) on an unbiased and challenging test dataset with 1.2 s wall-clock
computational time per system. Its performances on DUD-E dataset and the
time-split PDBBind dataset used in EquiBind, TankBind, and DiffDock are also
effective, presenting a 57.2% and 41.8% top-1 success rate with 0.8 s and 1.0 s
per system, respectively.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 07:00:21 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Huang",
"YuPeng",
""
],
[
"Zhang",
"Hong",
""
],
[
"Jiang",
"Siyuan",
""
],
[
"Yue",
"Dajiong",
""
],
[
"Lin",
"Xiaohan",
""
],
[
"Zhang",
"Jun",
""
],
[
"Gao",
"Yi Qin",
""
]
] | TITLE: DSDP: A Blind Docking Strategy Accelerated by GPUs
ABSTRACT: Virtual screening, including molecular docking, plays an essential role in
drug discovery. Many traditional and machine-learning based methods are
available to fulfil the docking task. The traditional docking methods are
normally extensively time-consuming, and their performance in blind docking
remains to be improved. Although the runtime of docking based on machine
learning is significantly decreased, their accuracy is still limited. In this
study, we take the advantage of both traditional and machine-learning based
methods, and present a method Deep Site and Docking Pose (DSDP) to improve the
performance of blind docking. For the traditional blind docking, the entire
protein is covered by a cube, and the initial positions of ligands are randomly
generated in the cube. In contract, DSDP can predict the binding site of
proteins and provide an accurate searching space and initial positions for the
further conformational sampling. The docking task of DSDP makes use of the
score function and a similar but modified searching strategy of AutoDock Vina,
accelerated by implementation in GPUs. We systematically compare its
performance with the state-of-the-art methods, including Autodock Vina, GNINA,
QuickVina, SMINA, and DiffDock. DSDP reaches a 29.8% top-1 success rate (RMSD <
2 {\AA}) on an unbiased and challenging test dataset with 1.2 s wall-clock
computational time per system. Its performances on DUD-E dataset and the
time-split PDBBind dataset used in EquiBind, TankBind, and DiffDock are also
effective, presenting a 57.2% and 41.8% top-1 success rate with 0.8 s and 1.0 s
per system, respectively.
| no_new_dataset | 0.9455 |
2303.09919 | Dongsheng Wang | Dongsheng Wang, Xu Jia, Yang Zhang, Xinyu Zhang, Yaoyuan Wang, Ziyang
Zhang, Dong Wang, Huchuan Lu | Dual Memory Aggregation Network for Event-Based Object Detection with
Learnable Representation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Event-based cameras are bio-inspired sensors that capture brightness change
of every pixel in an asynchronous manner. Compared with frame-based sensors,
event cameras have microsecond-level latency and high dynamic range, hence
showing great potential for object detection under high-speed motion and poor
illumination conditions. Due to sparsity and asynchronism nature with event
streams, most of existing approaches resort to hand-crafted methods to convert
event data into 2D grid representation. However, they are sub-optimal in
aggregating information from event stream for object detection. In this work,
we propose to learn an event representation optimized for event-based object
detection. Specifically, event streams are divided into grids in the x-y-t
coordinates for both positive and negative polarity, producing a set of pillars
as 3D tensor representation. To fully exploit information with event streams to
detect objects, a dual-memory aggregation network (DMANet) is proposed to
leverage both long and short memory along event streams to aggregate effective
information for object detection. Long memory is encoded in the hidden state of
adaptive convLSTMs while short memory is modeled by computing spatial-temporal
correlation between event pillars at neighboring time intervals. Extensive
experiments on the recently released event-based automotive detection dataset
demonstrate the effectiveness of the proposed method.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 12:12:41 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Wang",
"Dongsheng",
""
],
[
"Jia",
"Xu",
""
],
[
"Zhang",
"Yang",
""
],
[
"Zhang",
"Xinyu",
""
],
[
"Wang",
"Yaoyuan",
""
],
[
"Zhang",
"Ziyang",
""
],
[
"Wang",
"Dong",
""
],
[
"Lu",
"Huchuan",
""
]
] | TITLE: Dual Memory Aggregation Network for Event-Based Object Detection with
Learnable Representation
ABSTRACT: Event-based cameras are bio-inspired sensors that capture brightness change
of every pixel in an asynchronous manner. Compared with frame-based sensors,
event cameras have microsecond-level latency and high dynamic range, hence
showing great potential for object detection under high-speed motion and poor
illumination conditions. Due to sparsity and asynchronism nature with event
streams, most of existing approaches resort to hand-crafted methods to convert
event data into 2D grid representation. However, they are sub-optimal in
aggregating information from event stream for object detection. In this work,
we propose to learn an event representation optimized for event-based object
detection. Specifically, event streams are divided into grids in the x-y-t
coordinates for both positive and negative polarity, producing a set of pillars
as 3D tensor representation. To fully exploit information with event streams to
detect objects, a dual-memory aggregation network (DMANet) is proposed to
leverage both long and short memory along event streams to aggregate effective
information for object detection. Long memory is encoded in the hidden state of
adaptive convLSTMs while short memory is modeled by computing spatial-temporal
correlation between event pillars at neighboring time intervals. Extensive
experiments on the recently released event-based automotive detection dataset
demonstrate the effectiveness of the proposed method.
| no_new_dataset | 0.894835 |
2303.09930 | NIkhil Cherian Kurian | Nikhil Cherian Kurian, Varsha S, Abhijit Patil, Shashikant Khade, Amit
Sethi | Robust Semi-Supervised Learning for Histopathology Images through
Self-Supervision Guided Out-of-Distribution Scoring | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Semi-supervised learning (semi-SL) is a promising alternative to supervised
learning for medical image analysis when obtaining good quality supervision for
medical imaging is difficult. However, semi-SL assumes that the underlying
distribution of unaudited data matches that of the few labeled samples, which
is often violated in practical settings, particularly in medical images. The
presence of out-of-distribution (OOD) samples in the unlabeled training pool of
semi-SL is inevitable and can reduce the efficiency of the algorithm. Common
preprocessing methods to filter out outlier samples may not be suitable for
medical images that involve a wide range of anatomical structures and rare
morphologies. In this paper, we propose a novel pipeline for addressing
open-set supervised learning challenges in digital histology images. Our
pipeline efficiently estimates an OOD score for each unlabelled data point
based on self-supervised learning to calibrate the knowledge needed for a
subsequent semi-SL framework. The outlier score derived from the OOD detector
is used to modulate sample selection for the subsequent semi-SL stage, ensuring
that samples conforming to the distribution of the few labeled samples are more
frequently exposed to the subsequent semi-SL framework. Our framework is
compatible with any semi-SL framework, and we base our experiments on the
popular Mixmatch semi-SL framework. We conduct extensive studies on two digital
pathology datasets, Kather colorectal histology dataset and a dataset derived
from TCGA-BRCA whole slide images, and establish the effectiveness of our
method by comparing with popular methods and frameworks in semi-SL algorithms
through various experiments.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 12:38:28 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Kurian",
"Nikhil Cherian",
""
],
[
"S",
"Varsha",
""
],
[
"Patil",
"Abhijit",
""
],
[
"Khade",
"Shashikant",
""
],
[
"Sethi",
"Amit",
""
]
] | TITLE: Robust Semi-Supervised Learning for Histopathology Images through
Self-Supervision Guided Out-of-Distribution Scoring
ABSTRACT: Semi-supervised learning (semi-SL) is a promising alternative to supervised
learning for medical image analysis when obtaining good quality supervision for
medical imaging is difficult. However, semi-SL assumes that the underlying
distribution of unaudited data matches that of the few labeled samples, which
is often violated in practical settings, particularly in medical images. The
presence of out-of-distribution (OOD) samples in the unlabeled training pool of
semi-SL is inevitable and can reduce the efficiency of the algorithm. Common
preprocessing methods to filter out outlier samples may not be suitable for
medical images that involve a wide range of anatomical structures and rare
morphologies. In this paper, we propose a novel pipeline for addressing
open-set supervised learning challenges in digital histology images. Our
pipeline efficiently estimates an OOD score for each unlabelled data point
based on self-supervised learning to calibrate the knowledge needed for a
subsequent semi-SL framework. The outlier score derived from the OOD detector
is used to modulate sample selection for the subsequent semi-SL stage, ensuring
that samples conforming to the distribution of the few labeled samples are more
frequently exposed to the subsequent semi-SL framework. Our framework is
compatible with any semi-SL framework, and we base our experiments on the
popular Mixmatch semi-SL framework. We conduct extensive studies on two digital
pathology datasets, Kather colorectal histology dataset and a dataset derived
from TCGA-BRCA whole slide images, and establish the effectiveness of our
method by comparing with popular methods and frameworks in semi-SL algorithms
through various experiments.
| no_new_dataset | 0.953579 |
2303.09946 | Wail Gueaieb | Shuzheng Qu, Mohammed Abouheaf, Wail Gueaieb, and Davide Spinello | An Adaptive Fuzzy Reinforcement Learning Cooperative Approach for the
Autonomous Control of Flock Systems | 7 pages, 2 figures | IEEE International Conference on Robotics and Automation (ICRA)
2021 | 10.1109/ICRA48506.2021.9561204 | null | eess.SY cs.LG cs.MA cs.RO cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The flock-guidance problem enjoys a challenging structure where multiple
optimization objectives are solved simultaneously. This usually necessitates
different control approaches to tackle various objectives, such as guidance,
collision avoidance, and cohesion. The guidance schemes, in particular, have
long suffered from complex tracking-error dynamics. Furthermore, techniques
that are based on linear feedback strategies obtained at equilibrium conditions
either may not hold or degrade when applied to uncertain dynamic environments.
Pre-tuned fuzzy inference architectures lack robustness under such unmodeled
conditions. This work introduces an adaptive distributed technique for the
autonomous control of flock systems. Its relatively flexible structure is based
on online fuzzy reinforcement learning schemes which simultaneously target a
number of objectives; namely, following a leader, avoiding collision, and
reaching a flock velocity consensus. In addition to its resilience in the face
of dynamic disturbances, the algorithm does not require more than the agent
position as a feedback signal. The effectiveness of the proposed method is
validated with two simulation scenarios and benchmarked against a similar
technique from the literature.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:07:35 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Qu",
"Shuzheng",
""
],
[
"Abouheaf",
"Mohammed",
""
],
[
"Gueaieb",
"Wail",
""
],
[
"Spinello",
"Davide",
""
]
] | TITLE: An Adaptive Fuzzy Reinforcement Learning Cooperative Approach for the
Autonomous Control of Flock Systems
ABSTRACT: The flock-guidance problem enjoys a challenging structure where multiple
optimization objectives are solved simultaneously. This usually necessitates
different control approaches to tackle various objectives, such as guidance,
collision avoidance, and cohesion. The guidance schemes, in particular, have
long suffered from complex tracking-error dynamics. Furthermore, techniques
that are based on linear feedback strategies obtained at equilibrium conditions
either may not hold or degrade when applied to uncertain dynamic environments.
Pre-tuned fuzzy inference architectures lack robustness under such unmodeled
conditions. This work introduces an adaptive distributed technique for the
autonomous control of flock systems. Its relatively flexible structure is based
on online fuzzy reinforcement learning schemes which simultaneously target a
number of objectives; namely, following a leader, avoiding collision, and
reaching a flock velocity consensus. In addition to its resilience in the face
of dynamic disturbances, the algorithm does not require more than the agent
position as a feedback signal. The effectiveness of the proposed method is
validated with two simulation scenarios and benchmarked against a similar
technique from the literature.
| no_new_dataset | 0.942981 |
2303.09950 | Zheng Qin | Zheng Qin, Hao Yu, Changjian Wang, Yuxing Peng and Kai Xu | Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud
Registration | Accepted by CVPR 2023. Our code and models are available at
https://github.com/qinzheng93/GraphSCNet | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of outlier correspondence pruning for non-rigid point
cloud registration. In rigid registration, spatial consistency has been a
commonly used criterion to discriminate outliers from inliers. It measures the
compatibility of two correspondences by the discrepancy between the respective
distances in two point clouds. However, spatial consistency no longer holds in
non-rigid cases and outlier rejection for non-rigid registration has not been
well studied. In this work, we propose Graph-based Spatial Consistency Network
(GraphSCNet) to filter outliers for non-rigid registration. Our method is based
on the fact that non-rigid deformations are usually locally rigid, or local
shape preserving. We first design a local spatial consistency measure over the
deformation graph of the point cloud, which evaluates the spatial compatibility
only between the correspondences in the vicinity of a graph node. An
attention-based non-rigid correspondence embedding module is then devised to
learn a robust representation of non-rigid correspondences from local spatial
consistency. Despite its simplicity, GraphSCNet effectively improves the
quality of the putative correspondences and attains state-of-the-art
performance on three challenging benchmarks. Our code and models are available
at https://github.com/qinzheng93/GraphSCNet.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:15:03 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Qin",
"Zheng",
""
],
[
"Yu",
"Hao",
""
],
[
"Wang",
"Changjian",
""
],
[
"Peng",
"Yuxing",
""
],
[
"Xu",
"Kai",
""
]
] | TITLE: Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud
Registration
ABSTRACT: We study the problem of outlier correspondence pruning for non-rigid point
cloud registration. In rigid registration, spatial consistency has been a
commonly used criterion to discriminate outliers from inliers. It measures the
compatibility of two correspondences by the discrepancy between the respective
distances in two point clouds. However, spatial consistency no longer holds in
non-rigid cases and outlier rejection for non-rigid registration has not been
well studied. In this work, we propose Graph-based Spatial Consistency Network
(GraphSCNet) to filter outliers for non-rigid registration. Our method is based
on the fact that non-rigid deformations are usually locally rigid, or local
shape preserving. We first design a local spatial consistency measure over the
deformation graph of the point cloud, which evaluates the spatial compatibility
only between the correspondences in the vicinity of a graph node. An
attention-based non-rigid correspondence embedding module is then devised to
learn a robust representation of non-rigid correspondences from local spatial
consistency. Despite its simplicity, GraphSCNet effectively improves the
quality of the putative correspondences and attains state-of-the-art
performance on three challenging benchmarks. Our code and models are available
at https://github.com/qinzheng93/GraphSCNet.
| no_new_dataset | 0.950641 |
2303.09956 | Yangfan Zhou | Yang-Fan Zhou, Kai-Lang Yao, Wu-Jun Li | GNNFormer: A Graph-based Framework for Cytopathology Report Generation | 12 pages, 6 figures | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cytopathology report generation is a necessary step for the standardized
examination of pathology images. However, manually writing detailed reports
brings heavy workloads for pathologists. To improve efficiency, some existing
works have studied automatic generation of cytopathology reports, mainly by
applying image caption generation frameworks with visual encoders originally
proposed for natural images. A common weakness of these works is that they do
not explicitly model the structural information among cells, which is a key
feature of pathology images and provides significant information for making
diagnoses. In this paper, we propose a novel graph-based framework called
GNNFormer, which seamlessly integrates graph neural network (GNN) and
Transformer into the same framework, for cytopathology report generation. To
the best of our knowledge, GNNFormer is the first report generation method that
explicitly models the structural information among cells in pathology images.
It also effectively fuses structural information among cells, fine-grained
morphology features of cells and background features to generate high-quality
reports. Experimental results on the NMI-WSI dataset show that GNNFormer can
outperform other state-of-the-art baselines.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:25:29 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Zhou",
"Yang-Fan",
""
],
[
"Yao",
"Kai-Lang",
""
],
[
"Li",
"Wu-Jun",
""
]
] | TITLE: GNNFormer: A Graph-based Framework for Cytopathology Report Generation
ABSTRACT: Cytopathology report generation is a necessary step for the standardized
examination of pathology images. However, manually writing detailed reports
brings heavy workloads for pathologists. To improve efficiency, some existing
works have studied automatic generation of cytopathology reports, mainly by
applying image caption generation frameworks with visual encoders originally
proposed for natural images. A common weakness of these works is that they do
not explicitly model the structural information among cells, which is a key
feature of pathology images and provides significant information for making
diagnoses. In this paper, we propose a novel graph-based framework called
GNNFormer, which seamlessly integrates graph neural network (GNN) and
Transformer into the same framework, for cytopathology report generation. To
the best of our knowledge, GNNFormer is the first report generation method that
explicitly models the structural information among cells in pathology images.
It also effectively fuses structural information among cells, fine-grained
morphology features of cells and background features to generate high-quality
reports. Experimental results on the NMI-WSI dataset show that GNNFormer can
outperform other state-of-the-art baselines.
| no_new_dataset | 0.948346 |
2303.09957 | Norman Meuschke | Norman Meuschke, Apurva Jagdale, Timo Spinde, Jelena Mitrovi\'c, Bela
Gipp | A Benchmark of PDF Information Extraction Tools using a Multi-Task and
Multi-Domain Evaluation Framework for Academic Documents | iConference 2023 | null | 10.1007/978-3-031-28032-0_31 | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Extracting information from academic PDF documents is crucial for numerous
indexing, retrieval, and analysis use cases. Choosing the best tool to extract
specific content elements is difficult because many, technically diverse tools
are available, but recent performance benchmarks are rare. Moreover, such
benchmarks typically cover only a few content elements like header metadata or
bibliographic references and use smaller datasets from specific academic
disciplines. We provide a large and diverse evaluation framework that supports
more extraction tasks than most related datasets. Our framework builds upon
DocBank, a multi-domain dataset of 1.5M annotated content elements extracted
from 500K pages of research papers on arXiv. Using the new framework, we
benchmark ten freely available tools in extracting document metadata,
bibliographic references, tables, and other content elements from academic PDF
documents. GROBID achieves the best metadata and reference extraction results,
followed by CERMINE and Science Parse. For table extraction, Adobe Extract
outperforms other tools, even though the performance is much lower than for
other content elements. All tools struggle to extract lists, footers, and
equations. We conclude that more research on improving and combining tools is
necessary to achieve satisfactory extraction quality for most content elements.
Evaluation datasets and frameworks like the one we present support this line of
research. We make our data and code publicly available to contribute toward
this goal.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:26:33 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Meuschke",
"Norman",
""
],
[
"Jagdale",
"Apurva",
""
],
[
"Spinde",
"Timo",
""
],
[
"Mitrović",
"Jelena",
""
],
[
"Gipp",
"Bela",
""
]
] | TITLE: A Benchmark of PDF Information Extraction Tools using a Multi-Task and
Multi-Domain Evaluation Framework for Academic Documents
ABSTRACT: Extracting information from academic PDF documents is crucial for numerous
indexing, retrieval, and analysis use cases. Choosing the best tool to extract
specific content elements is difficult because many, technically diverse tools
are available, but recent performance benchmarks are rare. Moreover, such
benchmarks typically cover only a few content elements like header metadata or
bibliographic references and use smaller datasets from specific academic
disciplines. We provide a large and diverse evaluation framework that supports
more extraction tasks than most related datasets. Our framework builds upon
DocBank, a multi-domain dataset of 1.5M annotated content elements extracted
from 500K pages of research papers on arXiv. Using the new framework, we
benchmark ten freely available tools in extracting document metadata,
bibliographic references, tables, and other content elements from academic PDF
documents. GROBID achieves the best metadata and reference extraction results,
followed by CERMINE and Science Parse. For table extraction, Adobe Extract
outperforms other tools, even though the performance is much lower than for
other content elements. All tools struggle to extract lists, footers, and
equations. We conclude that more research on improving and combining tools is
necessary to achieve satisfactory extraction quality for most content elements.
Evaluation datasets and frameworks like the one we present support this line of
research. We make our data and code publicly available to contribute toward
this goal.
| new_dataset | 0.960063 |
2303.09972 | Jiawei Yang | Jiawei Yang, Susanto Rahardja, Pasi Franti | Neighborhood Averaging for Improving Outlier Detectors | null | null | null | null | cs.LG cs.CV cs.DS | http://creativecommons.org/publicdomain/zero/1.0/ | We hypothesize that similar objects should have similar outlier scores. To
our knowledge, all existing outlier detectors calculate the outlier score for
each object independently regardless of the outlier scores of the other
objects. Therefore, they do not guarantee that similar objects have similar
outlier scores. To verify our proposed hypothesis, we propose an outlier score
post-processing technique for outlier detectors, called neighborhood
averaging(NA), which pays attention to objects and their neighbors and
guarantees them to have more similar outlier scores than their original scores.
Given an object and its outlier score from any outlier detector, NA modifies
its outlier score by combining it with its k nearest neighbors' scores. We
demonstrate the effectivity of NA by using the well-known k-nearest neighbors
(k-NN). Experimental results show that NA improves all 10 tested baseline
detectors by 13% (from 0.70 to 0.79 AUC) on average evaluated on nine
real-world datasets. Moreover, even outlier detectors that are already based on
k-NN are also improved. The experiments also show that in some applications,
the choice of detector is no more significant when detectors are jointly used
with NA, which may pose a challenge to the generally considered idea that the
data model is the most important factor. We open our code on www.outlierNet.com
for reproducibility.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:44:52 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Yang",
"Jiawei",
""
],
[
"Rahardja",
"Susanto",
""
],
[
"Franti",
"Pasi",
""
]
] | TITLE: Neighborhood Averaging for Improving Outlier Detectors
ABSTRACT: We hypothesize that similar objects should have similar outlier scores. To
our knowledge, all existing outlier detectors calculate the outlier score for
each object independently regardless of the outlier scores of the other
objects. Therefore, they do not guarantee that similar objects have similar
outlier scores. To verify our proposed hypothesis, we propose an outlier score
post-processing technique for outlier detectors, called neighborhood
averaging(NA), which pays attention to objects and their neighbors and
guarantees them to have more similar outlier scores than their original scores.
Given an object and its outlier score from any outlier detector, NA modifies
its outlier score by combining it with its k nearest neighbors' scores. We
demonstrate the effectivity of NA by using the well-known k-nearest neighbors
(k-NN). Experimental results show that NA improves all 10 tested baseline
detectors by 13% (from 0.70 to 0.79 AUC) on average evaluated on nine
real-world datasets. Moreover, even outlier detectors that are already based on
k-NN are also improved. The experiments also show that in some applications,
the choice of detector is no more significant when detectors are jointly used
with NA, which may pose a challenge to the generally considered idea that the
data model is the most important factor. We open our code on www.outlierNet.com
for reproducibility.
| no_new_dataset | 0.951006 |
2303.09977 | Fengyun Wang | Fengyun Wang, Dong Zhang, Hanwang Zhang, Jinhui Tang, and Qianru Sun | Semantic Scene Completion with Cleaner Self | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Semantic Scene Completion (SSC) transforms an image of single-view depth
and/or RGB 2D pixels into 3D voxels, each of whose semantic labels are
predicted. SSC is a well-known ill-posed problem as the prediction model has to
"imagine" what is behind the visible surface, which is usually represented by
Truncated Signed Distance Function (TSDF). Due to the sensory imperfection of
the depth camera, most existing methods based on the noisy TSDF estimated from
depth values suffer from 1) incomplete volumetric predictions and 2) confused
semantic labels. To this end, we use the ground-truth 3D voxels to generate a
perfect visible surface, called TSDF-CAD, and then train a "cleaner" SSC model.
As the model is noise-free, it is expected to focus more on the "imagination"
of unseen voxels. Then, we propose to distill the intermediate "cleaner"
knowledge into another model with noisy TSDF input. In particular, we use the
3D occupancy feature and the semantic relations of the "cleaner self" to
supervise the counterparts of the "noisy self" to respectively address the
above two incorrect predictions. Experimental results validate that our method
improves the noisy counterparts with 3.1% IoU and 2.2% mIoU for measuring scene
completion and SSC, and also achieves new state-of-the-art accuracy on the
popular NYU dataset.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 13:50:18 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Wang",
"Fengyun",
""
],
[
"Zhang",
"Dong",
""
],
[
"Zhang",
"Hanwang",
""
],
[
"Tang",
"Jinhui",
""
],
[
"Sun",
"Qianru",
""
]
] | TITLE: Semantic Scene Completion with Cleaner Self
ABSTRACT: Semantic Scene Completion (SSC) transforms an image of single-view depth
and/or RGB 2D pixels into 3D voxels, each of whose semantic labels are
predicted. SSC is a well-known ill-posed problem as the prediction model has to
"imagine" what is behind the visible surface, which is usually represented by
Truncated Signed Distance Function (TSDF). Due to the sensory imperfection of
the depth camera, most existing methods based on the noisy TSDF estimated from
depth values suffer from 1) incomplete volumetric predictions and 2) confused
semantic labels. To this end, we use the ground-truth 3D voxels to generate a
perfect visible surface, called TSDF-CAD, and then train a "cleaner" SSC model.
As the model is noise-free, it is expected to focus more on the "imagination"
of unseen voxels. Then, we propose to distill the intermediate "cleaner"
knowledge into another model with noisy TSDF input. In particular, we use the
3D occupancy feature and the semantic relations of the "cleaner self" to
supervise the counterparts of the "noisy self" to respectively address the
above two incorrect predictions. Experimental results validate that our method
improves the noisy counterparts with 3.1% IoU and 2.2% mIoU for measuring scene
completion and SSC, and also achieves new state-of-the-art accuracy on the
popular NYU dataset.
| no_new_dataset | 0.948537 |
2303.09988 | Yuanqi Chang | Jiawei Mao, Yuanqi Chang, Xuesong Yin, Binling Nie | Star-Net: Improving Single Image Desnowing Model With More Efficient
Connection and Diverse Feature Interaction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Compared to other severe weather image restoration tasks, single image
desnowing is a more challenging task. This is mainly due to the diversity and
irregularity of snow shape, which makes it extremely difficult to restore
images in snowy scenes. Moreover, snow particles also have a veiling effect
similar to haze or mist. Although current works can effectively remove snow
particles with various shapes, they also bring distortion to the restored
image. To address these issues, we propose a novel single image desnowing
network called Star-Net. First, we design a Star type Skip Connection (SSC) to
establish information channels for all different scale features, which can deal
with the complex shape of snow particles.Second, we present a Multi-Stage
Interactive Transformer (MIT) as the base module of Star-Net, which is designed
to better understand snow particle shapes and to address image distortion by
explicitly modeling a variety of important image recovery features. Finally, we
propose a Degenerate Filter Module (DFM) to filter the snow particle and snow
fog residual in the SSC on the spatial and channel domains. Extensive
experiments show that our Star-Net achieves state-of-the-art snow removal
performances on three standard snow removal datasets and retains the original
sharpness of the images.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 14:03:49 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Mao",
"Jiawei",
""
],
[
"Chang",
"Yuanqi",
""
],
[
"Yin",
"Xuesong",
""
],
[
"Nie",
"Binling",
""
]
] | TITLE: Star-Net: Improving Single Image Desnowing Model With More Efficient
Connection and Diverse Feature Interaction
ABSTRACT: Compared to other severe weather image restoration tasks, single image
desnowing is a more challenging task. This is mainly due to the diversity and
irregularity of snow shape, which makes it extremely difficult to restore
images in snowy scenes. Moreover, snow particles also have a veiling effect
similar to haze or mist. Although current works can effectively remove snow
particles with various shapes, they also bring distortion to the restored
image. To address these issues, we propose a novel single image desnowing
network called Star-Net. First, we design a Star type Skip Connection (SSC) to
establish information channels for all different scale features, which can deal
with the complex shape of snow particles.Second, we present a Multi-Stage
Interactive Transformer (MIT) as the base module of Star-Net, which is designed
to better understand snow particle shapes and to address image distortion by
explicitly modeling a variety of important image recovery features. Finally, we
propose a Degenerate Filter Module (DFM) to filter the snow particle and snow
fog residual in the SSC on the spatial and channel domains. Extensive
experiments show that our Star-Net achieves state-of-the-art snow removal
performances on three standard snow removal datasets and retains the original
sharpness of the images.
| no_new_dataset | 0.950041 |
2303.10067 | Zeyd Boukhers | Zeyd Boukhers and Nagaraj Bahubali Asundi | Deep Author Name Disambiguation using DBLP Data | Accepted for publication in the International Journal on Digital
Libraries. arXiv admin note: substantial text overlap with arXiv:2207.04772 | null | null | null | cs.DL cs.LG | http://creativecommons.org/licenses/by/4.0/ | In the academic world, the number of scientists grows every year and so does
the number of authors sharing the same names. Consequently, it challenging to
assign newly published papers to their respective authors. Therefore, Author
Name Ambiguity (ANA) is considered a critical open problem in digital
libraries. This paper proposes an Author Name Disambiguation (AND) approach
that links author names to their real-world entities by leveraging their
co-authors and domain of research. To this end, we use data collected from the
DBLP repository that contains more than 5 million bibliographic records
authored by around 2.6 million co-authors. Our approach first groups authors
who share the same last names and same first name initials. The author within
each group is identified by capturing the relation with his/her co-authors and
area of research, represented by the titles of the validated publications of
the corresponding author. To this end, we train a neural network model that
learns from the representations of the co-authors and titles. We validated the
effectiveness of our approach by conducting extensive experiments on a large
dataset.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 15:50:00 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Boukhers",
"Zeyd",
""
],
[
"Asundi",
"Nagaraj Bahubali",
""
]
] | TITLE: Deep Author Name Disambiguation using DBLP Data
ABSTRACT: In the academic world, the number of scientists grows every year and so does
the number of authors sharing the same names. Consequently, it challenging to
assign newly published papers to their respective authors. Therefore, Author
Name Ambiguity (ANA) is considered a critical open problem in digital
libraries. This paper proposes an Author Name Disambiguation (AND) approach
that links author names to their real-world entities by leveraging their
co-authors and domain of research. To this end, we use data collected from the
DBLP repository that contains more than 5 million bibliographic records
authored by around 2.6 million co-authors. Our approach first groups authors
who share the same last names and same first name initials. The author within
each group is identified by capturing the relation with his/her co-authors and
area of research, represented by the titles of the validated publications of
the corresponding author. To this end, we train a neural network model that
learns from the representations of the co-authors and titles. We validated the
effectiveness of our approach by conducting extensive experiments on a large
dataset.
| no_new_dataset | 0.958421 |
2303.10068 | Pengfei Xu | Pengfei Xu, Zhiyong Peng, Liwei Wang | Proactive Rumor Control: When Impression Counts (Full Version) | null | null | null | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | The spread of rumors in online networks threatens public safety and results
in economic losses. To overcome this problem, a lot of work studies the problem
of rumor control which aims at limiting the spread of rumors. However, all
previous work ignores the relationship between the influence block effect and
counts of impressions on the user. In this paper, we study the problem of
minimizing the spread of rumors when impression counts. Given a graph $G(V,E)$,
a rumor set $R \in V$ and a budget $k$, it aims to find a protector set $P \in
V \backslash R$ to minimize the spread of the rumor set $R$ under the budget
$k$. Due to the impression counts, two following challenges of our problem need
to be overcome: (1) our problem is NP-hard; (2) the influence block is
non-submodular, which means a straightforward greedy approach is not
applicable. Hence, we devise a branch-and-bound framework for this problem with
a ($1-1/e-\epsilon$) approximation ratio. To further improve the efficiency, we
speed up our framework with a progressive upper bound estimation method, which
achieves a ($1-1/e-\epsilon - \rho$) approximation ratio. We conduct
experiments on real-world datasets to verify the efficiency, effectiveness, and
scalability of our methods.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 15:51:28 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Xu",
"Pengfei",
""
],
[
"Peng",
"Zhiyong",
""
],
[
"Wang",
"Liwei",
""
]
] | TITLE: Proactive Rumor Control: When Impression Counts (Full Version)
ABSTRACT: The spread of rumors in online networks threatens public safety and results
in economic losses. To overcome this problem, a lot of work studies the problem
of rumor control which aims at limiting the spread of rumors. However, all
previous work ignores the relationship between the influence block effect and
counts of impressions on the user. In this paper, we study the problem of
minimizing the spread of rumors when impression counts. Given a graph $G(V,E)$,
a rumor set $R \in V$ and a budget $k$, it aims to find a protector set $P \in
V \backslash R$ to minimize the spread of the rumor set $R$ under the budget
$k$. Due to the impression counts, two following challenges of our problem need
to be overcome: (1) our problem is NP-hard; (2) the influence block is
non-submodular, which means a straightforward greedy approach is not
applicable. Hence, we devise a branch-and-bound framework for this problem with
a ($1-1/e-\epsilon$) approximation ratio. To further improve the efficiency, we
speed up our framework with a progressive upper bound estimation method, which
achieves a ($1-1/e-\epsilon - \rho$) approximation ratio. We conduct
experiments on real-world datasets to verify the efficiency, effectiveness, and
scalability of our methods.
| no_new_dataset | 0.935582 |
2303.10094 | Alessandro Benedetti | Alessandro Benedetti, Anna Ruggero | Stat-weight: Improving the Estimator of Interleaved Methods Outcomes
with Statistical Hypothesis Testing | This preprint has not undergone peer review (when applicable) or any
post-submission improvements or corrections. The Version of Record of this
contribution is published in Advances in Information Retrieval 45th European
Conference on Information Retrieval, ECIR 2023, Dublin, Ireland, April, 2023,
Proceedings, Part III, and is available online at
https://doi.org/10.1007/978-3-031-28241-6_2 | Advances in Information Retrieval 45th European Conference on
Information Retrieval, ECIR 2023, Dublin, Ireland, 2023, Proceedings, Part
III | 10.1007/978-3-031-28241-6_2 | null | cs.IR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Interleaving is an online evaluation approach for information retrieval
systems that compares the effectiveness of ranking functions in interpreting
the users' implicit feedback. Previous work such as Hofmann et al (2011) has
evaluated the most promising interleaved methods at the time, on uniform
distributions of queries. In the real world, ordinarily, there is an unbalanced
distribution of repeated queries that follows a long-tailed users' search
demand curve. The more a query is executed, by different users (or in different
sessions), the higher the probability of collecting implicit feedback
(interactions/clicks) on the related search results. This paper first aims to
replicate the Team Draft Interleaving accuracy evaluation on uniform query
distributions and then focuses on assessing how this method generalizes to
long-tailed real-world scenarios. The reproducibility work raised interesting
considerations on how the winning ranking function for each query should impact
the overall winner for the entire evaluation. Based on what was observed, we
propose that not all the queries should contribute to the final decision in
equal proportion. As a result of these insights, we designed two variations of
the $\Delta_{AB}$ score winner estimator that assign to each query a credit
based on statistical hypothesis testing. To replicate, reproduce and extend the
original work, we have developed from scratch a system that simulates a search
engine and users' interactions from datasets from the industry. Our experiments
confirm our intuition and show that our methods are promising in terms of
accuracy, sensitivity, and robustness to noise.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 16:15:13 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Benedetti",
"Alessandro",
""
],
[
"Ruggero",
"Anna",
""
]
] | TITLE: Stat-weight: Improving the Estimator of Interleaved Methods Outcomes
with Statistical Hypothesis Testing
ABSTRACT: Interleaving is an online evaluation approach for information retrieval
systems that compares the effectiveness of ranking functions in interpreting
the users' implicit feedback. Previous work such as Hofmann et al (2011) has
evaluated the most promising interleaved methods at the time, on uniform
distributions of queries. In the real world, ordinarily, there is an unbalanced
distribution of repeated queries that follows a long-tailed users' search
demand curve. The more a query is executed, by different users (or in different
sessions), the higher the probability of collecting implicit feedback
(interactions/clicks) on the related search results. This paper first aims to
replicate the Team Draft Interleaving accuracy evaluation on uniform query
distributions and then focuses on assessing how this method generalizes to
long-tailed real-world scenarios. The reproducibility work raised interesting
considerations on how the winning ranking function for each query should impact
the overall winner for the entire evaluation. Based on what was observed, we
propose that not all the queries should contribute to the final decision in
equal proportion. As a result of these insights, we designed two variations of
the $\Delta_{AB}$ score winner estimator that assign to each query a credit
based on statistical hypothesis testing. To replicate, reproduce and extend the
original work, we have developed from scratch a system that simulates a search
engine and users' interactions from datasets from the industry. Our experiments
confirm our intuition and show that our methods are promising in terms of
accuracy, sensitivity, and robustness to noise.
| no_new_dataset | 0.943295 |
2303.10100 | Liulei Li | Liulei Li, Wenguan Wang, Tianfei Zhou, Jianwu Li, Yi Yang | Unified Mask Embedding and Correspondence Learning for Self-Supervised
Video Segmentation | CVPR 2023. code: https://github.com/0liliulei/Mask-VOS | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The objective of this paper is self-supervised learning of video object
segmentation. We develop a unified framework which simultaneously models
cross-frame dense correspondence for locally discriminative feature learning
and embeds object-level context for target-mask decoding. As a result, it is
able to directly learn to perform mask-guided sequential segmentation from
unlabeled videos, in contrast to previous efforts usually relying on an oblique
solution - cheaply "copying" labels according to pixel-wise correlations.
Concretely, our algorithm alternates between i) clustering video pixels for
creating pseudo segmentation labels ex nihilo; and ii) utilizing the pseudo
labels to learn mask encoding and decoding for VOS. Unsupervised correspondence
learning is further incorporated into this self-taught, mask embedding scheme,
so as to ensure the generic nature of the learnt representation and avoid
cluster degeneracy. Our algorithm sets state-of-the-arts on two standard
benchmarks (i.e., DAVIS17 and YouTube-VOS), narrowing the gap between self- and
fully-supervised VOS, in terms of both performance and network architecture
design.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 16:23:36 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Li",
"Liulei",
""
],
[
"Wang",
"Wenguan",
""
],
[
"Zhou",
"Tianfei",
""
],
[
"Li",
"Jianwu",
""
],
[
"Yang",
"Yi",
""
]
] | TITLE: Unified Mask Embedding and Correspondence Learning for Self-Supervised
Video Segmentation
ABSTRACT: The objective of this paper is self-supervised learning of video object
segmentation. We develop a unified framework which simultaneously models
cross-frame dense correspondence for locally discriminative feature learning
and embeds object-level context for target-mask decoding. As a result, it is
able to directly learn to perform mask-guided sequential segmentation from
unlabeled videos, in contrast to previous efforts usually relying on an oblique
solution - cheaply "copying" labels according to pixel-wise correlations.
Concretely, our algorithm alternates between i) clustering video pixels for
creating pseudo segmentation labels ex nihilo; and ii) utilizing the pseudo
labels to learn mask encoding and decoding for VOS. Unsupervised correspondence
learning is further incorporated into this self-taught, mask embedding scheme,
so as to ensure the generic nature of the learnt representation and avoid
cluster degeneracy. Our algorithm sets state-of-the-arts on two standard
benchmarks (i.e., DAVIS17 and YouTube-VOS), narrowing the gap between self- and
fully-supervised VOS, in terms of both performance and network architecture
design.
| no_new_dataset | 0.945701 |
2303.10144 | Nicolai Dorka | Nicolai Dorka, Tim Welschehold, Wolfram Burgard | Dynamic Update-to-Data Ratio: Minimizing World Model Overfitting | ICLR 2023 | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Early stopping based on the validation set performance is a popular approach
to find the right balance between under- and overfitting in the context of
supervised learning. However, in reinforcement learning, even for supervised
sub-problems such as world model learning, early stopping is not applicable as
the dataset is continually evolving. As a solution, we propose a new general
method that dynamically adjusts the update to data (UTD) ratio during training
based on under- and overfitting detection on a small subset of the continuously
collected experience not used for training. We apply our method to DreamerV2, a
state-of-the-art model-based reinforcement learning algorithm, and evaluate it
on the DeepMind Control Suite and the Atari $100$k benchmark. The results
demonstrate that one can better balance under- and overestimation by adjusting
the UTD ratio with our approach compared to the default setting in DreamerV2
and that it is competitive with an extensive hyperparameter search which is not
feasible for many applications. Our method eliminates the need to set the UTD
hyperparameter by hand and even leads to a higher robustness with regard to
other learning-related hyperparameters further reducing the amount of necessary
tuning.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 17:29:02 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Dorka",
"Nicolai",
""
],
[
"Welschehold",
"Tim",
""
],
[
"Burgard",
"Wolfram",
""
]
] | TITLE: Dynamic Update-to-Data Ratio: Minimizing World Model Overfitting
ABSTRACT: Early stopping based on the validation set performance is a popular approach
to find the right balance between under- and overfitting in the context of
supervised learning. However, in reinforcement learning, even for supervised
sub-problems such as world model learning, early stopping is not applicable as
the dataset is continually evolving. As a solution, we propose a new general
method that dynamically adjusts the update to data (UTD) ratio during training
based on under- and overfitting detection on a small subset of the continuously
collected experience not used for training. We apply our method to DreamerV2, a
state-of-the-art model-based reinforcement learning algorithm, and evaluate it
on the DeepMind Control Suite and the Atari $100$k benchmark. The results
demonstrate that one can better balance under- and overestimation by adjusting
the UTD ratio with our approach compared to the default setting in DreamerV2
and that it is competitive with an extensive hyperparameter search which is not
feasible for many applications. Our method eliminates the need to set the UTD
hyperparameter by hand and even leads to a higher robustness with regard to
other learning-related hyperparameters further reducing the amount of necessary
tuning.
| no_new_dataset | 0.9463 |
2303.10145 | Sudarshan Ambasamudram Rajagopalan | Kitty Varghese, Sudarshan Rajagopalan, Mohit Lamba, Kaushik Mitra | Spectrum-inspired Low-light Image Translation for Saliency Detection | Presented at The Indian Conference on Computer Vision, Graphics and
Image Processing (ICVGIP) 2022 | null | 10.1145/3571600.3571634 | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Saliency detection methods are central to several real-world applications
such as robot navigation and satellite imagery. However, the performance of
existing methods deteriorate under low-light conditions because training
datasets mostly comprise of well-lit images. One possible solution is to
collect a new dataset for low-light conditions. This involves pixel-level
annotations, which is not only tedious and time-consuming but also infeasible
if a huge training corpus is required. We propose a technique that performs
classical band-pass filtering in the Fourier space to transform well-lit images
to low-light images and use them as a proxy for real low-light images. Unlike
popular deep learning approaches which require learning thousands of parameters
and enormous amounts of training data, the proposed transformation is fast and
simple and easy to extend to other tasks such as low-light depth estimation.
Our experiments show that the state-of-the-art saliency detection and depth
estimation networks trained on our proxy low-light images perform significantly
better on real low-light images than networks trained using existing
strategies.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 17:30:42 GMT"
}
] | 2023-03-20T00:00:00 | [
[
"Varghese",
"Kitty",
""
],
[
"Rajagopalan",
"Sudarshan",
""
],
[
"Lamba",
"Mohit",
""
],
[
"Mitra",
"Kaushik",
""
]
] | TITLE: Spectrum-inspired Low-light Image Translation for Saliency Detection
ABSTRACT: Saliency detection methods are central to several real-world applications
such as robot navigation and satellite imagery. However, the performance of
existing methods deteriorate under low-light conditions because training
datasets mostly comprise of well-lit images. One possible solution is to
collect a new dataset for low-light conditions. This involves pixel-level
annotations, which is not only tedious and time-consuming but also infeasible
if a huge training corpus is required. We propose a technique that performs
classical band-pass filtering in the Fourier space to transform well-lit images
to low-light images and use them as a proxy for real low-light images. Unlike
popular deep learning approaches which require learning thousands of parameters
and enormous amounts of training data, the proposed transformation is fast and
simple and easy to extend to other tasks such as low-light depth estimation.
Our experiments show that the state-of-the-art saliency detection and depth
estimation networks trained on our proxy low-light images perform significantly
better on real low-light images than networks trained using existing
strategies.
| new_dataset | 0.778818 |
2101.02390 | Junjie Huang | Junjie Huang, Huawei Shen, Liang Hou, Xueqi Cheng | SDGNN: Learning Node Representation for Signed Directed Networks | Accepted and to appear at AAAI2021 | null | null | null | cs.SI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Network embedding is aimed at mapping nodes in a network into low-dimensional
vector representations. Graph Neural Networks (GNNs) have received widespread
attention and lead to state-of-the-art performance in learning node
representations. However, most GNNs only work in unsigned networks, where only
positive links exist. It is not trivial to transfer these models to signed
directed networks, which are widely observed in the real world yet less
studied. In this paper, we first review two fundamental sociological theories
(i.e., status theory and balance theory) and conduct empirical studies on
real-world datasets to analyze the social mechanism in signed directed
networks. Guided by related sociological theories, we propose a novel Signed
Directed Graph Neural Networks model named SDGNN to learn node embeddings for
signed directed networks. The proposed model simultaneously reconstructs link
signs, link directions, and signed directed triangles. We validate our model's
effectiveness on five real-world datasets, which are commonly used as the
benchmark for signed network embedding. Experiments demonstrate the proposed
model outperforms existing models, including feature-based methods, network
embedding methods, and several GNN methods.
| [
{
"version": "v1",
"created": "Thu, 7 Jan 2021 06:15:07 GMT"
},
{
"version": "v2",
"created": "Wed, 3 Mar 2021 02:23:07 GMT"
},
{
"version": "v3",
"created": "Sat, 27 Mar 2021 11:45:02 GMT"
},
{
"version": "v4",
"created": "Thu, 16 Mar 2023 09:01:55 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Huang",
"Junjie",
""
],
[
"Shen",
"Huawei",
""
],
[
"Hou",
"Liang",
""
],
[
"Cheng",
"Xueqi",
""
]
] | TITLE: SDGNN: Learning Node Representation for Signed Directed Networks
ABSTRACT: Network embedding is aimed at mapping nodes in a network into low-dimensional
vector representations. Graph Neural Networks (GNNs) have received widespread
attention and lead to state-of-the-art performance in learning node
representations. However, most GNNs only work in unsigned networks, where only
positive links exist. It is not trivial to transfer these models to signed
directed networks, which are widely observed in the real world yet less
studied. In this paper, we first review two fundamental sociological theories
(i.e., status theory and balance theory) and conduct empirical studies on
real-world datasets to analyze the social mechanism in signed directed
networks. Guided by related sociological theories, we propose a novel Signed
Directed Graph Neural Networks model named SDGNN to learn node embeddings for
signed directed networks. The proposed model simultaneously reconstructs link
signs, link directions, and signed directed triangles. We validate our model's
effectiveness on five real-world datasets, which are commonly used as the
benchmark for signed network embedding. Experiments demonstrate the proposed
model outperforms existing models, including feature-based methods, network
embedding methods, and several GNN methods.
| no_new_dataset | 0.949902 |
2104.06116 | Constantinos Siettos | Gianluca Fabiani, Francesco Calabr\`o, Lucia Russo, Constantinos
Siettos | Numerical Solution and Bifurcation Analysis of Nonlinear Partial
Differential Equations with Extreme Learning Machines | null | Journal of Scientific Computing 89, 44 (2021) | 10.1007/s10915-021-01650-5 | null | math.NA cs.NA math.AP math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address a new numerical scheme based on a class of machine learning
methods, the so-called Extreme Learning Machines with both sigmoidal and
radial-basis functions, for the computation of steady-state solutions and the
construction of (one dimensional) bifurcation diagrams of nonlinear partial
differential equations (PDEs). For our illustrations, we considered two
benchmark problems, namely (a) the one-dimensional viscous Burgers with both
homogeneous (Dirichlet) and non-homogeneous mixed boundary conditions, and, (b)
the one and two-dimensional Liouville-Bratu-Gelfand PDEs with homogeneous
Dirichlet boundary conditions. For the one-dimensional Burgers and Bratu PDEs,
exact analytical solutions are available and used for comparison purposes
against the numerical derived solutions. Furthermore, the numerical efficiency
(in terms of accuracy and size of the grid) of the proposed numerical machine
learning scheme is compared against central finite differences (FD) and
Galerkin weighted-residuals finite-element methods (FEM). We show that the
proposed ELM numerical method outperforms both FD and FEM methods for medium to
large sized grids, while provides equivalent results with the FEM for low to
medium sized grids.
| [
{
"version": "v1",
"created": "Tue, 13 Apr 2021 11:45:39 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Fabiani",
"Gianluca",
""
],
[
"Calabrò",
"Francesco",
""
],
[
"Russo",
"Lucia",
""
],
[
"Siettos",
"Constantinos",
""
]
] | TITLE: Numerical Solution and Bifurcation Analysis of Nonlinear Partial
Differential Equations with Extreme Learning Machines
ABSTRACT: We address a new numerical scheme based on a class of machine learning
methods, the so-called Extreme Learning Machines with both sigmoidal and
radial-basis functions, for the computation of steady-state solutions and the
construction of (one dimensional) bifurcation diagrams of nonlinear partial
differential equations (PDEs). For our illustrations, we considered two
benchmark problems, namely (a) the one-dimensional viscous Burgers with both
homogeneous (Dirichlet) and non-homogeneous mixed boundary conditions, and, (b)
the one and two-dimensional Liouville-Bratu-Gelfand PDEs with homogeneous
Dirichlet boundary conditions. For the one-dimensional Burgers and Bratu PDEs,
exact analytical solutions are available and used for comparison purposes
against the numerical derived solutions. Furthermore, the numerical efficiency
(in terms of accuracy and size of the grid) of the proposed numerical machine
learning scheme is compared against central finite differences (FD) and
Galerkin weighted-residuals finite-element methods (FEM). We show that the
proposed ELM numerical method outperforms both FD and FEM methods for medium to
large sized grids, while provides equivalent results with the FEM for low to
medium sized grids.
| no_new_dataset | 0.950319 |
2106.13638 | Jochen Stiasny | Jochen Stiasny, Georgios S. Misyris, Spyros Chatzivasileiadis | Transient Stability Analysis with Physics-Informed Neural Networks | 9 pages, 8 figures | null | null | null | cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | We explore the possibility to use physics-informed neural networks to
drastically accelerate the solution of ordinary differential-algebraic
equations that govern the power system dynamics. When it comes to transient
stability assessment, the traditionally applied methods either carry a
significant computational burden, require model simplifications, or use overly
conservative surrogate models. Conventional neural networks can circumvent
these limitations but are faced with high demand of high-quality training
datasets, while they ignore the underlying governing equations.
Physics-informed neural networks are different: they incorporate the power
system differential algebraic equations directly into the neural network
training and drastically reduce the need for training data. This paper takes a
deep dive into the performance of physics-informed neural networks for power
system transient stability assessment. Introducing a new neural network
training procedure to facilitate a thorough comparison, we explore how
physics-informed neural networks compare with conventional
differential-algebraic solvers and classical neural networks in terms of
computation time, requirements in data, and prediction accuracy. We illustrate
the findings on the Kundur two-area system, and assess the opportunities and
challenges of physics-informed neural networks to serve as a transient
stability analysis tool, highlighting possible pathways to further develop this
method.
| [
{
"version": "v1",
"created": "Fri, 25 Jun 2021 13:42:09 GMT"
},
{
"version": "v2",
"created": "Mon, 15 Nov 2021 12:23:16 GMT"
},
{
"version": "v3",
"created": "Wed, 15 Mar 2023 22:50:15 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Stiasny",
"Jochen",
""
],
[
"Misyris",
"Georgios S.",
""
],
[
"Chatzivasileiadis",
"Spyros",
""
]
] | TITLE: Transient Stability Analysis with Physics-Informed Neural Networks
ABSTRACT: We explore the possibility to use physics-informed neural networks to
drastically accelerate the solution of ordinary differential-algebraic
equations that govern the power system dynamics. When it comes to transient
stability assessment, the traditionally applied methods either carry a
significant computational burden, require model simplifications, or use overly
conservative surrogate models. Conventional neural networks can circumvent
these limitations but are faced with high demand of high-quality training
datasets, while they ignore the underlying governing equations.
Physics-informed neural networks are different: they incorporate the power
system differential algebraic equations directly into the neural network
training and drastically reduce the need for training data. This paper takes a
deep dive into the performance of physics-informed neural networks for power
system transient stability assessment. Introducing a new neural network
training procedure to facilitate a thorough comparison, we explore how
physics-informed neural networks compare with conventional
differential-algebraic solvers and classical neural networks in terms of
computation time, requirements in data, and prediction accuracy. We illustrate
the findings on the Kundur two-area system, and assess the opportunities and
challenges of physics-informed neural networks to serve as a transient
stability analysis tool, highlighting possible pathways to further develop this
method.
| no_new_dataset | 0.950411 |
2106.15278 | Geeho Kim | Geeho Kim and Junoh Kang and Bohyung Han | Open-Set Representation Learning through Combinatorial Embedding | CVPR 2023 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual recognition tasks are often limited to dealing with a small subset of
classes simply because the labels for the remaining classes are unavailable. We
are interested in identifying novel concepts in a dataset through
representation learning based on both labeled and unlabeled examples, and
extending the horizon of recognition to both known and novel classes. To
address this challenging task, we propose a combinatorial learning approach,
which naturally clusters the examples in unseen classes using the compositional
knowledge given by multiple supervised meta-classifiers on heterogeneous label
spaces. The representations given by the combinatorial embedding are made more
robust by unsupervised pairwise relation learning. The proposed algorithm
discovers novel concepts via a joint optimization for enhancing the
discrimitiveness of unseen classes as well as learning the representations of
known classes generalizable to novel ones. Our extensive experiments
demonstrate remarkable performance gains by the proposed approach on public
datasets for image retrieval and image categorization with novel class
discovery.
| [
{
"version": "v1",
"created": "Tue, 29 Jun 2021 11:51:57 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Dec 2021 13:03:08 GMT"
},
{
"version": "v3",
"created": "Wed, 15 Mar 2023 20:53:33 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Kim",
"Geeho",
""
],
[
"Kang",
"Junoh",
""
],
[
"Han",
"Bohyung",
""
]
] | TITLE: Open-Set Representation Learning through Combinatorial Embedding
ABSTRACT: Visual recognition tasks are often limited to dealing with a small subset of
classes simply because the labels for the remaining classes are unavailable. We
are interested in identifying novel concepts in a dataset through
representation learning based on both labeled and unlabeled examples, and
extending the horizon of recognition to both known and novel classes. To
address this challenging task, we propose a combinatorial learning approach,
which naturally clusters the examples in unseen classes using the compositional
knowledge given by multiple supervised meta-classifiers on heterogeneous label
spaces. The representations given by the combinatorial embedding are made more
robust by unsupervised pairwise relation learning. The proposed algorithm
discovers novel concepts via a joint optimization for enhancing the
discrimitiveness of unseen classes as well as learning the representations of
known classes generalizable to novel ones. Our extensive experiments
demonstrate remarkable performance gains by the proposed approach on public
datasets for image retrieval and image categorization with novel class
discovery.
| no_new_dataset | 0.946892 |
2201.08845 | Qiangeng Xu | Qiangeng Xu and Zexiang Xu and Julien Philip and Sai Bi and Zhixin Shu
and Kalyan Sunkavalli and Ulrich Neumann | Point-NeRF: Point-based Neural Radiance Fields | Accepted to CVPR 2022 (Oral) | In Proceedings of the IEEE/CVF Conference on Computer Vision and
Pattern Recognition (pp. 5438-5448) (2022) | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | Volumetric neural rendering methods like NeRF generate high-quality view
synthesis results but are optimized per-scene leading to prohibitive
reconstruction time. On the other hand, deep multi-view stereo methods can
quickly reconstruct scene geometry via direct network inference. Point-NeRF
combines the advantages of these two approaches by using neural 3D point
clouds, with associated neural features, to model a radiance field. Point-NeRF
can be rendered efficiently by aggregating neural point features near scene
surfaces, in a ray marching-based rendering pipeline. Moreover, Point-NeRF can
be initialized via direct inference of a pre-trained deep network to produce a
neural point cloud; this point cloud can be finetuned to surpass the visual
quality of NeRF with 30X faster training time. Point-NeRF can be combined with
other 3D reconstruction methods and handles the errors and outliers in such
methods via a novel pruning and growing mechanism. The experiments on the DTU,
the NeRF Synthetics , the ScanNet and the Tanks and Temples datasets
demonstrate Point-NeRF can surpass the existing methods and achieve the
state-of-the-art results.
| [
{
"version": "v1",
"created": "Fri, 21 Jan 2022 18:59:20 GMT"
},
{
"version": "v2",
"created": "Fri, 4 Mar 2022 23:55:50 GMT"
},
{
"version": "v3",
"created": "Tue, 8 Mar 2022 09:19:35 GMT"
},
{
"version": "v4",
"created": "Wed, 16 Mar 2022 07:24:49 GMT"
},
{
"version": "v5",
"created": "Fri, 18 Mar 2022 23:32:17 GMT"
},
{
"version": "v6",
"created": "Fri, 2 Dec 2022 09:32:51 GMT"
},
{
"version": "v7",
"created": "Wed, 15 Mar 2023 21:51:08 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Xu",
"Qiangeng",
""
],
[
"Xu",
"Zexiang",
""
],
[
"Philip",
"Julien",
""
],
[
"Bi",
"Sai",
""
],
[
"Shu",
"Zhixin",
""
],
[
"Sunkavalli",
"Kalyan",
""
],
[
"Neumann",
"Ulrich",
""
]
] | TITLE: Point-NeRF: Point-based Neural Radiance Fields
ABSTRACT: Volumetric neural rendering methods like NeRF generate high-quality view
synthesis results but are optimized per-scene leading to prohibitive
reconstruction time. On the other hand, deep multi-view stereo methods can
quickly reconstruct scene geometry via direct network inference. Point-NeRF
combines the advantages of these two approaches by using neural 3D point
clouds, with associated neural features, to model a radiance field. Point-NeRF
can be rendered efficiently by aggregating neural point features near scene
surfaces, in a ray marching-based rendering pipeline. Moreover, Point-NeRF can
be initialized via direct inference of a pre-trained deep network to produce a
neural point cloud; this point cloud can be finetuned to surpass the visual
quality of NeRF with 30X faster training time. Point-NeRF can be combined with
other 3D reconstruction methods and handles the errors and outliers in such
methods via a novel pruning and growing mechanism. The experiments on the DTU,
the NeRF Synthetics , the ScanNet and the Tanks and Temples datasets
demonstrate Point-NeRF can surpass the existing methods and achieve the
state-of-the-art results.
| no_new_dataset | 0.951188 |
2202.00307 | Qiujie Dong | Qiujie Dong, Zixiong Wang, Manyi Li, Junjie Gao, Shuangmin Chen,
Zhenyu Shu, Shiqing Xin, Changhe Tu, Wenping Wang | Laplacian2Mesh: Laplacian-Based Mesh Understanding | Accepted by IEEE Transactions on Visualization and Computer Graphics
(TVCG) | null | 10.1109/TVCG.2023.3259044 | null | cs.CV cs.GR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Geometric deep learning has sparked a rising interest in computer graphics to
perform shape understanding tasks, such as shape classification and semantic
segmentation. When the input is a polygonal surface, one has to suffer from the
irregular mesh structure. Motivated by the geometric spectral theory, we
introduce Laplacian2Mesh, a novel and flexible convolutional neural network
(CNN) framework for coping with irregular triangle meshes (vertices may have
any valence). By mapping the input mesh surface to the multi-dimensional
Laplacian-Beltrami space, Laplacian2Mesh enables one to perform shape analysis
tasks directly using the mature CNNs, without the need to deal with the
irregular connectivity of the mesh structure. We further define a mesh pooling
operation such that the receptive field of the network can be expanded while
retaining the original vertex set as well as the connections between them.
Besides, we introduce a channel-wise self-attention block to learn the
individual importance of feature ingredients. Laplacian2Mesh not only decouples
the geometry from the irregular connectivity of the mesh structure but also
better captures the global features that are central to shape classification
and segmentation. Extensive tests on various datasets demonstrate the
effectiveness and efficiency of Laplacian2Mesh, particularly in terms of the
capability of being vulnerable to noise to fulfill various learning tasks.
| [
{
"version": "v1",
"created": "Tue, 1 Feb 2022 10:10:13 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 10:57:44 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Dong",
"Qiujie",
""
],
[
"Wang",
"Zixiong",
""
],
[
"Li",
"Manyi",
""
],
[
"Gao",
"Junjie",
""
],
[
"Chen",
"Shuangmin",
""
],
[
"Shu",
"Zhenyu",
""
],
[
"Xin",
"Shiqing",
""
],
[
"Tu",
"Changhe",
""
],
[
"Wang",
"Wenping",
""
]
] | TITLE: Laplacian2Mesh: Laplacian-Based Mesh Understanding
ABSTRACT: Geometric deep learning has sparked a rising interest in computer graphics to
perform shape understanding tasks, such as shape classification and semantic
segmentation. When the input is a polygonal surface, one has to suffer from the
irregular mesh structure. Motivated by the geometric spectral theory, we
introduce Laplacian2Mesh, a novel and flexible convolutional neural network
(CNN) framework for coping with irregular triangle meshes (vertices may have
any valence). By mapping the input mesh surface to the multi-dimensional
Laplacian-Beltrami space, Laplacian2Mesh enables one to perform shape analysis
tasks directly using the mature CNNs, without the need to deal with the
irregular connectivity of the mesh structure. We further define a mesh pooling
operation such that the receptive field of the network can be expanded while
retaining the original vertex set as well as the connections between them.
Besides, we introduce a channel-wise self-attention block to learn the
individual importance of feature ingredients. Laplacian2Mesh not only decouples
the geometry from the irregular connectivity of the mesh structure but also
better captures the global features that are central to shape classification
and segmentation. Extensive tests on various datasets demonstrate the
effectiveness and efficiency of Laplacian2Mesh, particularly in terms of the
capability of being vulnerable to noise to fulfill various learning tasks.
| no_new_dataset | 0.953405 |
2202.03382 | Yuxin Fang | Yuxin Fang, Li Dong, Hangbo Bao, Xinggang Wang, Furu Wei | Corrupted Image Modeling for Self-Supervised Visual Pre-Training | ICLR 2023 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce Corrupted Image Modeling (CIM) for self-supervised visual
pre-training. CIM uses an auxiliary generator with a small trainable BEiT to
corrupt the input image instead of using artificial [MASK] tokens, where some
patches are randomly selected and replaced with plausible alternatives sampled
from the BEiT output distribution. Given this corrupted image, an enhancer
network learns to either recover all the original image pixels, or predict
whether each visual token is replaced by a generator sample or not. The
generator and the enhancer are simultaneously trained and synergistically
updated. After pre-training, the enhancer can be used as a high-capacity visual
encoder for downstream tasks. CIM is a general and flexible visual pre-training
framework that is suitable for various network architectures. For the first
time, CIM demonstrates that both ViT and CNN can learn rich visual
representations using a unified, non-Siamese framework. Experimental results
show that our approach achieves compelling results in vision benchmarks, such
as ImageNet classification and ADE20K semantic segmentation.
| [
{
"version": "v1",
"created": "Mon, 7 Feb 2022 17:59:04 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 05:58:10 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Fang",
"Yuxin",
""
],
[
"Dong",
"Li",
""
],
[
"Bao",
"Hangbo",
""
],
[
"Wang",
"Xinggang",
""
],
[
"Wei",
"Furu",
""
]
] | TITLE: Corrupted Image Modeling for Self-Supervised Visual Pre-Training
ABSTRACT: We introduce Corrupted Image Modeling (CIM) for self-supervised visual
pre-training. CIM uses an auxiliary generator with a small trainable BEiT to
corrupt the input image instead of using artificial [MASK] tokens, where some
patches are randomly selected and replaced with plausible alternatives sampled
from the BEiT output distribution. Given this corrupted image, an enhancer
network learns to either recover all the original image pixels, or predict
whether each visual token is replaced by a generator sample or not. The
generator and the enhancer are simultaneously trained and synergistically
updated. After pre-training, the enhancer can be used as a high-capacity visual
encoder for downstream tasks. CIM is a general and flexible visual pre-training
framework that is suitable for various network architectures. For the first
time, CIM demonstrates that both ViT and CNN can learn rich visual
representations using a unified, non-Siamese framework. Experimental results
show that our approach achieves compelling results in vision benchmarks, such
as ImageNet classification and ADE20K semantic segmentation.
| no_new_dataset | 0.949949 |
2202.07206 | Yasaman Razeghi | Yasaman Razeghi, Robert L. Logan IV, Matt Gardner, Sameer Singh | Impact of Pretraining Term Frequencies on Few-Shot Reasoning | null | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pretrained Language Models (LMs) have demonstrated ability to perform
numerical reasoning by extrapolating from a few examples in few-shot settings.
However, the extent to which this extrapolation relies on robust reasoning is
unclear. In this paper, we investigate how well these models reason with terms
that are less frequent in the pretraining data. In particular, we examine the
correlations between the model performance on test instances and the frequency
of terms from those instances in the pretraining data. We measure the strength
of this correlation for a number of GPT-based language models (pretrained on
the Pile dataset) on various numerical deduction tasks (e.g., arithmetic and
unit conversion). Our results consistently demonstrate that models are more
accurate on instances whose terms are more prevalent, in some cases above
$70\%$ (absolute) more accurate on the top 10\% frequent terms in comparison to
the bottom 10\%. Overall, although LMs exhibit strong performance at few-shot
numerical reasoning tasks, our results raise the question of how much models
actually generalize beyond pretraining data, and we encourage researchers to
take the pretraining data into account when interpreting evaluation results.
| [
{
"version": "v1",
"created": "Tue, 15 Feb 2022 05:43:54 GMT"
},
{
"version": "v2",
"created": "Tue, 24 May 2022 02:08:16 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Razeghi",
"Yasaman",
""
],
[
"Logan",
"Robert L.",
"IV"
],
[
"Gardner",
"Matt",
""
],
[
"Singh",
"Sameer",
""
]
] | TITLE: Impact of Pretraining Term Frequencies on Few-Shot Reasoning
ABSTRACT: Pretrained Language Models (LMs) have demonstrated ability to perform
numerical reasoning by extrapolating from a few examples in few-shot settings.
However, the extent to which this extrapolation relies on robust reasoning is
unclear. In this paper, we investigate how well these models reason with terms
that are less frequent in the pretraining data. In particular, we examine the
correlations between the model performance on test instances and the frequency
of terms from those instances in the pretraining data. We measure the strength
of this correlation for a number of GPT-based language models (pretrained on
the Pile dataset) on various numerical deduction tasks (e.g., arithmetic and
unit conversion). Our results consistently demonstrate that models are more
accurate on instances whose terms are more prevalent, in some cases above
$70\%$ (absolute) more accurate on the top 10\% frequent terms in comparison to
the bottom 10\%. Overall, although LMs exhibit strong performance at few-shot
numerical reasoning tasks, our results raise the question of how much models
actually generalize beyond pretraining data, and we encourage researchers to
take the pretraining data into account when interpreting evaluation results.
| no_new_dataset | 0.951774 |
2202.08192 | Zitong Yu | Zitong Yu, Ajian Liu, Chenxu Zhao, Kevin H. M. Cheng, Xu Cheng,
Guoying Zhao | Flexible-Modal Face Anti-Spoofing: A Benchmark | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Face anti-spoofing (FAS) plays a vital role in securing face recognition
systems from presentation attacks. Benefitted from the maturing camera sensors,
single-modal (RGB) and multi-modal (e.g., RGB+Depth) FAS has been applied in
various scenarios with different configurations of sensors/modalities. Existing
single- and multi-modal FAS methods usually separately train and deploy models
for each possible modality scenario, which might be redundant and inefficient.
Can we train a unified model, and flexibly deploy it under various modality
scenarios? In this paper, we establish the first flexible-modal FAS benchmark
with the principle `train one for all'. To be specific, with trained
multi-modal (RGB+Depth+IR) FAS models, both intra- and cross-dataset testings
are conducted on four flexible-modal sub-protocols (RGB, RGB+Depth, RGB+IR, and
RGB+Depth+IR). We also investigate prevalent deep models and feature fusion
strategies for flexible-modal FAS. We hope this new benchmark will facilitate
the future research of the multi-modal FAS. The protocols and codes are
available at https://github.com/ZitongYu/Flex-Modal-FAS.
| [
{
"version": "v1",
"created": "Wed, 16 Feb 2022 16:55:39 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 16:04:15 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 00:52:59 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Yu",
"Zitong",
""
],
[
"Liu",
"Ajian",
""
],
[
"Zhao",
"Chenxu",
""
],
[
"Cheng",
"Kevin H. M.",
""
],
[
"Cheng",
"Xu",
""
],
[
"Zhao",
"Guoying",
""
]
] | TITLE: Flexible-Modal Face Anti-Spoofing: A Benchmark
ABSTRACT: Face anti-spoofing (FAS) plays a vital role in securing face recognition
systems from presentation attacks. Benefitted from the maturing camera sensors,
single-modal (RGB) and multi-modal (e.g., RGB+Depth) FAS has been applied in
various scenarios with different configurations of sensors/modalities. Existing
single- and multi-modal FAS methods usually separately train and deploy models
for each possible modality scenario, which might be redundant and inefficient.
Can we train a unified model, and flexibly deploy it under various modality
scenarios? In this paper, we establish the first flexible-modal FAS benchmark
with the principle `train one for all'. To be specific, with trained
multi-modal (RGB+Depth+IR) FAS models, both intra- and cross-dataset testings
are conducted on four flexible-modal sub-protocols (RGB, RGB+Depth, RGB+IR, and
RGB+Depth+IR). We also investigate prevalent deep models and feature fusion
strategies for flexible-modal FAS. We hope this new benchmark will facilitate
the future research of the multi-modal FAS. The protocols and codes are
available at https://github.com/ZitongYu/Flex-Modal-FAS.
| no_new_dataset | 0.714827 |
2202.10037 | N. Sukumar | Alvin Chen, N. Sukumar | Stabilization-free virtual element method for plane elasticity | 38 pages, 11 figures | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the construction and application of a first order
stabilization-free virtual element method to problems in plane elasticity.
Well-posedness and error estimates of the discrete problem are established. The
method is assessed on a series of well-known benchmark problems from linear
elasticity and numerical results are presented that affirm the optimal
convergence rate of the virtual element method in the $L^2$ norm and the energy
seminorm.
| [
{
"version": "v1",
"created": "Mon, 21 Feb 2022 08:12:32 GMT"
},
{
"version": "v2",
"created": "Fri, 16 Sep 2022 17:42:03 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 16:42:03 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Chen",
"Alvin",
""
],
[
"Sukumar",
"N.",
""
]
] | TITLE: Stabilization-free virtual element method for plane elasticity
ABSTRACT: We present the construction and application of a first order
stabilization-free virtual element method to problems in plane elasticity.
Well-posedness and error estimates of the discrete problem are established. The
method is assessed on a series of well-known benchmark problems from linear
elasticity and numerical results are presented that affirm the optimal
convergence rate of the virtual element method in the $L^2$ norm and the energy
seminorm.
| no_new_dataset | 0.942029 |
2203.14360 | Jinkun Cao | Jinkun Cao, Jiangmiao Pang, Xinshuo Weng, Rawal Khirodkar, Kris Kitani | Observation-Centric SORT: Rethinking SORT for Robust Multi-Object
Tracking | Accepted by CVPR 2023. 8 pages + 10 pages of appendix. Renamed OOS as
Observation-centric Re-Update (ORU) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Kalman filter (KF) based methods for multi-object tracking (MOT) make an
assumption that objects move linearly. While this assumption is acceptable for
very short periods of occlusion, linear estimates of motion for prolonged time
can be highly inaccurate. Moreover, when there is no measurement available to
update Kalman filter parameters, the standard convention is to trust the priori
state estimations for posteriori update. This leads to the accumulation of
errors during a period of occlusion. The error causes significant motion
direction variance in practice. In this work, we show that a basic Kalman
filter can still obtain state-of-the-art tracking performance if proper care is
taken to fix the noise accumulated during occlusion. Instead of relying only on
the linear state estimate (i.e., estimation-centric approach), we use object
observations (i.e., the measurements by object detector) to compute a virtual
trajectory over the occlusion period to fix the error accumulation of filter
parameters during the occlusion period. This allows more time steps to correct
errors accumulated during occlusion. We name our method Observation-Centric
SORT (OC-SORT). It remains Simple, Online, and Real-Time but improves
robustness during occlusion and non-linear motion. Given off-the-shelf
detections as input, OC-SORT runs at 700+ FPS on a single CPU. It achieves
state-of-the-art on multiple datasets, including MOT17, MOT20, KITTI, head
tracking, and especially DanceTrack where the object motion is highly
non-linear. The code and models are available at
\url{https://github.com/noahcao/OC_SORT}.
| [
{
"version": "v1",
"created": "Sun, 27 Mar 2022 17:57:08 GMT"
},
{
"version": "v2",
"created": "Tue, 14 Mar 2023 03:32:35 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 03:49:36 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Cao",
"Jinkun",
""
],
[
"Pang",
"Jiangmiao",
""
],
[
"Weng",
"Xinshuo",
""
],
[
"Khirodkar",
"Rawal",
""
],
[
"Kitani",
"Kris",
""
]
] | TITLE: Observation-Centric SORT: Rethinking SORT for Robust Multi-Object
Tracking
ABSTRACT: Kalman filter (KF) based methods for multi-object tracking (MOT) make an
assumption that objects move linearly. While this assumption is acceptable for
very short periods of occlusion, linear estimates of motion for prolonged time
can be highly inaccurate. Moreover, when there is no measurement available to
update Kalman filter parameters, the standard convention is to trust the priori
state estimations for posteriori update. This leads to the accumulation of
errors during a period of occlusion. The error causes significant motion
direction variance in practice. In this work, we show that a basic Kalman
filter can still obtain state-of-the-art tracking performance if proper care is
taken to fix the noise accumulated during occlusion. Instead of relying only on
the linear state estimate (i.e., estimation-centric approach), we use object
observations (i.e., the measurements by object detector) to compute a virtual
trajectory over the occlusion period to fix the error accumulation of filter
parameters during the occlusion period. This allows more time steps to correct
errors accumulated during occlusion. We name our method Observation-Centric
SORT (OC-SORT). It remains Simple, Online, and Real-Time but improves
robustness during occlusion and non-linear motion. Given off-the-shelf
detections as input, OC-SORT runs at 700+ FPS on a single CPU. It achieves
state-of-the-art on multiple datasets, including MOT17, MOT20, KITTI, head
tracking, and especially DanceTrack where the object motion is highly
non-linear. The code and models are available at
\url{https://github.com/noahcao/OC_SORT}.
| no_new_dataset | 0.953492 |
2204.06108 | Giang Tran | Nicholas Richardson, Hayden Schaeffer, Giang Tran | SRMD: Sparse Random Mode Decomposition | null | null | null | null | eess.SP cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Signal decomposition and multiscale signal analysis provide many useful tools
for time-frequency analysis. We proposed a random feature method for analyzing
time-series data by constructing a sparse approximation to the spectrogram. The
randomization is both in the time window locations and the frequency sampling,
which lowers the overall sampling and computational cost. The sparsification of
the spectrogram leads to a sharp separation between time-frequency clusters
which makes it easier to identify intrinsic modes, and thus leads to a new
data-driven mode decomposition. The applications include signal representation,
outlier removal, and mode decomposition. On the benchmark tests, we show that
our approach outperforms other state-of-the-art decomposition methods.
| [
{
"version": "v1",
"created": "Tue, 12 Apr 2022 22:40:10 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 19:06:38 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Richardson",
"Nicholas",
""
],
[
"Schaeffer",
"Hayden",
""
],
[
"Tran",
"Giang",
""
]
] | TITLE: SRMD: Sparse Random Mode Decomposition
ABSTRACT: Signal decomposition and multiscale signal analysis provide many useful tools
for time-frequency analysis. We proposed a random feature method for analyzing
time-series data by constructing a sparse approximation to the spectrogram. The
randomization is both in the time window locations and the frequency sampling,
which lowers the overall sampling and computational cost. The sparsification of
the spectrogram leads to a sharp separation between time-frequency clusters
which makes it easier to identify intrinsic modes, and thus leads to a new
data-driven mode decomposition. The applications include signal representation,
outlier removal, and mode decomposition. On the benchmark tests, we show that
our approach outperforms other state-of-the-art decomposition methods.
| no_new_dataset | 0.949623 |
2205.00760 | Zhili Chen | Zhili Chen, Zian Qian, Sukai Wang, Qifeng Chen | Point Cloud Compression with Sibling Context and Surface Priors | null | null | 10.1007/978-3-031-19839-7_43 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel octree-based multi-level framework for large-scale point
cloud compression, which can organize sparse and unstructured point clouds in a
memory-efficient way. In this framework, we propose a new entropy model that
explores the hierarchical dependency in an octree using the context of
siblings' children, ancestors, and neighbors to encode the occupancy
information of each non-leaf octree node into a bitstream. Moreover, we locally
fit quadratic surfaces with a voxel-based geometry-aware module to provide
geometric priors in entropy encoding. These strong priors empower our entropy
framework to encode the octree into a more compact bitstream. In the decoding
stage, we apply a two-step heuristic strategy to restore point clouds with
better reconstruction quality. The quantitative evaluation shows that our
method outperforms state-of-the-art baselines with a bitrate improvement of
11-16% and 12-14% on the KITTI Odometry and nuScenes datasets, respectively.
| [
{
"version": "v1",
"created": "Mon, 2 May 2022 09:13:26 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Chen",
"Zhili",
""
],
[
"Qian",
"Zian",
""
],
[
"Wang",
"Sukai",
""
],
[
"Chen",
"Qifeng",
""
]
] | TITLE: Point Cloud Compression with Sibling Context and Surface Priors
ABSTRACT: We present a novel octree-based multi-level framework for large-scale point
cloud compression, which can organize sparse and unstructured point clouds in a
memory-efficient way. In this framework, we propose a new entropy model that
explores the hierarchical dependency in an octree using the context of
siblings' children, ancestors, and neighbors to encode the occupancy
information of each non-leaf octree node into a bitstream. Moreover, we locally
fit quadratic surfaces with a voxel-based geometry-aware module to provide
geometric priors in entropy encoding. These strong priors empower our entropy
framework to encode the octree into a more compact bitstream. In the decoding
stage, we apply a two-step heuristic strategy to restore point clouds with
better reconstruction quality. The quantitative evaluation shows that our
method outperforms state-of-the-art baselines with a bitrate improvement of
11-16% and 12-14% on the KITTI Odometry and nuScenes datasets, respectively.
| no_new_dataset | 0.948537 |
2206.06513 | Luis Scoccola | Luis Scoccola and Jose A. Perea | FibeRed: Fiberwise Dimensionality Reduction of Topologically Complex
Data with Vector Bundles | 15 pages + 10 page appendix, 15 figures + 1 table. To appear in
proceedings of 39th International Symposium on Computational Geometry | null | null | null | cs.CG cs.LG math.AT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Datasets with non-trivial large scale topology can be hard to embed in
low-dimensional Euclidean space with existing dimensionality reduction
algorithms. We propose to model topologically complex datasets using vector
bundles, in such a way that the base space accounts for the large scale
topology, while the fibers account for the local geometry. This allows one to
reduce the dimensionality of the fibers, while preserving the large scale
topology. We formalize this point of view, and, as an application, we describe
an algorithm which takes as input a dataset together with an initial
representation of it in Euclidean space, assumed to recover part of its large
scale topology, and outputs a new representation that integrates local
representations, obtained through local linear dimensionality reduction, along
the initial global representation. We demonstrate this algorithm on examples
coming from dynamical systems and chemistry. In these examples, our algorithm
is able to learn topologically faithful embeddings of the data in lower target
dimension than various well known metric-based dimensionality reduction
algorithms.
| [
{
"version": "v1",
"created": "Mon, 13 Jun 2022 22:53:46 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 19:41:03 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Scoccola",
"Luis",
""
],
[
"Perea",
"Jose A.",
""
]
] | TITLE: FibeRed: Fiberwise Dimensionality Reduction of Topologically Complex
Data with Vector Bundles
ABSTRACT: Datasets with non-trivial large scale topology can be hard to embed in
low-dimensional Euclidean space with existing dimensionality reduction
algorithms. We propose to model topologically complex datasets using vector
bundles, in such a way that the base space accounts for the large scale
topology, while the fibers account for the local geometry. This allows one to
reduce the dimensionality of the fibers, while preserving the large scale
topology. We formalize this point of view, and, as an application, we describe
an algorithm which takes as input a dataset together with an initial
representation of it in Euclidean space, assumed to recover part of its large
scale topology, and outputs a new representation that integrates local
representations, obtained through local linear dimensionality reduction, along
the initial global representation. We demonstrate this algorithm on examples
coming from dynamical systems and chemistry. In these examples, our algorithm
is able to learn topologically faithful embeddings of the data in lower target
dimension than various well known metric-based dimensionality reduction
algorithms.
| no_new_dataset | 0.947866 |
2208.07441 | Myounggyu Won | Jibran Ali Abbasi, Navid Mohammad Imran, Lokesh Chandra Das, Myounggyu
Won | WatchPed: Pedestrian Crossing Intention Prediction Using Embedded
Sensors of Smartwatch | null | null | null | null | cs.CV cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The pedestrian crossing intention prediction problem is to estimate whether
or not the target pedestrian will cross the street. State-of-the-art techniques
heavily depend on visual data acquired through the front camera of the
ego-vehicle to make a prediction of the pedestrian's crossing intention. Hence,
the efficiency of current methodologies tends to decrease notably in situations
where visual input is imprecise, for instance, when the distance between the
pedestrian and ego-vehicle is considerable or the illumination levels are
inadequate. To address the limitation, in this paper, we present the design,
implementation, and evaluation of the first-of-its-kind pedestrian crossing
intention prediction model based on integration of motion sensor data gathered
through the smartwatch (or smartphone) of the pedestrian. We propose an
innovative machine learning framework that effectively integrates motion sensor
data with visual input to enhance the predictive accuracy significantly,
particularly in scenarios where visual data may be unreliable. Moreover, we
perform an extensive data collection process and introduce the first pedestrian
intention prediction dataset that features synchronized motion sensor data. The
dataset comprises 255 video clips that encompass diverse distances and lighting
conditions. We trained our model using the widely-used JAAD and our own
datasets and compare the performance with a state-of-the-art model. The results
demonstrate that our model outperforms the current state-of-the-art method,
particularly in cases where the distance between the pedestrian and the
observer is considerable (more than 70 meters) and the lighting conditions are
inadequate.
| [
{
"version": "v1",
"created": "Mon, 15 Aug 2022 21:27:21 GMT"
},
{
"version": "v2",
"created": "Sun, 12 Mar 2023 04:16:26 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 03:50:31 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Abbasi",
"Jibran Ali",
""
],
[
"Imran",
"Navid Mohammad",
""
],
[
"Das",
"Lokesh Chandra",
""
],
[
"Won",
"Myounggyu",
""
]
] | TITLE: WatchPed: Pedestrian Crossing Intention Prediction Using Embedded
Sensors of Smartwatch
ABSTRACT: The pedestrian crossing intention prediction problem is to estimate whether
or not the target pedestrian will cross the street. State-of-the-art techniques
heavily depend on visual data acquired through the front camera of the
ego-vehicle to make a prediction of the pedestrian's crossing intention. Hence,
the efficiency of current methodologies tends to decrease notably in situations
where visual input is imprecise, for instance, when the distance between the
pedestrian and ego-vehicle is considerable or the illumination levels are
inadequate. To address the limitation, in this paper, we present the design,
implementation, and evaluation of the first-of-its-kind pedestrian crossing
intention prediction model based on integration of motion sensor data gathered
through the smartwatch (or smartphone) of the pedestrian. We propose an
innovative machine learning framework that effectively integrates motion sensor
data with visual input to enhance the predictive accuracy significantly,
particularly in scenarios where visual data may be unreliable. Moreover, we
perform an extensive data collection process and introduce the first pedestrian
intention prediction dataset that features synchronized motion sensor data. The
dataset comprises 255 video clips that encompass diverse distances and lighting
conditions. We trained our model using the widely-used JAAD and our own
datasets and compare the performance with a state-of-the-art model. The results
demonstrate that our model outperforms the current state-of-the-art method,
particularly in cases where the distance between the pedestrian and the
observer is considerable (more than 70 meters) and the lighting conditions are
inadequate.
| new_dataset | 0.966505 |
2208.08664 | Bahjat Kawar | Bahjat Kawar, Roy Ganz, Michael Elad | Enhancing Diffusion-Based Image Synthesis with Robust Classifier
Guidance | Accepted to TMLR | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Denoising diffusion probabilistic models (DDPMs) are a recent family of
generative models that achieve state-of-the-art results. In order to obtain
class-conditional generation, it was suggested to guide the diffusion process
by gradients from a time-dependent classifier. While the idea is theoretically
sound, deep learning-based classifiers are infamously susceptible to
gradient-based adversarial attacks. Therefore, while traditional classifiers
may achieve good accuracy scores, their gradients are possibly unreliable and
might hinder the improvement of the generation results. Recent work discovered
that adversarially robust classifiers exhibit gradients that are aligned with
human perception, and these could better guide a generative process towards
semantically meaningful images. We utilize this observation by defining and
training a time-dependent adversarially robust classifier and use it as
guidance for a generative diffusion model. In experiments on the highly
challenging and diverse ImageNet dataset, our scheme introduces significantly
more intelligible intermediate gradients, better alignment with theoretical
findings, as well as improved generation results under several evaluation
metrics. Furthermore, we conduct an opinion survey whose findings indicate that
human raters prefer our method's results.
| [
{
"version": "v1",
"created": "Thu, 18 Aug 2022 06:51:23 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 18:10:43 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Kawar",
"Bahjat",
""
],
[
"Ganz",
"Roy",
""
],
[
"Elad",
"Michael",
""
]
] | TITLE: Enhancing Diffusion-Based Image Synthesis with Robust Classifier
Guidance
ABSTRACT: Denoising diffusion probabilistic models (DDPMs) are a recent family of
generative models that achieve state-of-the-art results. In order to obtain
class-conditional generation, it was suggested to guide the diffusion process
by gradients from a time-dependent classifier. While the idea is theoretically
sound, deep learning-based classifiers are infamously susceptible to
gradient-based adversarial attacks. Therefore, while traditional classifiers
may achieve good accuracy scores, their gradients are possibly unreliable and
might hinder the improvement of the generation results. Recent work discovered
that adversarially robust classifiers exhibit gradients that are aligned with
human perception, and these could better guide a generative process towards
semantically meaningful images. We utilize this observation by defining and
training a time-dependent adversarially robust classifier and use it as
guidance for a generative diffusion model. In experiments on the highly
challenging and diverse ImageNet dataset, our scheme introduces significantly
more intelligible intermediate gradients, better alignment with theoretical
findings, as well as improved generation results under several evaluation
metrics. Furthermore, we conduct an opinion survey whose findings indicate that
human raters prefer our method's results.
| no_new_dataset | 0.946597 |
2208.14888 | JoonHo Lee | JoonHo Lee and Gyemin Lee | Feature Alignment by Uncertainty and Self-Training for Source-Free
Unsupervised Domain Adaptation | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most unsupervised domain adaptation (UDA) methods assume that labeled source
images are available during model adaptation. However, this assumption is often
infeasible owing to confidentiality issues or memory constraints on mobile
devices. Some recently developed approaches do not require source images during
adaptation, but they show limited performance on perturbed images. To address
these problems, we propose a novel source-free UDA method that uses only a
pre-trained source model and unlabeled target images. Our method captures the
aleatoric uncertainty by incorporating data augmentation and trains the feature
generator with two consistency objectives. The feature generator is encouraged
to learn consistent visual features away from the decision boundaries of the
head classifier. Thus, the adapted model becomes more robust to image
perturbations. Inspired by self-supervised learning, our method promotes
inter-space alignment between the prediction space and the feature space while
incorporating intra-space consistency within the feature space to reduce the
domain gap between the source and target domains. We also consider epistemic
uncertainty to boost the model adaptation performance. Extensive experiments on
popular UDA benchmark datasets demonstrate that the proposed source-free method
is comparable or even superior to vanilla UDA methods. Moreover, the adapted
models show more robust results when input images are perturbed.
| [
{
"version": "v1",
"created": "Wed, 31 Aug 2022 14:28:36 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 11:48:19 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Lee",
"JoonHo",
""
],
[
"Lee",
"Gyemin",
""
]
] | TITLE: Feature Alignment by Uncertainty and Self-Training for Source-Free
Unsupervised Domain Adaptation
ABSTRACT: Most unsupervised domain adaptation (UDA) methods assume that labeled source
images are available during model adaptation. However, this assumption is often
infeasible owing to confidentiality issues or memory constraints on mobile
devices. Some recently developed approaches do not require source images during
adaptation, but they show limited performance on perturbed images. To address
these problems, we propose a novel source-free UDA method that uses only a
pre-trained source model and unlabeled target images. Our method captures the
aleatoric uncertainty by incorporating data augmentation and trains the feature
generator with two consistency objectives. The feature generator is encouraged
to learn consistent visual features away from the decision boundaries of the
head classifier. Thus, the adapted model becomes more robust to image
perturbations. Inspired by self-supervised learning, our method promotes
inter-space alignment between the prediction space and the feature space while
incorporating intra-space consistency within the feature space to reduce the
domain gap between the source and target domains. We also consider epistemic
uncertainty to boost the model adaptation performance. Extensive experiments on
popular UDA benchmark datasets demonstrate that the proposed source-free method
is comparable or even superior to vanilla UDA methods. Moreover, the adapted
models show more robust results when input images are perturbed.
| no_new_dataset | 0.944382 |
2210.04994 | Javier Antor\'an | Javier Antor\'an, Shreyas Padhy, Riccardo Barbano, Eric Nalisnick,
David Janz, Jos\'e Miguel Hern\'andez-Lobato | Sampling-based inference for large linear models, with application to
linearised Laplace | Published at ICLR 2023. This latest Arxiv version is extended with a
demonstration of the proposed methods on the Imagenet dataset | null | null | null | stat.ML cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large-scale linear models are ubiquitous throughout machine learning, with
contemporary application as surrogate models for neural network uncertainty
quantification; that is, the linearised Laplace method. Alas, the computational
cost associated with Bayesian linear models constrains this method's
application to small networks, small output spaces and small datasets. We
address this limitation by introducing a scalable sample-based Bayesian
inference method for conjugate Gaussian multi-output linear models, together
with a matching method for hyperparameter (regularisation) selection.
Furthermore, we use a classic feature normalisation method (the g-prior) to
resolve a previously highlighted pathology of the linearised Laplace method.
Together, these contributions allow us to perform linearised neural network
inference with ResNet-18 on CIFAR100 (11M parameters, 100 outputs x 50k
datapoints), with ResNet-50 on Imagenet (50M parameters, 1000 outputs x 1.2M
datapoints) and with a U-Net on a high-resolution tomographic reconstruction
task (2M parameters, 251k output~dimensions).
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2022 19:59:40 GMT"
},
{
"version": "v2",
"created": "Fri, 3 Mar 2023 14:40:23 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 13:24:29 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Antorán",
"Javier",
""
],
[
"Padhy",
"Shreyas",
""
],
[
"Barbano",
"Riccardo",
""
],
[
"Nalisnick",
"Eric",
""
],
[
"Janz",
"David",
""
],
[
"Hernández-Lobato",
"José Miguel",
""
]
] | TITLE: Sampling-based inference for large linear models, with application to
linearised Laplace
ABSTRACT: Large-scale linear models are ubiquitous throughout machine learning, with
contemporary application as surrogate models for neural network uncertainty
quantification; that is, the linearised Laplace method. Alas, the computational
cost associated with Bayesian linear models constrains this method's
application to small networks, small output spaces and small datasets. We
address this limitation by introducing a scalable sample-based Bayesian
inference method for conjugate Gaussian multi-output linear models, together
with a matching method for hyperparameter (regularisation) selection.
Furthermore, we use a classic feature normalisation method (the g-prior) to
resolve a previously highlighted pathology of the linearised Laplace method.
Together, these contributions allow us to perform linearised neural network
inference with ResNet-18 on CIFAR100 (11M parameters, 100 outputs x 50k
datapoints), with ResNet-50 on Imagenet (50M parameters, 1000 outputs x 1.2M
datapoints) and with a U-Net on a high-resolution tomographic reconstruction
task (2M parameters, 251k output~dimensions).
| no_new_dataset | 0.943764 |
2210.15586 | Huayi Zhou | Huayi Zhou, Fei Jiang, Jiaxin Si, Hongtao Lu | Joint Multi-Person Body Detection and Orientation Estimation via One
Unified Embedding | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human body orientation estimation (HBOE) is widely applied into various
applications, including robotics, surveillance, pedestrian analysis and
autonomous driving. Although many approaches have been addressing the HBOE
problem from specific under-controlled scenes to challenging in-the-wild
environments, they assume human instances are already detected and take a well
cropped sub-image as the input. This setting is less efficient and prone to
errors in real application, such as crowds of people. In the paper, we propose
a single-stage end-to-end trainable framework for tackling the HBOE problem
with multi-persons. By integrating the prediction of bounding boxes and
direction angles in one embedding, our method can jointly estimate the location
and orientation of all bodies in one image directly. Our key idea is to
integrate the HBOE task into the multi-scale anchor channel predictions of
persons for concurrently benefiting from engaged intermediate features.
Therefore, our approach can naturally adapt to difficult instances involving
low resolution and occlusion as in object detection. We validated the
efficiency and effectiveness of our method in the recently presented benchmark
MEBOW with extensive experiments. Besides, we completed ambiguous instances
ignored by the MEBOW dataset, and provided corresponding weak body-orientation
labels to keep the integrity and consistency of it for supporting studies
toward multi-persons. Our work is available at
https://github.com/hnuzhy/JointBDOE.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2022 16:22:50 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 08:25:00 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhou",
"Huayi",
""
],
[
"Jiang",
"Fei",
""
],
[
"Si",
"Jiaxin",
""
],
[
"Lu",
"Hongtao",
""
]
] | TITLE: Joint Multi-Person Body Detection and Orientation Estimation via One
Unified Embedding
ABSTRACT: Human body orientation estimation (HBOE) is widely applied into various
applications, including robotics, surveillance, pedestrian analysis and
autonomous driving. Although many approaches have been addressing the HBOE
problem from specific under-controlled scenes to challenging in-the-wild
environments, they assume human instances are already detected and take a well
cropped sub-image as the input. This setting is less efficient and prone to
errors in real application, such as crowds of people. In the paper, we propose
a single-stage end-to-end trainable framework for tackling the HBOE problem
with multi-persons. By integrating the prediction of bounding boxes and
direction angles in one embedding, our method can jointly estimate the location
and orientation of all bodies in one image directly. Our key idea is to
integrate the HBOE task into the multi-scale anchor channel predictions of
persons for concurrently benefiting from engaged intermediate features.
Therefore, our approach can naturally adapt to difficult instances involving
low resolution and occlusion as in object detection. We validated the
efficiency and effectiveness of our method in the recently presented benchmark
MEBOW with extensive experiments. Besides, we completed ambiguous instances
ignored by the MEBOW dataset, and provided corresponding weak body-orientation
labels to keep the integrity and consistency of it for supporting studies
toward multi-persons. Our work is available at
https://github.com/hnuzhy/JointBDOE.
| no_new_dataset | 0.909667 |
2211.01452 | Dacheng Li | Dacheng Li, Rulin Shao, Hongyi Wang, Han Guo, Eric P. Xing, Hao Zhang | MPCFormer: fast, performant and private Transformer inference with MPC | null | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Enabling private inference is crucial for many cloud inference services that
are based on Transformer models. However, existing private inference solutions
can increase the inference latency by more than 60x or significantly compromise
the inference quality. In this paper, we design the framework MPCFORMER as a
practical solution, using Secure Multi-Party Computation (MPC) and Knowledge
Distillation (KD). Through extensive evaluations, we show that MPCFORMER
significantly speeds up Transformer inference in MPC settings while achieving
similar ML performance to the input model. On the IMDb dataset, it achieves
similar performance to BERTBASE, while being 5.3x faster. On the GLUE
benchmark, it achieves 97% performance of BERTBASE with a 2.2x speedup.
MPCFORMER remains effective with different trained Transformer weights such as
ROBERTABASE and larger models including BERTLarge. Code is available at
https://github.com/MccRee177/MPCFormer.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2022 19:43:22 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 06:51:31 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Li",
"Dacheng",
""
],
[
"Shao",
"Rulin",
""
],
[
"Wang",
"Hongyi",
""
],
[
"Guo",
"Han",
""
],
[
"Xing",
"Eric P.",
""
],
[
"Zhang",
"Hao",
""
]
] | TITLE: MPCFormer: fast, performant and private Transformer inference with MPC
ABSTRACT: Enabling private inference is crucial for many cloud inference services that
are based on Transformer models. However, existing private inference solutions
can increase the inference latency by more than 60x or significantly compromise
the inference quality. In this paper, we design the framework MPCFORMER as a
practical solution, using Secure Multi-Party Computation (MPC) and Knowledge
Distillation (KD). Through extensive evaluations, we show that MPCFORMER
significantly speeds up Transformer inference in MPC settings while achieving
similar ML performance to the input model. On the IMDb dataset, it achieves
similar performance to BERTBASE, while being 5.3x faster. On the GLUE
benchmark, it achieves 97% performance of BERTBASE with a 2.2x speedup.
MPCFORMER remains effective with different trained Transformer weights such as
ROBERTABASE and larger models including BERTLarge. Code is available at
https://github.com/MccRee177/MPCFormer.
| no_new_dataset | 0.94625 |
2211.03284 | Zhengkun Tian | Zhengkun Tian, Hongyu Xiang, Min Li, Feifei Lin, Ke Ding, Guanglu Wan | Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying
Peak-First Regularization | Accepted by ICASSP 2023(5 pages, 2 figures) | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The CTC model has been widely applied to many application scenarios because
of its simple structure, excellent performance, and fast inference speed. There
are many peaks in the probability distribution predicted by the CTC models, and
each peak represents a non-blank token. The recognition latency of CTC models
can be reduced by encouraging the model to predict peaks earlier. Existing
methods to reduce latency require modifying the transition relationship between
tokens in the forward-backward algorithm, and the gradient calculation. Some of
these methods even depend on the forced alignment results provided by other
pretrained models. The above methods are complex to implement. To reduce the
peak latency, we propose a simple and novel method named peak-first
regularization, which utilizes a frame-wise knowledge distillation function to
force the probability distribution of the CTC model to shift left along the
time axis instead of directly modifying the calculation process of CTC loss and
gradients. All the experiments are conducted on a Chinese Mandarin dataset
AISHELL-1. We have verified the effectiveness of the proposed regularization on
both streaming and non-streaming CTC models respectively. The results show that
the proposed method can reduce the average peak latency by about 100 to 200
milliseconds with almost no degradation of recognition accuracy.
| [
{
"version": "v1",
"created": "Mon, 7 Nov 2022 03:36:00 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 03:21:26 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Tian",
"Zhengkun",
""
],
[
"Xiang",
"Hongyu",
""
],
[
"Li",
"Min",
""
],
[
"Lin",
"Feifei",
""
],
[
"Ding",
"Ke",
""
],
[
"Wan",
"Guanglu",
""
]
] | TITLE: Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying
Peak-First Regularization
ABSTRACT: The CTC model has been widely applied to many application scenarios because
of its simple structure, excellent performance, and fast inference speed. There
are many peaks in the probability distribution predicted by the CTC models, and
each peak represents a non-blank token. The recognition latency of CTC models
can be reduced by encouraging the model to predict peaks earlier. Existing
methods to reduce latency require modifying the transition relationship between
tokens in the forward-backward algorithm, and the gradient calculation. Some of
these methods even depend on the forced alignment results provided by other
pretrained models. The above methods are complex to implement. To reduce the
peak latency, we propose a simple and novel method named peak-first
regularization, which utilizes a frame-wise knowledge distillation function to
force the probability distribution of the CTC model to shift left along the
time axis instead of directly modifying the calculation process of CTC loss and
gradients. All the experiments are conducted on a Chinese Mandarin dataset
AISHELL-1. We have verified the effectiveness of the proposed regularization on
both streaming and non-streaming CTC models respectively. The results show that
the proposed method can reduce the average peak latency by about 100 to 200
milliseconds with almost no degradation of recognition accuracy.
| no_new_dataset | 0.949012 |
2211.11950 | Sunwook Hwang | Sunwook Hwang, Youngseok Kim, Seongwon Kim, Saewoong Bahk, Hyung-Sin
Kim | UpCycling: Semi-supervised 3D Object Detection without Sharing Raw-level
Unlabeled Scenes | We have updated the results to fix errors in the experimental
process, which resulted in some logical changes. We also have added new
experiments related to privacy protection. The previous version (v1) has been
discarded | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-supervised Learning (SSL) has received increasing attention in
autonomous driving to reduce the enormous burden of 3D annotation. In this
paper, we propose UpCycling, a novel SSL framework for 3D object detection with
zero additional raw-level point cloud: learning from unlabeled de-identified
intermediate features (i.e., smashed data) to preserve privacy. Since these
intermediate features are naturally produced by the inference pipeline, no
additional computation is required on autonomous vehicles. However, generating
effective consistency loss for unlabeled feature-level scene turns out to be a
critical challenge. The latest SSL frameworks for 3D object detection that
enforce consistency regularization between different augmentations of an
unlabeled raw-point scene become detrimental when applied to intermediate
features. To solve the problem, we introduce a novel combination of hybrid
pseudo labels and feature-level Ground Truth sampling (F-GT), which safely
augments unlabeled multi-type 3D scene features and provides high-quality
supervision. We implement UpCycling on two representative 3D object detection
models: SECOND-IoU and PV-RCNN. Experiments on widely-used datasets (Waymo,
KITTI, and Lyft) verify that UpCycling outperforms other augmentation methods
applied at the feature level. In addition, while preserving privacy, UpCycling
performs better or comparably to the state-of-the-art methods that utilize
raw-level unlabeled data in both domain adaptation and partial-label scenarios.
| [
{
"version": "v1",
"created": "Tue, 22 Nov 2022 02:04:09 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 11:43:07 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Hwang",
"Sunwook",
""
],
[
"Kim",
"Youngseok",
""
],
[
"Kim",
"Seongwon",
""
],
[
"Bahk",
"Saewoong",
""
],
[
"Kim",
"Hyung-Sin",
""
]
] | TITLE: UpCycling: Semi-supervised 3D Object Detection without Sharing Raw-level
Unlabeled Scenes
ABSTRACT: Semi-supervised Learning (SSL) has received increasing attention in
autonomous driving to reduce the enormous burden of 3D annotation. In this
paper, we propose UpCycling, a novel SSL framework for 3D object detection with
zero additional raw-level point cloud: learning from unlabeled de-identified
intermediate features (i.e., smashed data) to preserve privacy. Since these
intermediate features are naturally produced by the inference pipeline, no
additional computation is required on autonomous vehicles. However, generating
effective consistency loss for unlabeled feature-level scene turns out to be a
critical challenge. The latest SSL frameworks for 3D object detection that
enforce consistency regularization between different augmentations of an
unlabeled raw-point scene become detrimental when applied to intermediate
features. To solve the problem, we introduce a novel combination of hybrid
pseudo labels and feature-level Ground Truth sampling (F-GT), which safely
augments unlabeled multi-type 3D scene features and provides high-quality
supervision. We implement UpCycling on two representative 3D object detection
models: SECOND-IoU and PV-RCNN. Experiments on widely-used datasets (Waymo,
KITTI, and Lyft) verify that UpCycling outperforms other augmentation methods
applied at the feature level. In addition, while preserving privacy, UpCycling
performs better or comparably to the state-of-the-art methods that utilize
raw-level unlabeled data in both domain adaptation and partial-label scenarios.
| no_new_dataset | 0.947137 |
2211.12254 | Ashkan Mirzaei | Ashkan Mirzaei, Tristan Aumentado-Armstrong, Konstantinos G. Derpanis,
Jonathan Kelly, Marcus A. Brubaker, Igor Gilitschenski, Alex Levinshtein | SPIn-NeRF: Multiview Segmentation and Perceptual Inpainting with Neural
Radiance Fields | Project Page: https://spinnerf3d.github.io | The IEEE/CVF Conference on Computer Vision and Pattern Recognition
(CVPR) 2023 | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural Radiance Fields (NeRFs) have emerged as a popular approach for novel
view synthesis. While NeRFs are quickly being adapted for a wider set of
applications, intuitively editing NeRF scenes is still an open challenge. One
important editing task is the removal of unwanted objects from a 3D scene, such
that the replaced region is visually plausible and consistent with its context.
We refer to this task as 3D inpainting. In 3D, solutions must be both
consistent across multiple views and geometrically valid. In this paper, we
propose a novel 3D inpainting method that addresses these challenges. Given a
small set of posed images and sparse annotations in a single input image, our
framework first rapidly obtains a 3D segmentation mask for a target object.
Using the mask, a perceptual optimizationbased approach is then introduced that
leverages learned 2D image inpainters, distilling their information into 3D
space, while ensuring view consistency. We also address the lack of a diverse
benchmark for evaluating 3D scene inpainting methods by introducing a dataset
comprised of challenging real-world scenes. In particular, our dataset contains
views of the same scene with and without a target object, enabling more
principled benchmarking of the 3D inpainting task. We first demonstrate the
superiority of our approach on multiview segmentation, comparing to NeRFbased
methods and 2D segmentation approaches. We then evaluate on the task of 3D
inpainting, establishing state-ofthe-art performance against other NeRF
manipulation algorithms, as well as a strong 2D image inpainter baseline.
Project Page: https://spinnerf3d.github.io
| [
{
"version": "v1",
"created": "Tue, 22 Nov 2022 13:14:50 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 21:11:11 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Mirzaei",
"Ashkan",
""
],
[
"Aumentado-Armstrong",
"Tristan",
""
],
[
"Derpanis",
"Konstantinos G.",
""
],
[
"Kelly",
"Jonathan",
""
],
[
"Brubaker",
"Marcus A.",
""
],
[
"Gilitschenski",
"Igor",
""
],
[
"Levinshtein",
"Alex",
""
]
] | TITLE: SPIn-NeRF: Multiview Segmentation and Perceptual Inpainting with Neural
Radiance Fields
ABSTRACT: Neural Radiance Fields (NeRFs) have emerged as a popular approach for novel
view synthesis. While NeRFs are quickly being adapted for a wider set of
applications, intuitively editing NeRF scenes is still an open challenge. One
important editing task is the removal of unwanted objects from a 3D scene, such
that the replaced region is visually plausible and consistent with its context.
We refer to this task as 3D inpainting. In 3D, solutions must be both
consistent across multiple views and geometrically valid. In this paper, we
propose a novel 3D inpainting method that addresses these challenges. Given a
small set of posed images and sparse annotations in a single input image, our
framework first rapidly obtains a 3D segmentation mask for a target object.
Using the mask, a perceptual optimizationbased approach is then introduced that
leverages learned 2D image inpainters, distilling their information into 3D
space, while ensuring view consistency. We also address the lack of a diverse
benchmark for evaluating 3D scene inpainting methods by introducing a dataset
comprised of challenging real-world scenes. In particular, our dataset contains
views of the same scene with and without a target object, enabling more
principled benchmarking of the 3D inpainting task. We first demonstrate the
superiority of our approach on multiview segmentation, comparing to NeRFbased
methods and 2D segmentation approaches. We then evaluate on the task of 3D
inpainting, establishing state-ofthe-art performance against other NeRF
manipulation algorithms, as well as a strong 2D image inpainter baseline.
Project Page: https://spinnerf3d.github.io
| new_dataset | 0.964921 |
2211.13357 | Kevin Lin | Kevin Lin, Chung-Ching Lin, Lin Liang, Zicheng Liu, Lijuan Wang | MPT: Mesh Pre-Training with Transformers for Human Pose and Mesh
Reconstruction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Traditional methods of reconstructing 3D human pose and mesh from single
images rely on paired image-mesh datasets, which can be difficult and expensive
to obtain. Due to this limitation, model scalability is constrained as well as
reconstruction performance. Towards addressing the challenge, we introduce Mesh
Pre-Training (MPT), an effective pre-training strategy that leverages large
amounts of MoCap data to effectively perform pre-training at scale. We
introduce the use of MoCap-generated heatmaps as input representations to the
mesh regression transformer and propose a Masked Heatmap Modeling approach for
improving pre-training performance. This study demonstrates that pre-training
using the proposed MPT allows our models to perform effective inference without
requiring fine-tuning. We further show that fine-tuning the pre-trained MPT
model considerably improves the accuracy of human mesh reconstruction from
single images. Experimental results show that MPT outperforms previous
state-of-the-art methods on Human3.6M and 3DPW datasets. As a further
application, we benchmark and study MPT on the task of 3D hand reconstruction,
showing that our generic pre-training scheme generalizes well to hand pose
estimation and achieves promising reconstruction performance.
| [
{
"version": "v1",
"created": "Thu, 24 Nov 2022 00:02:13 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 18:05:40 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Lin",
"Kevin",
""
],
[
"Lin",
"Chung-Ching",
""
],
[
"Liang",
"Lin",
""
],
[
"Liu",
"Zicheng",
""
],
[
"Wang",
"Lijuan",
""
]
] | TITLE: MPT: Mesh Pre-Training with Transformers for Human Pose and Mesh
Reconstruction
ABSTRACT: Traditional methods of reconstructing 3D human pose and mesh from single
images rely on paired image-mesh datasets, which can be difficult and expensive
to obtain. Due to this limitation, model scalability is constrained as well as
reconstruction performance. Towards addressing the challenge, we introduce Mesh
Pre-Training (MPT), an effective pre-training strategy that leverages large
amounts of MoCap data to effectively perform pre-training at scale. We
introduce the use of MoCap-generated heatmaps as input representations to the
mesh regression transformer and propose a Masked Heatmap Modeling approach for
improving pre-training performance. This study demonstrates that pre-training
using the proposed MPT allows our models to perform effective inference without
requiring fine-tuning. We further show that fine-tuning the pre-trained MPT
model considerably improves the accuracy of human mesh reconstruction from
single images. Experimental results show that MPT outperforms previous
state-of-the-art methods on Human3.6M and 3DPW datasets. As a further
application, we benchmark and study MPT on the task of 3D hand reconstruction,
showing that our generic pre-training scheme generalizes well to hand pose
estimation and achieves promising reconstruction performance.
| no_new_dataset | 0.946001 |
2211.14594 | Yuhui Li | Yuhui Li, Zejia Wu, Chao Zhang, Hongyang Zhang | Direct-Effect Risk Minimization for Domain Generalization | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of out-of-distribution (o.o.d.) generalization where
spurious correlations of attributes vary across training and test domains. This
is known as the problem of correlation shift and has posed concerns on the
reliability of machine learning. In this work, we introduce the concepts of
direct and indirect effects from causal inference to the domain generalization
problem. We argue that models that learn direct effects minimize the worst-case
risk across correlation-shifted domains. To eliminate the indirect effects, our
algorithm consists of two stages: in the first stage, we learn an
indirect-effect representation by minimizing the prediction error of domain
labels using the representation and the class labels; in the second stage, we
remove the indirect effects learned in the first stage by matching each data
with another data of similar indirect-effect representation but of different
class labels in the training and validation phase. Our approach is shown to be
compatible with existing methods and improve the generalization performance of
them on correlation-shifted datasets. Experiments on 5 correlation-shifted
datasets and the DomainBed benchmark verify the effectiveness of our approach.
| [
{
"version": "v1",
"created": "Sat, 26 Nov 2022 15:35:36 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Jan 2023 15:52:23 GMT"
},
{
"version": "v3",
"created": "Wed, 1 Feb 2023 11:22:57 GMT"
},
{
"version": "v4",
"created": "Thu, 16 Mar 2023 03:21:22 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Li",
"Yuhui",
""
],
[
"Wu",
"Zejia",
""
],
[
"Zhang",
"Chao",
""
],
[
"Zhang",
"Hongyang",
""
]
] | TITLE: Direct-Effect Risk Minimization for Domain Generalization
ABSTRACT: We study the problem of out-of-distribution (o.o.d.) generalization where
spurious correlations of attributes vary across training and test domains. This
is known as the problem of correlation shift and has posed concerns on the
reliability of machine learning. In this work, we introduce the concepts of
direct and indirect effects from causal inference to the domain generalization
problem. We argue that models that learn direct effects minimize the worst-case
risk across correlation-shifted domains. To eliminate the indirect effects, our
algorithm consists of two stages: in the first stage, we learn an
indirect-effect representation by minimizing the prediction error of domain
labels using the representation and the class labels; in the second stage, we
remove the indirect effects learned in the first stage by matching each data
with another data of similar indirect-effect representation but of different
class labels in the training and validation phase. Our approach is shown to be
compatible with existing methods and improve the generalization performance of
them on correlation-shifted datasets. Experiments on 5 correlation-shifted
datasets and the DomainBed benchmark verify the effectiveness of our approach.
| no_new_dataset | 0.950869 |
2211.15736 | Yuzhang Shang | Yuzhang Shang, Zhihang Yuan, Bin Xie, Bingzhe Wu, Yan Yan | Post-training Quantization on Diffusion Models | Accepted to CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Denoising diffusion (score-based) generative models have recently achieved
significant accomplishments in generating realistic and diverse data. These
approaches define a forward diffusion process for transforming data into noise
and a backward denoising process for sampling data from noise. Unfortunately,
the generation process of current denoising diffusion models is notoriously
slow due to the lengthy iterative noise estimations, which rely on cumbersome
neural networks. It prevents the diffusion models from being widely deployed,
especially on edge devices. Previous works accelerate the generation process of
diffusion model (DM) via finding shorter yet effective sampling trajectories.
However, they overlook the cost of noise estimation with a heavy network in
every iteration. In this work, we accelerate generation from the perspective of
compressing the noise estimation network. Due to the difficulty of retraining
DMs, we exclude mainstream training-aware compression paradigms and introduce
post-training quantization (PTQ) into DM acceleration. However, the output
distributions of noise estimation networks change with time-step, making
previous PTQ methods fail in DMs since they are designed for single-time step
scenarios. To devise a DM-specific PTQ method, we explore PTQ on DM in three
aspects: quantized operations, calibration dataset, and calibration metric. We
summarize and use several observations derived from all-inclusive
investigations to formulate our method, which especially targets the unique
multi-time-step structure of DMs. Experimentally, our method can directly
quantize full-precision DMs into 8-bit models while maintaining or even
improving their performance in a training-free manner. Importantly, our method
can serve as a plug-and-play module on other fast-sampling methods, e.g., DDIM.
The code is available at https://github.com/42Shawn/PTQ4DM .
| [
{
"version": "v1",
"created": "Mon, 28 Nov 2022 19:33:39 GMT"
},
{
"version": "v2",
"created": "Thu, 9 Mar 2023 21:02:41 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 16:51:02 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Shang",
"Yuzhang",
""
],
[
"Yuan",
"Zhihang",
""
],
[
"Xie",
"Bin",
""
],
[
"Wu",
"Bingzhe",
""
],
[
"Yan",
"Yan",
""
]
] | TITLE: Post-training Quantization on Diffusion Models
ABSTRACT: Denoising diffusion (score-based) generative models have recently achieved
significant accomplishments in generating realistic and diverse data. These
approaches define a forward diffusion process for transforming data into noise
and a backward denoising process for sampling data from noise. Unfortunately,
the generation process of current denoising diffusion models is notoriously
slow due to the lengthy iterative noise estimations, which rely on cumbersome
neural networks. It prevents the diffusion models from being widely deployed,
especially on edge devices. Previous works accelerate the generation process of
diffusion model (DM) via finding shorter yet effective sampling trajectories.
However, they overlook the cost of noise estimation with a heavy network in
every iteration. In this work, we accelerate generation from the perspective of
compressing the noise estimation network. Due to the difficulty of retraining
DMs, we exclude mainstream training-aware compression paradigms and introduce
post-training quantization (PTQ) into DM acceleration. However, the output
distributions of noise estimation networks change with time-step, making
previous PTQ methods fail in DMs since they are designed for single-time step
scenarios. To devise a DM-specific PTQ method, we explore PTQ on DM in three
aspects: quantized operations, calibration dataset, and calibration metric. We
summarize and use several observations derived from all-inclusive
investigations to formulate our method, which especially targets the unique
multi-time-step structure of DMs. Experimentally, our method can directly
quantize full-precision DMs into 8-bit models while maintaining or even
improving their performance in a training-free manner. Importantly, our method
can serve as a plug-and-play module on other fast-sampling methods, e.g., DDIM.
The code is available at https://github.com/42Shawn/PTQ4DM .
| no_new_dataset | 0.944842 |
2212.02400 | Mathilde Caron | Mathilde Caron, Neil Houlsby, Cordelia Schmid | Location-Aware Self-Supervised Transformers for Semantic Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Pixel-level labels are particularly expensive to acquire. Hence, pretraining
is a critical step to improve models on a task like semantic segmentation.
However, prominent algorithms for pretraining neural networks use image-level
objectives, e.g. image classification, image-text alignment a la CLIP, or
self-supervised contrastive learning. These objectives do not model spatial
information, which might be sub-optimal when finetuning on downstream tasks
with spatial reasoning. In this work, we pretrain network with a location-aware
(LOCA) self-supervised method which fosters the emergence of strong dense
features. Specifically, we use both a patch-level clustering scheme to mine
dense pseudo-labels and a relative location prediction task to encourage
learning about object parts and their spatial arrangements. Our experiments
show that LOCA pretraining leads to representations that transfer competitively
to challenging and diverse semantic segmentation datasets.
| [
{
"version": "v1",
"created": "Mon, 5 Dec 2022 16:24:29 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 20:08:12 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Caron",
"Mathilde",
""
],
[
"Houlsby",
"Neil",
""
],
[
"Schmid",
"Cordelia",
""
]
] | TITLE: Location-Aware Self-Supervised Transformers for Semantic Segmentation
ABSTRACT: Pixel-level labels are particularly expensive to acquire. Hence, pretraining
is a critical step to improve models on a task like semantic segmentation.
However, prominent algorithms for pretraining neural networks use image-level
objectives, e.g. image classification, image-text alignment a la CLIP, or
self-supervised contrastive learning. These objectives do not model spatial
information, which might be sub-optimal when finetuning on downstream tasks
with spatial reasoning. In this work, we pretrain network with a location-aware
(LOCA) self-supervised method which fosters the emergence of strong dense
features. Specifically, we use both a patch-level clustering scheme to mine
dense pseudo-labels and a relative location prediction task to encourage
learning about object parts and their spatial arrangements. Our experiments
show that LOCA pretraining leads to representations that transfer competitively
to challenging and diverse semantic segmentation datasets.
| no_new_dataset | 0.953101 |
2212.02932 | Alessandro Antonucci | Marco Zaffalon and Alessandro Antonucci and David Huber and Rafael
Caba\~nas | Learning to Bound Counterfactual Inference from Observational, Biased
and Randomised Data | null | null | null | null | cs.AI stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We address the problem of integrating data from multiple, possibly biased,
observational and interventional studies, to eventually compute counterfactuals
in structural causal models. We start from the case of a single observational
dataset affected by a selection bias. We show that the likelihood of the
available data has no local maxima. This enables us to use the causal
expectation-maximisation scheme to compute approximate bounds for partially
identifiable counterfactual queries, which are the focus of this paper. We then
show how the same approach can solve the general case of multiple datasets, no
matter whether interventional or observational, biased or unbiased, by
remapping it into the former one via graphical transformations. Systematic
numerical experiments and a case study on palliative care show the
effectiveness and accuracy of our approach, while hinting at the benefits of
integrating heterogeneous data to get informative bounds in case of partial
identifiability.
| [
{
"version": "v1",
"created": "Tue, 6 Dec 2022 12:42:11 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 14:41:18 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zaffalon",
"Marco",
""
],
[
"Antonucci",
"Alessandro",
""
],
[
"Huber",
"David",
""
],
[
"Cabañas",
"Rafael",
""
]
] | TITLE: Learning to Bound Counterfactual Inference from Observational, Biased
and Randomised Data
ABSTRACT: We address the problem of integrating data from multiple, possibly biased,
observational and interventional studies, to eventually compute counterfactuals
in structural causal models. We start from the case of a single observational
dataset affected by a selection bias. We show that the likelihood of the
available data has no local maxima. This enables us to use the causal
expectation-maximisation scheme to compute approximate bounds for partially
identifiable counterfactual queries, which are the focus of this paper. We then
show how the same approach can solve the general case of multiple datasets, no
matter whether interventional or observational, biased or unbiased, by
remapping it into the former one via graphical transformations. Systematic
numerical experiments and a case study on palliative care show the
effectiveness and accuracy of our approach, while hinting at the benefits of
integrating heterogeneous data to get informative bounds in case of partial
identifiability.
| no_new_dataset | 0.942188 |
2212.04267 | Mustafa Shukor | Mustafa Shukor, Nicolas Thome, Matthieu Cord | Vision and Structured-Language Pretraining for Cross-Modal Food
Retrieval | Code: https://github.com/mshukor/VLPCook | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-Language Pretraining (VLP) and Foundation models have been the go-to
recipe for achieving SoTA performance on general benchmarks. However,
leveraging these powerful techniques for more complex vision-language tasks,
such as cooking applications, with more structured input data, is still little
investigated. In this work, we propose to leverage these techniques for
structured-text based computational cuisine tasks. Our strategy, dubbed
VLPCook, first transforms existing image-text pairs to image and
structured-text pairs. This allows to pretrain our VLPCook model using VLP
objectives adapted to the strutured data of the resulting datasets, then
finetuning it on downstream computational cooking tasks. During finetuning, we
also enrich the visual encoder, leveraging pretrained foundation models (e.g.
CLIP) to provide local and global textual context. VLPCook outperforms current
SoTA by a significant margin (+3.3 Recall@1 absolute improvement) on the task
of Cross-Modal Food Retrieval on the large Recipe1M dataset. We conduct further
experiments on VLP to validate their importance, especially on the Recipe1M+
dataset. Finally, we validate the generalization of the approach to other tasks
(i.e, Food Recognition) and domains with structured text such as the Medical
domain on the ROCO dataset. The code is available here:
https://github.com/mshukor/VLPCook
| [
{
"version": "v1",
"created": "Thu, 8 Dec 2022 13:37:17 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 19:48:24 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Shukor",
"Mustafa",
""
],
[
"Thome",
"Nicolas",
""
],
[
"Cord",
"Matthieu",
""
]
] | TITLE: Vision and Structured-Language Pretraining for Cross-Modal Food
Retrieval
ABSTRACT: Vision-Language Pretraining (VLP) and Foundation models have been the go-to
recipe for achieving SoTA performance on general benchmarks. However,
leveraging these powerful techniques for more complex vision-language tasks,
such as cooking applications, with more structured input data, is still little
investigated. In this work, we propose to leverage these techniques for
structured-text based computational cuisine tasks. Our strategy, dubbed
VLPCook, first transforms existing image-text pairs to image and
structured-text pairs. This allows to pretrain our VLPCook model using VLP
objectives adapted to the strutured data of the resulting datasets, then
finetuning it on downstream computational cooking tasks. During finetuning, we
also enrich the visual encoder, leveraging pretrained foundation models (e.g.
CLIP) to provide local and global textual context. VLPCook outperforms current
SoTA by a significant margin (+3.3 Recall@1 absolute improvement) on the task
of Cross-Modal Food Retrieval on the large Recipe1M dataset. We conduct further
experiments on VLP to validate their importance, especially on the Recipe1M+
dataset. Finally, we validate the generalization of the approach to other tasks
(i.e, Food Recognition) and domains with structured text such as the Medical
domain on the ROCO dataset. The code is available here:
https://github.com/mshukor/VLPCook
| no_new_dataset | 0.948917 |
2212.07201 | Luis Scoccola | Luis Scoccola, Hitesh Gakhar, Johnathan Bush, Nikolas Schonsheck,
Tatum Rask, Ling Zhou, Jose A. Perea | Toroidal Coordinates: Decorrelating Circular Coordinates With Lattice
Reduction | 24 pages, 12 figures. To appear in proceedings of 39th International
Symposium on Computational Geometry | null | null | null | cs.CG cs.LG math.AT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The circular coordinates algorithm of de Silva, Morozov, and
Vejdemo-Johansson takes as input a dataset together with a cohomology class
representing a $1$-dimensional hole in the data; the output is a map from the
data into the circle that captures this hole, and that is of minimum energy in
a suitable sense. However, when applied to several cohomology classes, the
output circle-valued maps can be "geometrically correlated" even if the chosen
cohomology classes are linearly independent. It is shown in the original work
that less correlated maps can be obtained with suitable integer linear
combinations of the cohomology classes, with the linear combinations being
chosen by inspection. In this paper, we identify a formal notion of geometric
correlation between circle-valued maps which, in the Riemannian manifold case,
corresponds to the Dirichlet form, a bilinear form derived from the Dirichlet
energy. We describe a systematic procedure for constructing low energy
torus-valued maps on data, starting from a set of linearly independent
cohomology classes. We showcase our procedure with computational examples. Our
main algorithm is based on the Lenstra--Lenstra--Lov\'asz algorithm from
computational number theory.
| [
{
"version": "v1",
"created": "Wed, 14 Dec 2022 12:59:25 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 20:32:23 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Scoccola",
"Luis",
""
],
[
"Gakhar",
"Hitesh",
""
],
[
"Bush",
"Johnathan",
""
],
[
"Schonsheck",
"Nikolas",
""
],
[
"Rask",
"Tatum",
""
],
[
"Zhou",
"Ling",
""
],
[
"Perea",
"Jose A.",
""
]
] | TITLE: Toroidal Coordinates: Decorrelating Circular Coordinates With Lattice
Reduction
ABSTRACT: The circular coordinates algorithm of de Silva, Morozov, and
Vejdemo-Johansson takes as input a dataset together with a cohomology class
representing a $1$-dimensional hole in the data; the output is a map from the
data into the circle that captures this hole, and that is of minimum energy in
a suitable sense. However, when applied to several cohomology classes, the
output circle-valued maps can be "geometrically correlated" even if the chosen
cohomology classes are linearly independent. It is shown in the original work
that less correlated maps can be obtained with suitable integer linear
combinations of the cohomology classes, with the linear combinations being
chosen by inspection. In this paper, we identify a formal notion of geometric
correlation between circle-valued maps which, in the Riemannian manifold case,
corresponds to the Dirichlet form, a bilinear form derived from the Dirichlet
energy. We describe a systematic procedure for constructing low energy
torus-valued maps on data, starting from a set of linearly independent
cohomology classes. We showcase our procedure with computational examples. Our
main algorithm is based on the Lenstra--Lenstra--Lov\'asz algorithm from
computational number theory.
| no_new_dataset | 0.952574 |
2301.11705 | Hangdong Kuang | Kuang Hangdong and Mi Bo | FedPH: Privacy-enhanced Heterogeneous Federated Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Federated Learning is a distributed machine-learning environment that allows
clients to learn collaboratively without sharing private data. This is
accomplished by exchanging parameters. However, the differences in data
distributions and computing resources among clients make related studies
difficult. To address these heterogeneous problems, we propose a novel
Federated Learning method. Our method utilizes a pre-trained model as the
backbone of the local model, with fully connected layers comprising the head.
The backbone extracts features for the head, and the embedding vector of
classes is shared between clients to improve the head and enhance the
performance of the local model. By sharing the embedding vector of classes
instead of gradient-based parameters, clients can better adapt to private data,
and communication between the server and clients is more effective. To protect
privacy, we propose a privacy-preserving hybrid method that adds noise to the
embedding vector of classes. This method has a minimal effect on the
performance of the local model when differential privacy is met. We conduct a
comprehensive evaluation of our approach on a self-built vehicle dataset,
comparing it with other Federated Learning methods under non-independent
identically distributed(Non-IID).
| [
{
"version": "v1",
"created": "Fri, 27 Jan 2023 13:32:17 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 07:40:17 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Hangdong",
"Kuang",
""
],
[
"Bo",
"Mi",
""
]
] | TITLE: FedPH: Privacy-enhanced Heterogeneous Federated Learning
ABSTRACT: Federated Learning is a distributed machine-learning environment that allows
clients to learn collaboratively without sharing private data. This is
accomplished by exchanging parameters. However, the differences in data
distributions and computing resources among clients make related studies
difficult. To address these heterogeneous problems, we propose a novel
Federated Learning method. Our method utilizes a pre-trained model as the
backbone of the local model, with fully connected layers comprising the head.
The backbone extracts features for the head, and the embedding vector of
classes is shared between clients to improve the head and enhance the
performance of the local model. By sharing the embedding vector of classes
instead of gradient-based parameters, clients can better adapt to private data,
and communication between the server and clients is more effective. To protect
privacy, we propose a privacy-preserving hybrid method that adds noise to the
embedding vector of classes. This method has a minimal effect on the
performance of the local model when differential privacy is met. We conduct a
comprehensive evaluation of our approach on a self-built vehicle dataset,
comparing it with other Federated Learning methods under non-independent
identically distributed(Non-IID).
| new_dataset | 0.915242 |
2302.03507 | Chengcheng Han | Chengcheng Han, Yuhe Wang, Yingnan Fu, Xiang Li, Minghui Qiu, Ming
Gao, and Aoying Zhou | Meta-Learning Siamese Network for Few-Shot Text Classification | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-shot learning has been used to tackle the problem of label scarcity in
text classification, of which meta-learning based methods have shown to be
effective, such as the prototypical networks (PROTO). Despite the success of
PROTO, there still exist three main problems: (1) ignore the randomness of the
sampled support sets when computing prototype vectors; (2) disregard the
importance of labeled samples; (3) construct meta-tasks in a purely random
manner. In this paper, we propose a Meta-Learning Siamese Network, namely,
Meta-SN, to address these issues. Specifically, instead of computing prototype
vectors from the sampled support sets, Meta-SN utilizes external knowledge
(e.g. class names and descriptive texts) for class labels, which is encoded as
the low-dimensional embeddings of prototype vectors. In addition, Meta-SN
presents a novel sampling strategy for constructing meta-tasks, which gives
higher sampling probabilities to hard-to-classify samples. Extensive
experiments are conducted on six benchmark datasets to show the clear
superiority of Meta-SN over other state-of-the-art models. For reproducibility,
all the datasets and codes are provided at https://github.com/hccngu/Meta-SN.
| [
{
"version": "v1",
"created": "Sun, 5 Feb 2023 07:41:17 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 15:10:57 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Han",
"Chengcheng",
""
],
[
"Wang",
"Yuhe",
""
],
[
"Fu",
"Yingnan",
""
],
[
"Li",
"Xiang",
""
],
[
"Qiu",
"Minghui",
""
],
[
"Gao",
"Ming",
""
],
[
"Zhou",
"Aoying",
""
]
] | TITLE: Meta-Learning Siamese Network for Few-Shot Text Classification
ABSTRACT: Few-shot learning has been used to tackle the problem of label scarcity in
text classification, of which meta-learning based methods have shown to be
effective, such as the prototypical networks (PROTO). Despite the success of
PROTO, there still exist three main problems: (1) ignore the randomness of the
sampled support sets when computing prototype vectors; (2) disregard the
importance of labeled samples; (3) construct meta-tasks in a purely random
manner. In this paper, we propose a Meta-Learning Siamese Network, namely,
Meta-SN, to address these issues. Specifically, instead of computing prototype
vectors from the sampled support sets, Meta-SN utilizes external knowledge
(e.g. class names and descriptive texts) for class labels, which is encoded as
the low-dimensional embeddings of prototype vectors. In addition, Meta-SN
presents a novel sampling strategy for constructing meta-tasks, which gives
higher sampling probabilities to hard-to-classify samples. Extensive
experiments are conducted on six benchmark datasets to show the clear
superiority of Meta-SN over other state-of-the-art models. For reproducibility,
all the datasets and codes are provided at https://github.com/hccngu/Meta-SN.
| no_new_dataset | 0.951729 |
2302.04075 | Sudhanshu Mittal | Sudhanshu Mittal, Joshua Niemeijer, J\"org P. Sch\"afer, Thomas Brox | Best Practices in Active Learning for Semantic Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Active learning is particularly of interest for semantic segmentation, where
annotations are costly. Previous academic studies focused on datasets that are
already very diverse and where the model is trained in a supervised manner with
a large annotation budget. In contrast, data collected in many driving
scenarios is highly redundant, and most medical applications are subject to
very constrained annotation budgets. This work investigates the various types
of existing active learning methods for semantic segmentation under diverse
conditions across three dimensions - data distribution w.r.t. different
redundancy levels, integration of semi-supervised learning, and different
labeling budgets. We find that these three underlying factors are decisive for
the selection of the best active learning approach. As an outcome of our study,
we provide a comprehensive usage guide to obtain the best performance for each
case. We also propose an exemplary evaluation task for driving scenarios, where
data has high redundancy, to showcase the practical implications of our
research findings.
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2023 14:23:37 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 18:30:21 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Mittal",
"Sudhanshu",
""
],
[
"Niemeijer",
"Joshua",
""
],
[
"Schäfer",
"Jörg P.",
""
],
[
"Brox",
"Thomas",
""
]
] | TITLE: Best Practices in Active Learning for Semantic Segmentation
ABSTRACT: Active learning is particularly of interest for semantic segmentation, where
annotations are costly. Previous academic studies focused on datasets that are
already very diverse and where the model is trained in a supervised manner with
a large annotation budget. In contrast, data collected in many driving
scenarios is highly redundant, and most medical applications are subject to
very constrained annotation budgets. This work investigates the various types
of existing active learning methods for semantic segmentation under diverse
conditions across three dimensions - data distribution w.r.t. different
redundancy levels, integration of semi-supervised learning, and different
labeling budgets. We find that these three underlying factors are decisive for
the selection of the best active learning approach. As an outcome of our study,
we provide a comprehensive usage guide to obtain the best performance for each
case. We also propose an exemplary evaluation task for driving scenarios, where
data has high redundancy, to showcase the practical implications of our
research findings.
| no_new_dataset | 0.950915 |
2302.05017 | Bingnan Wang | Bingnan Wang, Fanjiang Xu and Quan Zheng | A survey on facial image deblurring | Accepted to computational visual media | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When a facial image is blurred, it significantly affects high-level vision
tasks such as face recognition. The purpose of facial image deblurring is to
recover a clear image from a blurry input image, which can improve the
recognition accuracy, etc. However, general deblurring methods do not perform
well on facial images. Therefore, some face deblurring methods have been
proposed to improve performance by adding semantic or structural information as
specific priors according to the characteristics of the facial images. In this
paper, we survey and summarize recently published methods for facial image
deblurring, most of which are based on deep learning. First, we provide a brief
introduction to the modeling of image blurring. Next, we summarize face
deblurring methods into two categories: model-based methods and deep
learning-based methods. Furthermore, we summarize the datasets, loss functions,
and performance evaluation metrics commonly used in the neural network training
process. We show the performance of classical methods on these datasets and
metrics and provide a brief discussion on the differences between model-based
and learning-based methods. Finally, we discuss the current challenges and
possible future research directions.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2023 02:24:56 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 11:03:09 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Wang",
"Bingnan",
""
],
[
"Xu",
"Fanjiang",
""
],
[
"Zheng",
"Quan",
""
]
] | TITLE: A survey on facial image deblurring
ABSTRACT: When a facial image is blurred, it significantly affects high-level vision
tasks such as face recognition. The purpose of facial image deblurring is to
recover a clear image from a blurry input image, which can improve the
recognition accuracy, etc. However, general deblurring methods do not perform
well on facial images. Therefore, some face deblurring methods have been
proposed to improve performance by adding semantic or structural information as
specific priors according to the characteristics of the facial images. In this
paper, we survey and summarize recently published methods for facial image
deblurring, most of which are based on deep learning. First, we provide a brief
introduction to the modeling of image blurring. Next, we summarize face
deblurring methods into two categories: model-based methods and deep
learning-based methods. Furthermore, we summarize the datasets, loss functions,
and performance evaluation metrics commonly used in the neural network training
process. We show the performance of classical methods on these datasets and
metrics and provide a brief discussion on the differences between model-based
and learning-based methods. Finally, we discuss the current challenges and
possible future research directions.
| no_new_dataset | 0.947527 |
2302.06885 | Zitao Liu | Jiahao Chen, Zitao Liu, Shuyan Huang, Qiongqiong Liu, Weiqi Luo | Improving Interpretability of Deep Sequential Knowledge Tracing Models
with Question-centric Cognitive Representations | Accepted at AAAI'23: The 37th AAAI Conference on Artificial
Intelligence, 2023 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Knowledge tracing (KT) is a crucial technique to predict students' future
performance by observing their historical learning processes. Due to the
powerful representation ability of deep neural networks, remarkable progress
has been made by using deep learning techniques to solve the KT problem. The
majority of existing approaches rely on the \emph{homogeneous question}
assumption that questions have equivalent contributions if they share the same
set of knowledge components. Unfortunately, this assumption is inaccurate in
real-world educational scenarios. Furthermore, it is very challenging to
interpret the prediction results from the existing deep learning based KT
models. Therefore, in this paper, we present QIKT, a question-centric
interpretable KT model to address the above challenges. The proposed QIKT
approach explicitly models students' knowledge state variations at a
fine-grained level with question-sensitive cognitive representations that are
jointly learned from a question-centric knowledge acquisition module and a
question-centric problem solving module. Meanwhile, the QIKT utilizes an item
response theory based prediction layer to generate interpretable prediction
results. The proposed QIKT model is evaluated on three public real-world
educational datasets. The results demonstrate that our approach is superior on
the KT prediction task, and it outperforms a wide range of deep learning based
KT models in terms of prediction accuracy with better model interpretability.
To encourage reproducible results, we have provided all the datasets and code
at \url{https://pykt.org/}.
| [
{
"version": "v1",
"created": "Tue, 14 Feb 2023 08:14:30 GMT"
},
{
"version": "v2",
"created": "Sun, 19 Feb 2023 09:46:47 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 06:46:14 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Chen",
"Jiahao",
""
],
[
"Liu",
"Zitao",
""
],
[
"Huang",
"Shuyan",
""
],
[
"Liu",
"Qiongqiong",
""
],
[
"Luo",
"Weiqi",
""
]
] | TITLE: Improving Interpretability of Deep Sequential Knowledge Tracing Models
with Question-centric Cognitive Representations
ABSTRACT: Knowledge tracing (KT) is a crucial technique to predict students' future
performance by observing their historical learning processes. Due to the
powerful representation ability of deep neural networks, remarkable progress
has been made by using deep learning techniques to solve the KT problem. The
majority of existing approaches rely on the \emph{homogeneous question}
assumption that questions have equivalent contributions if they share the same
set of knowledge components. Unfortunately, this assumption is inaccurate in
real-world educational scenarios. Furthermore, it is very challenging to
interpret the prediction results from the existing deep learning based KT
models. Therefore, in this paper, we present QIKT, a question-centric
interpretable KT model to address the above challenges. The proposed QIKT
approach explicitly models students' knowledge state variations at a
fine-grained level with question-sensitive cognitive representations that are
jointly learned from a question-centric knowledge acquisition module and a
question-centric problem solving module. Meanwhile, the QIKT utilizes an item
response theory based prediction layer to generate interpretable prediction
results. The proposed QIKT model is evaluated on three public real-world
educational datasets. The results demonstrate that our approach is superior on
the KT prediction task, and it outperforms a wide range of deep learning based
KT models in terms of prediction accuracy with better model interpretability.
To encourage reproducible results, we have provided all the datasets and code
at \url{https://pykt.org/}.
| no_new_dataset | 0.947186 |
2302.08287 | Yuhang Zhang | Yuhang Zhang, Weihong Deng, Liang Zheng | Unsupervised Evaluation of Out-of-distribution Detection: A Data-centric
Perspective | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Out-of-distribution (OOD) detection methods assume that they have test ground
truths, i.e., whether individual test samples are in-distribution (IND) or OOD.
However, in the real world, we do not always have such ground truths, and thus
do not know which sample is correctly detected and cannot compute the metric
like AUROC to evaluate the performance of different OOD detection methods. In
this paper, we are the first to introduce the unsupervised evaluation problem
in OOD detection, which aims to evaluate OOD detection methods in real-world
changing environments without OOD labels. We propose three methods to compute
Gscore as an unsupervised indicator of OOD detection performance. We further
introduce a new benchmark Gbench, which has 200 real-world OOD datasets of
various label spaces to train and evaluate our method. Through experiments, we
find a strong quantitative correlation betwwen Gscore and the OOD detection
performance. Extensive experiments demonstrate that our Gscore achieves
state-of-the-art performance. Gscore also generalizes well with different
IND/OOD datasets, OOD detection methods, backbones and dataset sizes. We
further provide interesting analyses of the effects of backbones and IND/OOD
datasets on OOD detection performance. The data and code will be available.
| [
{
"version": "v1",
"created": "Thu, 16 Feb 2023 13:34:35 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 01:32:53 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhang",
"Yuhang",
""
],
[
"Deng",
"Weihong",
""
],
[
"Zheng",
"Liang",
""
]
] | TITLE: Unsupervised Evaluation of Out-of-distribution Detection: A Data-centric
Perspective
ABSTRACT: Out-of-distribution (OOD) detection methods assume that they have test ground
truths, i.e., whether individual test samples are in-distribution (IND) or OOD.
However, in the real world, we do not always have such ground truths, and thus
do not know which sample is correctly detected and cannot compute the metric
like AUROC to evaluate the performance of different OOD detection methods. In
this paper, we are the first to introduce the unsupervised evaluation problem
in OOD detection, which aims to evaluate OOD detection methods in real-world
changing environments without OOD labels. We propose three methods to compute
Gscore as an unsupervised indicator of OOD detection performance. We further
introduce a new benchmark Gbench, which has 200 real-world OOD datasets of
various label spaces to train and evaluate our method. Through experiments, we
find a strong quantitative correlation betwwen Gscore and the OOD detection
performance. Extensive experiments demonstrate that our Gscore achieves
state-of-the-art performance. Gscore also generalizes well with different
IND/OOD datasets, OOD detection methods, backbones and dataset sizes. We
further provide interesting analyses of the effects of backbones and IND/OOD
datasets on OOD detection performance. The data and code will be available.
| new_dataset | 0.959421 |
2302.10276 | Likun Yang | Likun Yang, Xiaokun Feng, Xiaotang Chen, Shiyu Zhang, Kaiqi Huang | See Your Heart: Psychological states Interpretation through Visual
Creations | null | null | null | null | cs.CV cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In psychoanalysis, generating interpretations to one's psychological state
through visual creations is facing significant demands. The two main tasks of
existing studies in the field of computer vision, sentiment/emotion
classification and affective captioning, can hardly satisfy the requirement of
psychological interpreting. To meet the demands for psychoanalysis, we
introduce a challenging task, \textbf{V}isual \textbf{E}motion
\textbf{I}nterpretation \textbf{T}ask (VEIT). VEIT requires AI to generate
reasonable interpretations of creator's psychological state through visual
creations. To support the task, we present a multimodal dataset termed SpyIn
(\textbf{S}and\textbf{p}la\textbf{y} \textbf{In}terpretation Dataset), which is
psychological theory supported and professional annotated. Dataset analysis
illustrates that SpyIn is not only able to support VEIT, but also more
challenging compared with other captioning datasets. Building on SpyIn, we
conduct experiments of several image captioning method, and propose a
visual-semantic combined model which obtains a SOTA result on SpyIn. The
results indicate that VEIT is a more challenging task requiring scene graph
information and psychological knowledge. Our work also show a promise for AI to
analyze and explain inner world of humanity through visual creations.
| [
{
"version": "v1",
"created": "Sat, 11 Feb 2023 07:30:21 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 10:01:05 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Yang",
"Likun",
""
],
[
"Feng",
"Xiaokun",
""
],
[
"Chen",
"Xiaotang",
""
],
[
"Zhang",
"Shiyu",
""
],
[
"Huang",
"Kaiqi",
""
]
] | TITLE: See Your Heart: Psychological states Interpretation through Visual
Creations
ABSTRACT: In psychoanalysis, generating interpretations to one's psychological state
through visual creations is facing significant demands. The two main tasks of
existing studies in the field of computer vision, sentiment/emotion
classification and affective captioning, can hardly satisfy the requirement of
psychological interpreting. To meet the demands for psychoanalysis, we
introduce a challenging task, \textbf{V}isual \textbf{E}motion
\textbf{I}nterpretation \textbf{T}ask (VEIT). VEIT requires AI to generate
reasonable interpretations of creator's psychological state through visual
creations. To support the task, we present a multimodal dataset termed SpyIn
(\textbf{S}and\textbf{p}la\textbf{y} \textbf{In}terpretation Dataset), which is
psychological theory supported and professional annotated. Dataset analysis
illustrates that SpyIn is not only able to support VEIT, but also more
challenging compared with other captioning datasets. Building on SpyIn, we
conduct experiments of several image captioning method, and propose a
visual-semantic combined model which obtains a SOTA result on SpyIn. The
results indicate that VEIT is a more challenging task requiring scene graph
information and psychological knowledge. Our work also show a promise for AI to
analyze and explain inner world of humanity through visual creations.
| new_dataset | 0.964355 |
2302.11208 | Kaikai Zhao | Kaikai Zhao and Norimichi Ukita | KS-DETR: Knowledge Sharing in Attention Learning for Detection
Transformer | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Scaled dot-product attention applies a softmax function on the scaled
dot-product of queries and keys to calculate weights and then multiplies the
weights and values. In this work, we study how to improve the learning of
scaled dot-product attention to improve the accuracy of DETR. Our method is
based on the following observations: using ground truth foreground-background
mask (GT Fg-Bg Mask) as additional cues in the weights/values learning enables
learning much better weights/values; with better weights/values, better
values/weights can be learned. We propose a triple-attention module in which
the first attention is a plain scaled dot-product attention, the second/third
attention generates high-quality weights/values (with the assistance of GT
Fg-Bg Mask) and shares the values/weights with the first attention to improve
the quality of values/weights. The second and third attentions are removed
during inference. We call our method knowledge-sharing DETR (KS-DETR), which is
an extension of knowledge distillation (KD) in the way that the improved
weights and values of the teachers (the second and third attentions) are
directly shared, instead of mimicked, by the student (the first attention) to
enable more efficient knowledge transfer from the teachers to the student.
Experiments on various DETR-like methods show consistent improvements over the
baseline methods on the MS COCO benchmark. Code is available at
https://github.com/edocanonymous/KS-DETR.
| [
{
"version": "v1",
"created": "Wed, 22 Feb 2023 08:48:08 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 04:54:05 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhao",
"Kaikai",
""
],
[
"Ukita",
"Norimichi",
""
]
] | TITLE: KS-DETR: Knowledge Sharing in Attention Learning for Detection
Transformer
ABSTRACT: Scaled dot-product attention applies a softmax function on the scaled
dot-product of queries and keys to calculate weights and then multiplies the
weights and values. In this work, we study how to improve the learning of
scaled dot-product attention to improve the accuracy of DETR. Our method is
based on the following observations: using ground truth foreground-background
mask (GT Fg-Bg Mask) as additional cues in the weights/values learning enables
learning much better weights/values; with better weights/values, better
values/weights can be learned. We propose a triple-attention module in which
the first attention is a plain scaled dot-product attention, the second/third
attention generates high-quality weights/values (with the assistance of GT
Fg-Bg Mask) and shares the values/weights with the first attention to improve
the quality of values/weights. The second and third attentions are removed
during inference. We call our method knowledge-sharing DETR (KS-DETR), which is
an extension of knowledge distillation (KD) in the way that the improved
weights and values of the teachers (the second and third attentions) are
directly shared, instead of mimicked, by the student (the first attention) to
enable more efficient knowledge transfer from the teachers to the student.
Experiments on various DETR-like methods show consistent improvements over the
baseline methods on the MS COCO benchmark. Code is available at
https://github.com/edocanonymous/KS-DETR.
| no_new_dataset | 0.950595 |
2302.14237 | Zongyu Li | Kay Hutchinson, Zongyu Li, Ian Reyes, Homa Alemzadeh | Towards Surgical Context Inference and Translation to Gestures | accepted for the 2023 International Conference on Robotics and
Automation (ICRA) | null | null | null | cs.CV cs.LG cs.RO | http://creativecommons.org/licenses/by/4.0/ | Manual labeling of gestures in robot-assisted surgery is labor intensive,
prone to errors, and requires expertise or training. We propose a method for
automated and explainable generation of gesture transcripts that leverages the
abundance of data for image segmentation. Surgical context is detected using
segmentation masks by examining the distances and intersections between the
tools and objects. Next, context labels are translated into gesture transcripts
using knowledge-based Finite State Machine (FSM) and data-driven Long Short
Term Memory (LSTM) models. We evaluate the performance of each stage of our
method by comparing the results with the ground truth segmentation masks, the
consensus context labels, and the gesture labels in the JIGSAWS dataset. Our
results show that our segmentation models achieve state-of-the-art performance
in recognizing needle and thread in Suturing and we can automatically detect
important surgical states with high agreement with crowd-sourced labels (e.g.,
contact between graspers and objects in Suturing). We also find that the FSM
models are more robust to poor segmentation and labeling performance than
LSTMs. Our proposed method can significantly shorten the gesture labeling
process (~2.8 times).
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 01:39:36 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 19:07:36 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Hutchinson",
"Kay",
""
],
[
"Li",
"Zongyu",
""
],
[
"Reyes",
"Ian",
""
],
[
"Alemzadeh",
"Homa",
""
]
] | TITLE: Towards Surgical Context Inference and Translation to Gestures
ABSTRACT: Manual labeling of gestures in robot-assisted surgery is labor intensive,
prone to errors, and requires expertise or training. We propose a method for
automated and explainable generation of gesture transcripts that leverages the
abundance of data for image segmentation. Surgical context is detected using
segmentation masks by examining the distances and intersections between the
tools and objects. Next, context labels are translated into gesture transcripts
using knowledge-based Finite State Machine (FSM) and data-driven Long Short
Term Memory (LSTM) models. We evaluate the performance of each stage of our
method by comparing the results with the ground truth segmentation masks, the
consensus context labels, and the gesture labels in the JIGSAWS dataset. Our
results show that our segmentation models achieve state-of-the-art performance
in recognizing needle and thread in Suturing and we can automatically detect
important surgical states with high agreement with crowd-sourced labels (e.g.,
contact between graspers and objects in Suturing). We also find that the FSM
models are more robust to poor segmentation and labeling performance than
LSTMs. Our proposed method can significantly shorten the gesture labeling
process (~2.8 times).
| no_new_dataset | 0.951323 |
2303.01309 | Jiahong Zhang | Jiahong Zhang, Lihong Cao, Qiuxia Lai, Binyao Li, Yunxiao Qin | BIFRNet: A Brain-Inspired Feature Restoration DNN for Partially Occluded
Image Recognition | This paper has been accepted by AAAI-2023 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The partially occluded image recognition (POIR) problem has been a challenge
for artificial intelligence for a long time. A common strategy to handle the
POIR problem is using the non-occluded features for classification.
Unfortunately, this strategy will lose effectiveness when the image is severely
occluded, since the visible parts can only provide limited information. Several
studies in neuroscience reveal that feature restoration which fills in the
occluded information and is called amodal completion is essential for human
brains to recognize partially occluded images. However, feature restoration is
commonly ignored by CNNs, which may be the reason why CNNs are ineffective for
the POIR problem. Inspired by this, we propose a novel brain-inspired feature
restoration network (BIFRNet) to solve the POIR problem. It mimics a ventral
visual pathway to extract image features and a dorsal visual pathway to
distinguish occluded and visible image regions. In addition, it also uses a
knowledge module to store object prior knowledge and uses a completion module
to restore occluded features based on visible features and prior knowledge.
Thorough experiments on synthetic and real-world occluded image datasets show
that BIFRNet outperforms the existing methods in solving the POIR problem.
Especially for severely occluded images, BIRFRNet surpasses other methods by a
large margin and is close to the human brain performance. Furthermore, the
brain-inspired design makes BIFRNet more interpretable.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 14:33:43 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 02:19:51 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhang",
"Jiahong",
""
],
[
"Cao",
"Lihong",
""
],
[
"Lai",
"Qiuxia",
""
],
[
"Li",
"Binyao",
""
],
[
"Qin",
"Yunxiao",
""
]
] | TITLE: BIFRNet: A Brain-Inspired Feature Restoration DNN for Partially Occluded
Image Recognition
ABSTRACT: The partially occluded image recognition (POIR) problem has been a challenge
for artificial intelligence for a long time. A common strategy to handle the
POIR problem is using the non-occluded features for classification.
Unfortunately, this strategy will lose effectiveness when the image is severely
occluded, since the visible parts can only provide limited information. Several
studies in neuroscience reveal that feature restoration which fills in the
occluded information and is called amodal completion is essential for human
brains to recognize partially occluded images. However, feature restoration is
commonly ignored by CNNs, which may be the reason why CNNs are ineffective for
the POIR problem. Inspired by this, we propose a novel brain-inspired feature
restoration network (BIFRNet) to solve the POIR problem. It mimics a ventral
visual pathway to extract image features and a dorsal visual pathway to
distinguish occluded and visible image regions. In addition, it also uses a
knowledge module to store object prior knowledge and uses a completion module
to restore occluded features based on visible features and prior knowledge.
Thorough experiments on synthetic and real-world occluded image datasets show
that BIFRNet outperforms the existing methods in solving the POIR problem.
Especially for severely occluded images, BIRFRNet surpasses other methods by a
large margin and is close to the human brain performance. Furthermore, the
brain-inspired design makes BIFRNet more interpretable.
| no_new_dataset | 0.946843 |
2303.02455 | Suhang Ye | Suhang Ye, Yingyi Zhang, Jie Hu, Liujuan Cao, Shengchuan Zhang, Lei
Shen, Jun Wang, Shouhong Ding, Rongrong Ji | DistilPose: Tokenized Pose Regression with Heatmap Distillation | accepted by CVPR2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the field of human pose estimation, regression-based methods have been
dominated in terms of speed, while heatmap-based methods are far ahead in terms
of performance. How to take advantage of both schemes remains a challenging
problem. In this paper, we propose a novel human pose estimation framework
termed DistilPose, which bridges the gaps between heatmap-based and
regression-based methods. Specifically, DistilPose maximizes the transfer of
knowledge from the teacher model (heatmap-based) to the student model
(regression-based) through Token-distilling Encoder (TDE) and Simulated
Heatmaps. TDE aligns the feature spaces of heatmap-based and regression-based
models by introducing tokenization, while Simulated Heatmaps transfer explicit
guidance (distribution and confidence) from teacher heatmaps into student
models. Extensive experiments show that the proposed DistilPose can
significantly improve the performance of the regression-based models while
maintaining efficiency. Specifically, on the MSCOCO validation dataset,
DistilPose-S obtains 71.6% mAP with 5.36M parameter, 2.38 GFLOPs and 40.2 FPS,
which saves 12.95x, 7.16x computational cost and is 4.9x faster than its
teacher model with only 0.9 points performance drop. Furthermore, DistilPose-L
obtains 74.4% mAP on MSCOCO validation dataset, achieving a new
state-of-the-art among predominant regression-based models.
| [
{
"version": "v1",
"created": "Sat, 4 Mar 2023 16:56:29 GMT"
},
{
"version": "v2",
"created": "Wed, 8 Mar 2023 05:44:22 GMT"
},
{
"version": "v3",
"created": "Thu, 16 Mar 2023 11:17:20 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Ye",
"Suhang",
""
],
[
"Zhang",
"Yingyi",
""
],
[
"Hu",
"Jie",
""
],
[
"Cao",
"Liujuan",
""
],
[
"Zhang",
"Shengchuan",
""
],
[
"Shen",
"Lei",
""
],
[
"Wang",
"Jun",
""
],
[
"Ding",
"Shouhong",
""
],
[
"Ji",
"Rongrong",
""
]
] | TITLE: DistilPose: Tokenized Pose Regression with Heatmap Distillation
ABSTRACT: In the field of human pose estimation, regression-based methods have been
dominated in terms of speed, while heatmap-based methods are far ahead in terms
of performance. How to take advantage of both schemes remains a challenging
problem. In this paper, we propose a novel human pose estimation framework
termed DistilPose, which bridges the gaps between heatmap-based and
regression-based methods. Specifically, DistilPose maximizes the transfer of
knowledge from the teacher model (heatmap-based) to the student model
(regression-based) through Token-distilling Encoder (TDE) and Simulated
Heatmaps. TDE aligns the feature spaces of heatmap-based and regression-based
models by introducing tokenization, while Simulated Heatmaps transfer explicit
guidance (distribution and confidence) from teacher heatmaps into student
models. Extensive experiments show that the proposed DistilPose can
significantly improve the performance of the regression-based models while
maintaining efficiency. Specifically, on the MSCOCO validation dataset,
DistilPose-S obtains 71.6% mAP with 5.36M parameter, 2.38 GFLOPs and 40.2 FPS,
which saves 12.95x, 7.16x computational cost and is 4.9x faster than its
teacher model with only 0.9 points performance drop. Furthermore, DistilPose-L
obtains 74.4% mAP on MSCOCO validation dataset, achieving a new
state-of-the-art among predominant regression-based models.
| no_new_dataset | 0.950134 |
2303.03684 | Mingzhen Sun | Mingzhen Sun, Weining Wang, Xinxin Zhu and Jing Liu | MOSO: Decomposing MOtion, Scene and Object for Video Prediction | Accepted by CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Motion, scene and object are three primary visual components of a video. In
particular, objects represent the foreground, scenes represent the background,
and motion traces their dynamics. Based on this insight, we propose a two-stage
MOtion, Scene and Object decomposition framework (MOSO) for video prediction,
consisting of MOSO-VQVAE and MOSO-Transformer. In the first stage, MOSO-VQVAE
decomposes a previous video clip into the motion, scene and object components,
and represents them as distinct groups of discrete tokens. Then, in the second
stage, MOSO-Transformer predicts the object and scene tokens of the subsequent
video clip based on the previous tokens and adds dynamic motion at the token
level to the generated object and scene tokens. Our framework can be easily
extended to unconditional video generation and video frame interpolation tasks.
Experimental results demonstrate that our method achieves new state-of-the-art
performance on five challenging benchmarks for video prediction and
unconditional video generation: BAIR, RoboNet, KTH, KITTI and UCF101. In
addition, MOSO can produce realistic videos by combining objects and scenes
from different videos.
| [
{
"version": "v1",
"created": "Tue, 7 Mar 2023 06:54:48 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 08:41:44 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Sun",
"Mingzhen",
""
],
[
"Wang",
"Weining",
""
],
[
"Zhu",
"Xinxin",
""
],
[
"Liu",
"Jing",
""
]
] | TITLE: MOSO: Decomposing MOtion, Scene and Object for Video Prediction
ABSTRACT: Motion, scene and object are three primary visual components of a video. In
particular, objects represent the foreground, scenes represent the background,
and motion traces their dynamics. Based on this insight, we propose a two-stage
MOtion, Scene and Object decomposition framework (MOSO) for video prediction,
consisting of MOSO-VQVAE and MOSO-Transformer. In the first stage, MOSO-VQVAE
decomposes a previous video clip into the motion, scene and object components,
and represents them as distinct groups of discrete tokens. Then, in the second
stage, MOSO-Transformer predicts the object and scene tokens of the subsequent
video clip based on the previous tokens and adds dynamic motion at the token
level to the generated object and scene tokens. Our framework can be easily
extended to unconditional video generation and video frame interpolation tasks.
Experimental results demonstrate that our method achieves new state-of-the-art
performance on five challenging benchmarks for video prediction and
unconditional video generation: BAIR, RoboNet, KTH, KITTI and UCF101. In
addition, MOSO can produce realistic videos by combining objects and scenes
from different videos.
| no_new_dataset | 0.951051 |
2303.04748 | Junbo Zhang | Junbo Zhang, Runpei Dong, Kaisheng Ma | CLIP-FO3D: Learning Free Open-world 3D Scene Representations from 2D
Dense CLIP | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training a 3D scene understanding model requires complicated human
annotations, which are laborious to collect and result in a model only encoding
close-set object semantics. In contrast, vision-language pre-training models
(e.g., CLIP) have shown remarkable open-world reasoning properties. To this
end, we propose directly transferring CLIP's feature space to 3D scene
understanding model without any form of supervision. We first modify CLIP's
input and forwarding process so that it can be adapted to extract dense pixel
features for 3D scene contents. We then project multi-view image features to
the point cloud and train a 3D scene understanding model with feature
distillation. Without any annotations or additional training, our model
achieves promising annotation-free semantic segmentation results on
open-vocabulary semantics and long-tailed concepts. Besides, serving as a
cross-modal pre-training framework, our method can be used to improve data
efficiency during fine-tuning. Our model outperforms previous SOTA methods in
various zero-shot and data-efficient learning benchmarks. Most importantly, our
model successfully inherits CLIP's rich-structured knowledge, allowing 3D scene
understanding models to recognize not only object concepts but also open-world
semantics.
| [
{
"version": "v1",
"created": "Wed, 8 Mar 2023 17:30:58 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 04:52:06 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhang",
"Junbo",
""
],
[
"Dong",
"Runpei",
""
],
[
"Ma",
"Kaisheng",
""
]
] | TITLE: CLIP-FO3D: Learning Free Open-world 3D Scene Representations from 2D
Dense CLIP
ABSTRACT: Training a 3D scene understanding model requires complicated human
annotations, which are laborious to collect and result in a model only encoding
close-set object semantics. In contrast, vision-language pre-training models
(e.g., CLIP) have shown remarkable open-world reasoning properties. To this
end, we propose directly transferring CLIP's feature space to 3D scene
understanding model without any form of supervision. We first modify CLIP's
input and forwarding process so that it can be adapted to extract dense pixel
features for 3D scene contents. We then project multi-view image features to
the point cloud and train a 3D scene understanding model with feature
distillation. Without any annotations or additional training, our model
achieves promising annotation-free semantic segmentation results on
open-vocabulary semantics and long-tailed concepts. Besides, serving as a
cross-modal pre-training framework, our method can be used to improve data
efficiency during fine-tuning. Our model outperforms previous SOTA methods in
various zero-shot and data-efficient learning benchmarks. Most importantly, our
model successfully inherits CLIP's rich-structured knowledge, allowing 3D scene
understanding models to recognize not only object concepts but also open-world
semantics.
| no_new_dataset | 0.944995 |
2303.04910 | Emily First | Emily First, Markus N. Rabe, Talia Ringer, Yuriy Brun | Baldur: Whole-Proof Generation and Repair with Large Language Models | null | null | null | null | cs.LG cs.LO cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Formally verifying software properties is a highly desirable but
labor-intensive task. Recent work has developed methods to automate formal
verification using proof assistants, such as Coq and Isabelle/HOL, e.g., by
training a model to predict one proof step at a time, and using that model to
search through the space of possible proofs. This paper introduces a new method
to automate formal verification: We use large language models, trained on
natural language text and code and fine-tuned on proofs, to generate whole
proofs for theorems at once, rather than one step at a time. We combine this
proof generation model with a fine-tuned repair model to repair generated
proofs, further increasing proving power. As its main contributions, this paper
demonstrates for the first time that: (1) Whole-proof generation using
transformers is possible and is as effective as search-based techniques without
requiring costly search. (2) Giving the learned model additional context, such
as a prior failed proof attempt and the ensuing error message, results in proof
repair and further improves automated proof generation. (3) We establish a new
state of the art for fully automated proof synthesis. We reify our method in a
prototype, Baldur, and evaluate it on a benchmark of 6,336 Isabelle/HOL
theorems and their proofs. In addition to empirically showing the effectiveness
of whole-proof generation, repair, and added context, we show that Baldur
improves on the state-of-the-art tool, Thor, by automatically generating proofs
for an additional 8.7% of the theorems. Together, Baldur and Thor can prove
65.7% of the theorems fully automatically. This paper paves the way for new
research into using large language models for automating formal verification.
| [
{
"version": "v1",
"created": "Wed, 8 Mar 2023 22:00:15 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 01:09:08 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"First",
"Emily",
""
],
[
"Rabe",
"Markus N.",
""
],
[
"Ringer",
"Talia",
""
],
[
"Brun",
"Yuriy",
""
]
] | TITLE: Baldur: Whole-Proof Generation and Repair with Large Language Models
ABSTRACT: Formally verifying software properties is a highly desirable but
labor-intensive task. Recent work has developed methods to automate formal
verification using proof assistants, such as Coq and Isabelle/HOL, e.g., by
training a model to predict one proof step at a time, and using that model to
search through the space of possible proofs. This paper introduces a new method
to automate formal verification: We use large language models, trained on
natural language text and code and fine-tuned on proofs, to generate whole
proofs for theorems at once, rather than one step at a time. We combine this
proof generation model with a fine-tuned repair model to repair generated
proofs, further increasing proving power. As its main contributions, this paper
demonstrates for the first time that: (1) Whole-proof generation using
transformers is possible and is as effective as search-based techniques without
requiring costly search. (2) Giving the learned model additional context, such
as a prior failed proof attempt and the ensuing error message, results in proof
repair and further improves automated proof generation. (3) We establish a new
state of the art for fully automated proof synthesis. We reify our method in a
prototype, Baldur, and evaluate it on a benchmark of 6,336 Isabelle/HOL
theorems and their proofs. In addition to empirically showing the effectiveness
of whole-proof generation, repair, and added context, we show that Baldur
improves on the state-of-the-art tool, Thor, by automatically generating proofs
for an additional 8.7% of the theorems. Together, Baldur and Thor can prove
65.7% of the theorems fully automatically. This paper paves the way for new
research into using large language models for automating formal verification.
| no_new_dataset | 0.951774 |
2303.06832 | Seungpyo Choi | SP Choi, Jihun Lee, Hyeongseok Ahn, Sanghee Jung, Bumsoo Kang | ODIN: On-demand Data Formulation to Mitigate Dataset Lock-in | 15 pages | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | ODIN is an innovative approach that addresses the problem of dataset
constraints by integrating generative AI models. Traditional zero-shot learning
methods are constrained by the training dataset. To fundamentally overcome this
limitation, ODIN attempts to mitigate the dataset constraints by generating
on-demand datasets based on user requirements. ODIN consists of three main
modules: a prompt generator, a text-to-image generator, and an image
post-processor. To generate high-quality prompts and images, we adopted a large
language model (e.g., ChatGPT), and a text-to-image diffusion model (e.g.,
Stable Diffusion), respectively. We evaluated ODIN on various datasets in terms
of model accuracy and data diversity to demonstrate its potential, and
conducted post-experiments for further investigation. Overall, ODIN is a
feasible approach that enables Al to learn unseen knowledge beyond the training
dataset.
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 03:28:36 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 05:44:50 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Choi",
"SP",
""
],
[
"Lee",
"Jihun",
""
],
[
"Ahn",
"Hyeongseok",
""
],
[
"Jung",
"Sanghee",
""
],
[
"Kang",
"Bumsoo",
""
]
] | TITLE: ODIN: On-demand Data Formulation to Mitigate Dataset Lock-in
ABSTRACT: ODIN is an innovative approach that addresses the problem of dataset
constraints by integrating generative AI models. Traditional zero-shot learning
methods are constrained by the training dataset. To fundamentally overcome this
limitation, ODIN attempts to mitigate the dataset constraints by generating
on-demand datasets based on user requirements. ODIN consists of three main
modules: a prompt generator, a text-to-image generator, and an image
post-processor. To generate high-quality prompts and images, we adopted a large
language model (e.g., ChatGPT), and a text-to-image diffusion model (e.g.,
Stable Diffusion), respectively. We evaluated ODIN on various datasets in terms
of model accuracy and data diversity to demonstrate its potential, and
conducted post-experiments for further investigation. Overall, ODIN is a
feasible approach that enables Al to learn unseen knowledge beyond the training
dataset.
| no_new_dataset | 0.947769 |
2303.07347 | Dingfeng Shi | Dingfeng Shi, Yujie Zhong, Qiong Cao, Lin Ma, Jia Li, Dacheng Tao | TriDet: Temporal Action Detection with Relative Boundary Modeling | CVPR2023; Temporal Action Detection; Temporal Action Localization | null | null | null | cs.CV cs.AI cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a one-stage framework TriDet for temporal action
detection. Existing methods often suffer from imprecise boundary predictions
due to the ambiguous action boundaries in videos. To alleviate this problem, we
propose a novel Trident-head to model the action boundary via an estimated
relative probability distribution around the boundary. In the feature pyramid
of TriDet, we propose an efficient Scalable-Granularity Perception (SGP) layer
to mitigate the rank loss problem of self-attention that takes place in the
video features and aggregate information across different temporal
granularities. Benefiting from the Trident-head and the SGP-based feature
pyramid, TriDet achieves state-of-the-art performance on three challenging
benchmarks: THUMOS14, HACS and EPIC-KITCHEN 100, with lower computational
costs, compared to previous methods. For example, TriDet hits an average mAP of
$69.3\%$ on THUMOS14, outperforming the previous best by $2.5\%$, but with only
$74.6\%$ of its latency. The code is released to
https://github.com/sssste/TriDet.
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 17:59:59 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 11:26:39 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Shi",
"Dingfeng",
""
],
[
"Zhong",
"Yujie",
""
],
[
"Cao",
"Qiong",
""
],
[
"Ma",
"Lin",
""
],
[
"Li",
"Jia",
""
],
[
"Tao",
"Dacheng",
""
]
] | TITLE: TriDet: Temporal Action Detection with Relative Boundary Modeling
ABSTRACT: In this paper, we present a one-stage framework TriDet for temporal action
detection. Existing methods often suffer from imprecise boundary predictions
due to the ambiguous action boundaries in videos. To alleviate this problem, we
propose a novel Trident-head to model the action boundary via an estimated
relative probability distribution around the boundary. In the feature pyramid
of TriDet, we propose an efficient Scalable-Granularity Perception (SGP) layer
to mitigate the rank loss problem of self-attention that takes place in the
video features and aggregate information across different temporal
granularities. Benefiting from the Trident-head and the SGP-based feature
pyramid, TriDet achieves state-of-the-art performance on three challenging
benchmarks: THUMOS14, HACS and EPIC-KITCHEN 100, with lower computational
costs, compared to previous methods. For example, TriDet hits an average mAP of
$69.3\%$ on THUMOS14, outperforming the previous best by $2.5\%$, but with only
$74.6\%$ of its latency. The code is released to
https://github.com/sssste/TriDet.
| no_new_dataset | 0.94743 |
2303.07533 | Subhashini Venugopalan | Subhashini Venugopalan, Jimmy Tobin, Samuel J. Yang, Katie Seaver,
Richard J.N. Cave, Pan-Pan Jiang, Neil Zeghidour, Rus Heywood, Jordan Green,
Michael P. Brenner | Speech Intelligibility Classifiers from 550k Disordered Speech Samples | ICASSP 2023 camera-ready | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We developed dysarthric speech intelligibility classifiers on 551,176
disordered speech samples contributed by a diverse set of 468 speakers, with a
range of self-reported speaking disorders and rated for their overall
intelligibility on a five-point scale. We trained three models following
different deep learning approaches and evaluated them on ~94K utterances from
100 speakers. We further found the models to generalize well (without further
training) on the TORGO database (100% accuracy), UASpeech (0.93 correlation),
ALS-TDI PMP (0.81 AUC) datasets as well as on a dataset of realistic unprompted
speech we gathered (106 dysarthric and 76 control speakers,~2300 samples).
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 23:38:56 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Mar 2023 22:54:23 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Venugopalan",
"Subhashini",
""
],
[
"Tobin",
"Jimmy",
""
],
[
"Yang",
"Samuel J.",
""
],
[
"Seaver",
"Katie",
""
],
[
"Cave",
"Richard J. N.",
""
],
[
"Jiang",
"Pan-Pan",
""
],
[
"Zeghidour",
"Neil",
""
],
[
"Heywood",
"Rus",
""
],
[
"Green",
"Jordan",
""
],
[
"Brenner",
"Michael P.",
""
]
] | TITLE: Speech Intelligibility Classifiers from 550k Disordered Speech Samples
ABSTRACT: We developed dysarthric speech intelligibility classifiers on 551,176
disordered speech samples contributed by a diverse set of 468 speakers, with a
range of self-reported speaking disorders and rated for their overall
intelligibility on a five-point scale. We trained three models following
different deep learning approaches and evaluated them on ~94K utterances from
100 speakers. We further found the models to generalize well (without further
training) on the TORGO database (100% accuracy), UASpeech (0.93 correlation),
ALS-TDI PMP (0.81 AUC) datasets as well as on a dataset of realistic unprompted
speech we gathered (106 dysarthric and 76 control speakers,~2300 samples).
| new_dataset | 0.955152 |
2303.07653 | Yunfan Ye | Yunfan Ye, Renjiao Yi, Zhirui Gao, Chenyang Zhu, Zhiping Cai, Kai Xu | NEF: Neural Edge Fields for 3D Parametric Curve Reconstruction from
Multi-view Images | CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the problem of reconstructing 3D feature curves of an object from a
set of calibrated multi-view images. To do so, we learn a neural implicit field
representing the density distribution of 3D edges which we refer to as Neural
Edge Field (NEF). Inspired by NeRF, NEF is optimized with a view-based
rendering loss where a 2D edge map is rendered at a given view and is compared
to the ground-truth edge map extracted from the image of that view. The
rendering-based differentiable optimization of NEF fully exploits 2D edge
detection, without needing a supervision of 3D edges, a 3D geometric operator
or cross-view edge correspondence. Several technical designs are devised to
ensure learning a range-limited and view-independent NEF for robust edge
extraction. The final parametric 3D curves are extracted from NEF with an
iterative optimization method. On our benchmark with synthetic data, we
demonstrate that NEF outperforms existing state-of-the-art methods on all
metrics. Project page: https://yunfan1202.github.io/NEF/.
| [
{
"version": "v1",
"created": "Tue, 14 Mar 2023 06:45:13 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 12:22:50 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Ye",
"Yunfan",
""
],
[
"Yi",
"Renjiao",
""
],
[
"Gao",
"Zhirui",
""
],
[
"Zhu",
"Chenyang",
""
],
[
"Cai",
"Zhiping",
""
],
[
"Xu",
"Kai",
""
]
] | TITLE: NEF: Neural Edge Fields for 3D Parametric Curve Reconstruction from
Multi-view Images
ABSTRACT: We study the problem of reconstructing 3D feature curves of an object from a
set of calibrated multi-view images. To do so, we learn a neural implicit field
representing the density distribution of 3D edges which we refer to as Neural
Edge Field (NEF). Inspired by NeRF, NEF is optimized with a view-based
rendering loss where a 2D edge map is rendered at a given view and is compared
to the ground-truth edge map extracted from the image of that view. The
rendering-based differentiable optimization of NEF fully exploits 2D edge
detection, without needing a supervision of 3D edges, a 3D geometric operator
or cross-view edge correspondence. Several technical designs are devised to
ensure learning a range-limited and view-independent NEF for robust edge
extraction. The final parametric 3D curves are extracted from NEF with an
iterative optimization method. On our benchmark with synthetic data, we
demonstrate that NEF outperforms existing state-of-the-art methods on all
metrics. Project page: https://yunfan1202.github.io/NEF/.
| no_new_dataset | 0.943034 |
2303.07863 | Xiang Fang | Xiang Fang, Daizong Liu, Pan Zhou, Guoshun Nan | You Can Ground Earlier than See: An Effective and Efficient Pipeline for
Temporal Sentence Grounding in Compressed Videos | Accepted by CVPR-23 | null | null | null | cs.CV cs.AI cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given an untrimmed video, temporal sentence grounding (TSG) aims to locate a
target moment semantically according to a sentence query. Although previous
respectable works have made decent success, they only focus on high-level
visual features extracted from the consecutive decoded frames and fail to
handle the compressed videos for query modelling, suffering from insufficient
representation capability and significant computational complexity during
training and testing. In this paper, we pose a new setting, compressed-domain
TSG, which directly utilizes compressed videos rather than fully-decompressed
frames as the visual input. To handle the raw video bit-stream input, we
propose a novel Three-branch Compressed-domain Spatial-temporal Fusion (TCSF)
framework, which extracts and aggregates three kinds of low-level visual
features (I-frame, motion vector and residual features) for effective and
efficient grounding. Particularly, instead of encoding the whole decoded frames
like previous works, we capture the appearance representation by only learning
the I-frame feature to reduce delay or latency. Besides, we explore the motion
information not only by learning the motion vector feature, but also by
exploring the relations of neighboring frames via the residual feature. In this
way, a three-branch spatial-temporal attention layer with an adaptive
motion-appearance fusion module is further designed to extract and aggregate
both appearance and motion information for the final grounding. Experiments on
three challenging datasets shows that our TCSF achieves better performance than
other state-of-the-art methods with lower complexity.
| [
{
"version": "v1",
"created": "Tue, 14 Mar 2023 12:53:27 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 08:34:27 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Fang",
"Xiang",
""
],
[
"Liu",
"Daizong",
""
],
[
"Zhou",
"Pan",
""
],
[
"Nan",
"Guoshun",
""
]
] | TITLE: You Can Ground Earlier than See: An Effective and Efficient Pipeline for
Temporal Sentence Grounding in Compressed Videos
ABSTRACT: Given an untrimmed video, temporal sentence grounding (TSG) aims to locate a
target moment semantically according to a sentence query. Although previous
respectable works have made decent success, they only focus on high-level
visual features extracted from the consecutive decoded frames and fail to
handle the compressed videos for query modelling, suffering from insufficient
representation capability and significant computational complexity during
training and testing. In this paper, we pose a new setting, compressed-domain
TSG, which directly utilizes compressed videos rather than fully-decompressed
frames as the visual input. To handle the raw video bit-stream input, we
propose a novel Three-branch Compressed-domain Spatial-temporal Fusion (TCSF)
framework, which extracts and aggregates three kinds of low-level visual
features (I-frame, motion vector and residual features) for effective and
efficient grounding. Particularly, instead of encoding the whole decoded frames
like previous works, we capture the appearance representation by only learning
the I-frame feature to reduce delay or latency. Besides, we explore the motion
information not only by learning the motion vector feature, but also by
exploring the relations of neighboring frames via the residual feature. In this
way, a three-branch spatial-temporal attention layer with an adaptive
motion-appearance fusion module is further designed to extract and aggregate
both appearance and motion information for the final grounding. Experiments on
three challenging datasets shows that our TCSF achieves better performance than
other state-of-the-art methods with lower complexity.
| no_new_dataset | 0.947478 |
2303.07896 | Erik Ostrowski | Erik Ostrowski and Bharath Srinivas Prabakaran and Muhammad Shafique | Exploring Weakly Supervised Semantic Segmentation Ensembles for Medical
Imaging Systems | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Reliable classification and detection of certain medical conditions, in
images, with state-of-the-art semantic segmentation networks, require vast
amounts of pixel-wise annotation. However, the public availability of such
datasets is minimal. Therefore, semantic segmentation with image-level labels
presents a promising alternative to this problem. Nevertheless, very few works
have focused on evaluating this technique and its applicability to the medical
sector. Due to their complexity and the small number of training examples in
medical datasets, classifier-based weakly supervised networks like class
activation maps (CAMs) struggle to extract useful information from them.
However, most state-of-the-art approaches rely on them to achieve their
improvements. Therefore, we propose a framework that can still utilize the
low-quality CAM predictions of complicated datasets to improve the accuracy of
our results. Our framework achieves that by first utilizing lower threshold
CAMs to cover the target object with high certainty; second, by combining
multiple low-threshold CAMs that even out their errors while highlighting the
target object. We performed exhaustive experiments on the popular multi-modal
BRATS and prostate DECATHLON segmentation challenge datasets. Using the
proposed framework, we have demonstrated an improved dice score of up to 8% on
BRATS and 6% on DECATHLON datasets compared to the previous state-of-the-art.
| [
{
"version": "v1",
"created": "Tue, 14 Mar 2023 13:31:05 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 08:09:58 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Ostrowski",
"Erik",
""
],
[
"Prabakaran",
"Bharath Srinivas",
""
],
[
"Shafique",
"Muhammad",
""
]
] | TITLE: Exploring Weakly Supervised Semantic Segmentation Ensembles for Medical
Imaging Systems
ABSTRACT: Reliable classification and detection of certain medical conditions, in
images, with state-of-the-art semantic segmentation networks, require vast
amounts of pixel-wise annotation. However, the public availability of such
datasets is minimal. Therefore, semantic segmentation with image-level labels
presents a promising alternative to this problem. Nevertheless, very few works
have focused on evaluating this technique and its applicability to the medical
sector. Due to their complexity and the small number of training examples in
medical datasets, classifier-based weakly supervised networks like class
activation maps (CAMs) struggle to extract useful information from them.
However, most state-of-the-art approaches rely on them to achieve their
improvements. Therefore, we propose a framework that can still utilize the
low-quality CAM predictions of complicated datasets to improve the accuracy of
our results. Our framework achieves that by first utilizing lower threshold
CAMs to cover the target object with high certainty; second, by combining
multiple low-threshold CAMs that even out their errors while highlighting the
target object. We performed exhaustive experiments on the popular multi-modal
BRATS and prostate DECATHLON segmentation challenge datasets. Using the
proposed framework, we have demonstrated an improved dice score of up to 8% on
BRATS and 6% on DECATHLON datasets compared to the previous state-of-the-art.
| no_new_dataset | 0.950778 |
2303.08335 | Qianqian Xie | Qianqian Xie, Jiayu Zhou, Yifan Peng, Fei Wang | FactReranker: Fact-guided Reranker for Faithful Radiology Report
Summarization | 11 pages, 5 figures, 7 tables, submitted to KDD 2023 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Automatic radiology report summarization is a crucial clinical task, whose
key challenge is to maintain factual accuracy between produced summaries and
ground truth radiology findings. Existing research adopts reinforcement
learning to directly optimize factual consistency metrics such as CheXBert or
RadGraph score. However, their decoding method using greedy search or beam
search considers no factual consistency when picking the optimal candidate,
leading to limited factual consistency improvement. To address it, we propose a
novel second-stage summarizing approach FactReranker, the first attempt that
learns to choose the best summary from all candidates based on their estimated
factual consistency score. We propose to extract medical facts of the input
medical report, its gold summary, and candidate summaries based on the RadGraph
schema and design the fact-guided reranker to efficiently incorporate the
extracted medical facts for selecting the optimal summary. We decompose the
fact-guided reranker into the factual knowledge graph generation and the
factual scorer, which allows the reranker to model the mapping between the
medical facts of the input text and its gold summary, thus can select the
optimal summary even the gold summary can't be observed during inference. We
also present a fact-based ranking metric (RadMRR) for measuring the ability of
the reranker on selecting factual consistent candidates. Experimental results
on two benchmark datasets demonstrate the superiority of our method in
generating summaries with higher factual consistency scores when compared with
existing methods.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 02:51:57 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 03:09:09 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Xie",
"Qianqian",
""
],
[
"Zhou",
"Jiayu",
""
],
[
"Peng",
"Yifan",
""
],
[
"Wang",
"Fei",
""
]
] | TITLE: FactReranker: Fact-guided Reranker for Faithful Radiology Report
Summarization
ABSTRACT: Automatic radiology report summarization is a crucial clinical task, whose
key challenge is to maintain factual accuracy between produced summaries and
ground truth radiology findings. Existing research adopts reinforcement
learning to directly optimize factual consistency metrics such as CheXBert or
RadGraph score. However, their decoding method using greedy search or beam
search considers no factual consistency when picking the optimal candidate,
leading to limited factual consistency improvement. To address it, we propose a
novel second-stage summarizing approach FactReranker, the first attempt that
learns to choose the best summary from all candidates based on their estimated
factual consistency score. We propose to extract medical facts of the input
medical report, its gold summary, and candidate summaries based on the RadGraph
schema and design the fact-guided reranker to efficiently incorporate the
extracted medical facts for selecting the optimal summary. We decompose the
fact-guided reranker into the factual knowledge graph generation and the
factual scorer, which allows the reranker to model the mapping between the
medical facts of the input text and its gold summary, thus can select the
optimal summary even the gold summary can't be observed during inference. We
also present a fact-based ranking metric (RadMRR) for measuring the ability of
the reranker on selecting factual consistent candidates. Experimental results
on two benchmark datasets demonstrate the superiority of our method in
generating summaries with higher factual consistency scores when compared with
existing methods.
| no_new_dataset | 0.949669 |
2303.08450 | Ziyu Yao | Ziyu Yao, Xuxin Cheng, Yuexian Zou | PoseRAC: Pose Saliency Transformer for Repetitive Action Counting | 10 pages, 7 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a significant contribution to the field of repetitive
action counting through the introduction of a new approach called Pose Saliency
Representation. The proposed method efficiently represents each action using
only two salient poses instead of redundant frames, which significantly reduces
the computational cost while improving the performance. Moreover, we introduce
a pose-level method, PoseRAC, which is based on this representation and
achieves state-of-the-art performance on two new version datasets by using Pose
Saliency Annotation to annotate salient poses for training. Our lightweight
model is highly efficient, requiring only 20 minutes for training on a GPU, and
infers nearly 10x faster compared to previous methods. In addition, our
approach achieves a substantial improvement over the previous state-of-the-art
TransRAC, achieving an OBO metric of 0.56 compared to 0.29 of TransRAC. The
code and new dataset are available at https://github.com/MiracleDance/PoseRAC
for further research and experimentation, making our proposed approach highly
accessible to the research community.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 08:51:17 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 01:33:08 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Yao",
"Ziyu",
""
],
[
"Cheng",
"Xuxin",
""
],
[
"Zou",
"Yuexian",
""
]
] | TITLE: PoseRAC: Pose Saliency Transformer for Repetitive Action Counting
ABSTRACT: This paper presents a significant contribution to the field of repetitive
action counting through the introduction of a new approach called Pose Saliency
Representation. The proposed method efficiently represents each action using
only two salient poses instead of redundant frames, which significantly reduces
the computational cost while improving the performance. Moreover, we introduce
a pose-level method, PoseRAC, which is based on this representation and
achieves state-of-the-art performance on two new version datasets by using Pose
Saliency Annotation to annotate salient poses for training. Our lightweight
model is highly efficient, requiring only 20 minutes for training on a GPU, and
infers nearly 10x faster compared to previous methods. In addition, our
approach achieves a substantial improvement over the previous state-of-the-art
TransRAC, achieving an OBO metric of 0.56 compared to 0.29 of TransRAC. The
code and new dataset are available at https://github.com/MiracleDance/PoseRAC
for further research and experimentation, making our proposed approach highly
accessible to the research community.
| new_dataset | 0.908456 |
2303.08848 | Yang Zhou | Yang Zhou, Giuseppe Loianno | PENet: A Joint Panoptic Edge Detection Network | 7 pages, 5 figures, submitted to the IEEE/RSJ International
Conference on Intelligent Robots and Systems (IROS 2023) | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, compact and efficient scene understanding representations
have gained popularity in increasing situational awareness and autonomy of
robotic systems. In this work, we illustrate the concept of a panoptic edge
segmentation and propose PENet, a novel detection network called that combines
semantic edge detection and instance-level perception into a compact panoptic
edge representation. This is obtained through a joint network by multi-task
learning that concurrently predicts semantic edges, instance centers and offset
flow map without bounding box predictions exploiting the cross-task
correlations among the tasks. The proposed approach allows extending semantic
edge detection to panoptic edge detection which encapsulates both
category-aware and instance-aware segmentation. We validate the proposed
panoptic edge segmentation method and demonstrate its effectiveness on the
real-world Cityscapes dataset.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 18:01:01 GMT"
}
] | 2023-03-17T00:00:00 | [
[
"Zhou",
"Yang",
""
],
[
"Loianno",
"Giuseppe",
""
]
] | TITLE: PENet: A Joint Panoptic Edge Detection Network
ABSTRACT: In recent years, compact and efficient scene understanding representations
have gained popularity in increasing situational awareness and autonomy of
robotic systems. In this work, we illustrate the concept of a panoptic edge
segmentation and propose PENet, a novel detection network called that combines
semantic edge detection and instance-level perception into a compact panoptic
edge representation. This is obtained through a joint network by multi-task
learning that concurrently predicts semantic edges, instance centers and offset
flow map without bounding box predictions exploiting the cross-task
correlations among the tasks. The proposed approach allows extending semantic
edge detection to panoptic edge detection which encapsulates both
category-aware and instance-aware segmentation. We validate the proposed
panoptic edge segmentation method and demonstrate its effectiveness on the
real-world Cityscapes dataset.
| no_new_dataset | 0.948537 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.