id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.2k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-05-30 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
| label
stringclasses 2
values | prob
float64 0.5
0.73
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2007.03481
|
Vikram Krishnamurthy
|
Kunal Pattanayak and Vikram Krishnamurthy
|
Necessary and Sufficient Conditions for Inverse Reinforcement Learning
of Bayesian Stopping Time Problems
| null |
Journal of Machine Learning Research, Vol.24, 2023
| null | null |
cs.LG cs.SY econ.TH eess.SY stat.ML
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper presents an inverse reinforcement learning~(IRL) framework for
Bayesian stopping time problems. By observing the actions of a Bayesian
decision maker, we provide a necessary and sufficient condition to identify if
these actions are consistent with optimizing a cost function. In a Bayesian
(partially observed) setting, the inverse learner can at best identify
optimality wrt the observed strategies. Our IRL algorithm identifies optimality
and then constructs set-valued estimates of the cost function.To achieve this
IRL objective, we use novel ideas from Bayesian revealed preferences stemming
from microeconomics. We illustrate the proposed IRL scheme using two important
examples of stopping time problems, namely, sequential hypothesis testing and
Bayesian search. As a real-world example, we illustrate using a YouTube dataset
comprising metadata from 190000 videos how the proposed IRL method predicts
user engagement in online multimedia platforms with high accuracy. Finally, for
finite datasets, we propose an IRL detection algorithm and give finite sample
bounds on its error probabilities.
|
[
{
"version": "v1",
"created": "Tue, 7 Jul 2020 14:14:12 GMT"
},
{
"version": "v2",
"created": "Sun, 25 Oct 2020 00:40:40 GMT"
},
{
"version": "v3",
"created": "Thu, 22 Jul 2021 12:03:29 GMT"
},
{
"version": "v4",
"created": "Sun, 20 Feb 2022 14:54:13 GMT"
},
{
"version": "v5",
"created": "Tue, 17 Jan 2023 10:07:28 GMT"
},
{
"version": "v6",
"created": "Wed, 15 Mar 2023 23:57:38 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Pattanayak",
"Kunal",
""
],
[
"Krishnamurthy",
"Vikram",
""
]
] |
TITLE: Necessary and Sufficient Conditions for Inverse Reinforcement Learning
of Bayesian Stopping Time Problems
ABSTRACT: This paper presents an inverse reinforcement learning~(IRL) framework for
Bayesian stopping time problems. By observing the actions of a Bayesian
decision maker, we provide a necessary and sufficient condition to identify if
these actions are consistent with optimizing a cost function. In a Bayesian
(partially observed) setting, the inverse learner can at best identify
optimality wrt the observed strategies. Our IRL algorithm identifies optimality
and then constructs set-valued estimates of the cost function.To achieve this
IRL objective, we use novel ideas from Bayesian revealed preferences stemming
from microeconomics. We illustrate the proposed IRL scheme using two important
examples of stopping time problems, namely, sequential hypothesis testing and
Bayesian search. As a real-world example, we illustrate using a YouTube dataset
comprising metadata from 190000 videos how the proposed IRL method predicts
user engagement in online multimedia platforms with high accuracy. Finally, for
finite datasets, we propose an IRL detection algorithm and give finite sample
bounds on its error probabilities.
|
no_new_dataset
| 0.707582
|
2012.07962
|
Michalis Lazarou Mr
|
Michalis Lazarou, Tania Stathaki, Yannis Avrithis
|
Iterative label cleaning for transductive and semi-supervised few-shot
learning
|
published in ICCV 2021
| null | null | null |
cs.LG cs.AI cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Few-shot learning amounts to learning representations and acquiring knowledge
such that novel tasks may be solved with both supervision and data being
limited. Improved performance is possible by transductive inference, where the
entire test set is available concurrently, and semi-supervised learning, where
more unlabeled data is available. Focusing on these two settings, we introduce
a new algorithm that leverages the manifold structure of the labeled and
unlabeled data distribution to predict pseudo-labels, while balancing over
classes and using the loss value distribution of a limited-capacity classifier
to select the cleanest labels, iteratively improving the quality of
pseudo-labels. Our solution surpasses or matches the state of the art results
on four benchmark datasets, namely miniImageNet, tieredImageNet, CUB and
CIFAR-FS, while being robust over feature space pre-processing and the quantity
of available data. The publicly available source code can be found in
https://github.com/MichalisLazarou/iLPC.
|
[
{
"version": "v1",
"created": "Mon, 14 Dec 2020 21:54:11 GMT"
},
{
"version": "v2",
"created": "Sun, 31 Oct 2021 21:39:52 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 15:05:23 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Lazarou",
"Michalis",
""
],
[
"Stathaki",
"Tania",
""
],
[
"Avrithis",
"Yannis",
""
]
] |
TITLE: Iterative label cleaning for transductive and semi-supervised few-shot
learning
ABSTRACT: Few-shot learning amounts to learning representations and acquiring knowledge
such that novel tasks may be solved with both supervision and data being
limited. Improved performance is possible by transductive inference, where the
entire test set is available concurrently, and semi-supervised learning, where
more unlabeled data is available. Focusing on these two settings, we introduce
a new algorithm that leverages the manifold structure of the labeled and
unlabeled data distribution to predict pseudo-labels, while balancing over
classes and using the loss value distribution of a limited-capacity classifier
to select the cleanest labels, iteratively improving the quality of
pseudo-labels. Our solution surpasses or matches the state of the art results
on four benchmark datasets, namely miniImageNet, tieredImageNet, CUB and
CIFAR-FS, while being robust over feature space pre-processing and the quantity
of available data. The publicly available source code can be found in
https://github.com/MichalisLazarou/iLPC.
|
no_new_dataset
| 0.709837
|
2107.07075
|
Gintare Karolina Dziugaite
|
Mansheej Paul, Surya Ganguli, Gintare Karolina Dziugaite
|
Deep Learning on a Data Diet: Finding Important Examples Early in
Training
|
21 pages, 18 figures
|
Advances in Neural Information Processing Systems 34 (NeurIPS
2021)
| null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent success in deep learning has partially been driven by training
increasingly overparametrized networks on ever larger datasets. It is therefore
natural to ask: how much of the data is superfluous, which examples are
important for generalization, and how do we find them? In this work, we make
the striking observation that, in standard vision datasets, simple scores
averaged over several weight initializations can be used to identify important
examples very early in training. We propose two such scores -- the Gradient
Normed (GraNd) and the Error L2-Norm (EL2N) scores -- and demonstrate their
efficacy on a range of architectures and datasets by pruning significant
fractions of training data without sacrificing test accuracy. In fact, using
EL2N scores calculated a few epochs into training, we can prune half of the
CIFAR10 training set while slightly improving test accuracy. Furthermore, for a
given dataset, EL2N scores from one architecture or hyperparameter
configuration generalize to other configurations. Compared to recent work that
prunes data by discarding examples that are rarely forgotten over the course of
training, our scores use only local information early in training. We also use
our scores to detect noisy examples and study training dynamics through the
lens of important examples -- we investigate how the data distribution shapes
the loss surface and identify subspaces of the model's data representation that
are relatively stable over training.
|
[
{
"version": "v1",
"created": "Thu, 15 Jul 2021 02:12:20 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 13:51:14 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Paul",
"Mansheej",
""
],
[
"Ganguli",
"Surya",
""
],
[
"Dziugaite",
"Gintare Karolina",
""
]
] |
TITLE: Deep Learning on a Data Diet: Finding Important Examples Early in
Training
ABSTRACT: Recent success in deep learning has partially been driven by training
increasingly overparametrized networks on ever larger datasets. It is therefore
natural to ask: how much of the data is superfluous, which examples are
important for generalization, and how do we find them? In this work, we make
the striking observation that, in standard vision datasets, simple scores
averaged over several weight initializations can be used to identify important
examples very early in training. We propose two such scores -- the Gradient
Normed (GraNd) and the Error L2-Norm (EL2N) scores -- and demonstrate their
efficacy on a range of architectures and datasets by pruning significant
fractions of training data without sacrificing test accuracy. In fact, using
EL2N scores calculated a few epochs into training, we can prune half of the
CIFAR10 training set while slightly improving test accuracy. Furthermore, for a
given dataset, EL2N scores from one architecture or hyperparameter
configuration generalize to other configurations. Compared to recent work that
prunes data by discarding examples that are rarely forgotten over the course of
training, our scores use only local information early in training. We also use
our scores to detect noisy examples and study training dynamics through the
lens of important examples -- we investigate how the data distribution shapes
the loss surface and identify subspaces of the model's data representation that
are relatively stable over training.
|
no_new_dataset
| 0.708648
|
2109.00831
|
Davide Gurnari
|
Pawe{\l} D{\l}otko, Davide Gurnari and Radmila Sazdanovic
|
Mapper-type algorithms for complex data and relations
|
Changed title. Rewrote introduction and section 3. Added new examples
and applications. 20 pages, 16 figures
| null | null | null |
math.AT cs.AI math.CO math.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Mapper and Ball Mapper are Topological Data Analysis tools used for exploring
high dimensional point clouds and visualizing scalar-valued functions on those
point clouds. Inspired by open questions in knot theory, new features are added
to Ball Mapper that enable encoding of the structure, internal relations and
symmetries of the point cloud. Moreover, the strengths of Mapper and Ball
Mapper constructions are combined to create a tool for comparing high
dimensional data descriptors of a single dataset. This new hybrid algorithm,
Mapper on Ball Mapper, is applicable to high dimensional lens functions. As a
proof of concept we include applications to knot and game theory, as well as
material science and cancer research.
|
[
{
"version": "v1",
"created": "Thu, 2 Sep 2021 10:30:51 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 08:56:22 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Dłotko",
"Paweł",
""
],
[
"Gurnari",
"Davide",
""
],
[
"Sazdanovic",
"Radmila",
""
]
] |
TITLE: Mapper-type algorithms for complex data and relations
ABSTRACT: Mapper and Ball Mapper are Topological Data Analysis tools used for exploring
high dimensional point clouds and visualizing scalar-valued functions on those
point clouds. Inspired by open questions in knot theory, new features are added
to Ball Mapper that enable encoding of the structure, internal relations and
symmetries of the point cloud. Moreover, the strengths of Mapper and Ball
Mapper constructions are combined to create a tool for comparing high
dimensional data descriptors of a single dataset. This new hybrid algorithm,
Mapper on Ball Mapper, is applicable to high dimensional lens functions. As a
proof of concept we include applications to knot and game theory, as well as
material science and cancer research.
|
no_new_dataset
| 0.710025
|
2110.00681
|
Xuekui Zhang
|
Xiaowen Cao, Li Xing, Elham Majd, Hua He, Junhua Gu, Xuekui Zhang
|
A systematic evaluation of methods for cell phenotype classification
using single-cell RNA sequencing data
|
21 pages, 4 figures, 1 table
|
Front. Genet. 13:836798 (2022)
|
10.3389/fgene.2022.836798
| null |
q-bio.GN cs.LG stat.AP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Background: Single-cell RNA sequencing (scRNA-seq) yields valuable insights
about gene expression and gives critical information about complex tissue
cellular composition. In the analysis of single-cell RNA sequencing, the
annotations of cell subtypes are often done manually, which is time-consuming
and irreproducible. Garnett is a cell-type annotation software based the on
elastic net method. Besides cell-type annotation, supervised machine learning
methods can also be applied to predict other cell phenotypes from genomic data.
Despite the popularity of such applications, there is no existing study to
systematically investigate the performance of those supervised algorithms in
various sizes of scRNA-seq data sets.
Methods and Results: This study evaluates 13 popular supervised machine
learning algorithms to classify cell phenotypes, using published real and
simulated data sets with diverse cell sizes. The benchmark contained two parts.
In the first part, we used real data sets to assess the popular supervised
algorithms' computing speed and cell phenotype classification performance. The
classification performances were evaluated using AUC statistics, F1-score,
precision, recall, and false-positive rate. In the second part, we evaluated
gene selection performance using published simulated data sets with a known
list of real genes.
Conclusion: The study outcomes showed that ElasticNet with interactions
performed best in small and medium data sets. NB was another appropriate method
for medium data sets. In large data sets, XGB works excellent. Ensemble
algorithms were not significantly superior to individual machine learning
methods. Adding interactions to ElasticNet can help, and the improvement was
significant in small data sets.
|
[
{
"version": "v1",
"created": "Fri, 1 Oct 2021 23:24:15 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Cao",
"Xiaowen",
""
],
[
"Xing",
"Li",
""
],
[
"Majd",
"Elham",
""
],
[
"He",
"Hua",
""
],
[
"Gu",
"Junhua",
""
],
[
"Zhang",
"Xuekui",
""
]
] |
TITLE: A systematic evaluation of methods for cell phenotype classification
using single-cell RNA sequencing data
ABSTRACT: Background: Single-cell RNA sequencing (scRNA-seq) yields valuable insights
about gene expression and gives critical information about complex tissue
cellular composition. In the analysis of single-cell RNA sequencing, the
annotations of cell subtypes are often done manually, which is time-consuming
and irreproducible. Garnett is a cell-type annotation software based the on
elastic net method. Besides cell-type annotation, supervised machine learning
methods can also be applied to predict other cell phenotypes from genomic data.
Despite the popularity of such applications, there is no existing study to
systematically investigate the performance of those supervised algorithms in
various sizes of scRNA-seq data sets.
Methods and Results: This study evaluates 13 popular supervised machine
learning algorithms to classify cell phenotypes, using published real and
simulated data sets with diverse cell sizes. The benchmark contained two parts.
In the first part, we used real data sets to assess the popular supervised
algorithms' computing speed and cell phenotype classification performance. The
classification performances were evaluated using AUC statistics, F1-score,
precision, recall, and false-positive rate. In the second part, we evaluated
gene selection performance using published simulated data sets with a known
list of real genes.
Conclusion: The study outcomes showed that ElasticNet with interactions
performed best in small and medium data sets. NB was another appropriate method
for medium data sets. In large data sets, XGB works excellent. Ensemble
algorithms were not significantly superior to individual machine learning
methods. Adding interactions to ElasticNet can help, and the improvement was
significant in small data sets.
|
no_new_dataset
| 0.712263
|
2202.13517
|
Dayang Wang
|
Dayang Wang, Fenglei Fan, Zhan Wu, Rui Liu, Fei Wang, Hengyong Yu
|
CTformer: Convolution-free Token2Token Dilated Vision Transformer for
Low-dose CT Denoising
|
11 pages, 14 figures
| null |
10.1088/1361-6560/acc000
| null |
eess.IV cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Low-dose computed tomography (LDCT) denoising is an important problem in CT
research. Compared to the normal dose CT (NDCT), LDCT images are subjected to
severe noise and artifacts. Recently in many studies, vision transformers have
shown superior feature representation ability over convolutional neural
networks (CNNs). However, unlike CNNs, the potential of vision transformers in
LDCT denoising was little explored so far. To fill this gap, we propose a
Convolution-free Token2Token Dilated Vision Transformer for low-dose CT
denoising. The CTformer uses a more powerful token rearrangement to encompass
local contextual information and thus avoids convolution. It also dilates and
shifts feature maps to capture longer-range interaction. We interpret the
CTformer by statically inspecting patterns of its internal attention maps and
dynamically tracing the hierarchical attention flow with an explanatory graph.
Furthermore, an overlapped inference mechanism is introduced to effectively
eliminate the boundary artifacts that are common for encoder-decoder-based
denoising models. Experimental results on Mayo LDCT dataset suggest that the
CTformer outperforms the state-of-the-art denoising methods with a low
computation overhead.
|
[
{
"version": "v1",
"created": "Mon, 28 Feb 2022 02:58:16 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Dayang",
""
],
[
"Fan",
"Fenglei",
""
],
[
"Wu",
"Zhan",
""
],
[
"Liu",
"Rui",
""
],
[
"Wang",
"Fei",
""
],
[
"Yu",
"Hengyong",
""
]
] |
TITLE: CTformer: Convolution-free Token2Token Dilated Vision Transformer for
Low-dose CT Denoising
ABSTRACT: Low-dose computed tomography (LDCT) denoising is an important problem in CT
research. Compared to the normal dose CT (NDCT), LDCT images are subjected to
severe noise and artifacts. Recently in many studies, vision transformers have
shown superior feature representation ability over convolutional neural
networks (CNNs). However, unlike CNNs, the potential of vision transformers in
LDCT denoising was little explored so far. To fill this gap, we propose a
Convolution-free Token2Token Dilated Vision Transformer for low-dose CT
denoising. The CTformer uses a more powerful token rearrangement to encompass
local contextual information and thus avoids convolution. It also dilates and
shifts feature maps to capture longer-range interaction. We interpret the
CTformer by statically inspecting patterns of its internal attention maps and
dynamically tracing the hierarchical attention flow with an explanatory graph.
Furthermore, an overlapped inference mechanism is introduced to effectively
eliminate the boundary artifacts that are common for encoder-decoder-based
denoising models. Experimental results on Mayo LDCT dataset suggest that the
CTformer outperforms the state-of-the-art denoising methods with a low
computation overhead.
|
no_new_dataset
| 0.710007
|
2203.06111
|
Oleg Voynov
|
Oleg Voynov, Gleb Bobrovskikh, Pavel Karpyshev, Saveliy Galochkin,
Andrei-Timotei Ardelean, Arseniy Bozhenko, Ekaterina Karmanova, Pavel
Kopanev, Yaroslav Labutin-Rymsho, Ruslan Rakhimov, Aleksandr Safin, Valerii
Serpiva, Alexey Artemov, Evgeny Burnaev, Dzmitry Tsetserukou, Denis Zorin
|
Multi-sensor large-scale dataset for multi-view 3D reconstruction
|
v4: final camera-ready version
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a new multi-sensor dataset for multi-view 3D surface
reconstruction. It includes registered RGB and depth data from sensors of
different resolutions and modalities: smartphones, Intel RealSense, Microsoft
Kinect, industrial cameras, and structured-light scanner. The scenes are
selected to emphasize a diverse set of material properties challenging for
existing algorithms. We provide around 1.4 million images of 107 different
scenes acquired from 100 viewing directions under 14 lighting conditions. We
expect our dataset will be useful for evaluation and training of 3D
reconstruction algorithms and for related tasks. The dataset is available at
skoltech3d.appliedai.tech.
|
[
{
"version": "v1",
"created": "Fri, 11 Mar 2022 17:32:27 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Jan 2023 19:00:08 GMT"
},
{
"version": "v3",
"created": "Wed, 8 Feb 2023 10:33:55 GMT"
},
{
"version": "v4",
"created": "Tue, 28 Mar 2023 11:11:08 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Voynov",
"Oleg",
""
],
[
"Bobrovskikh",
"Gleb",
""
],
[
"Karpyshev",
"Pavel",
""
],
[
"Galochkin",
"Saveliy",
""
],
[
"Ardelean",
"Andrei-Timotei",
""
],
[
"Bozhenko",
"Arseniy",
""
],
[
"Karmanova",
"Ekaterina",
""
],
[
"Kopanev",
"Pavel",
""
],
[
"Labutin-Rymsho",
"Yaroslav",
""
],
[
"Rakhimov",
"Ruslan",
""
],
[
"Safin",
"Aleksandr",
""
],
[
"Serpiva",
"Valerii",
""
],
[
"Artemov",
"Alexey",
""
],
[
"Burnaev",
"Evgeny",
""
],
[
"Tsetserukou",
"Dzmitry",
""
],
[
"Zorin",
"Denis",
""
]
] |
TITLE: Multi-sensor large-scale dataset for multi-view 3D reconstruction
ABSTRACT: We present a new multi-sensor dataset for multi-view 3D surface
reconstruction. It includes registered RGB and depth data from sensors of
different resolutions and modalities: smartphones, Intel RealSense, Microsoft
Kinect, industrial cameras, and structured-light scanner. The scenes are
selected to emphasize a diverse set of material properties challenging for
existing algorithms. We provide around 1.4 million images of 107 different
scenes acquired from 100 viewing directions under 14 lighting conditions. We
expect our dataset will be useful for evaluation and training of 3D
reconstruction algorithms and for related tasks. The dataset is available at
skoltech3d.appliedai.tech.
|
new_dataset
| 0.713238
|
2204.01244
|
Bohan Zhuang
|
Haoyu He, Jianfei Cai, Zizheng Pan, Jing Liu, Jing Zhang, Dacheng Tao,
Bohan Zhuang
|
Dynamic Focus-aware Positional Queries for Semantic Segmentation
|
CVPR 2023
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
The DETR-like segmentors have underpinned the most recent breakthroughs in
semantic segmentation, which end-to-end train a set of queries representing the
class prototypes or target segments. Recently, masked attention is proposed to
restrict each query to only attend to the foreground regions predicted by the
preceding decoder block for easier optimization. Although promising, it relies
on the learnable parameterized positional queries which tend to encode the
dataset statistics, leading to inaccurate localization for distinct individual
queries. In this paper, we propose a simple yet effective query design for
semantic segmentation termed Dynamic Focus-aware Positional Queries (DFPQ),
which dynamically generates positional queries conditioned on the
cross-attention scores from the preceding decoder block and the positional
encodings for the corresponding image features, simultaneously. Therefore, our
DFPQ preserves rich localization information for the target segments and
provides accurate and fine-grained positional priors. In addition, we propose
to efficiently deal with high-resolution cross-attention by only aggregating
the contextual tokens based on the low-resolution cross-attention scores to
perform local relation aggregation. Extensive experiments on ADE20K and
Cityscapes show that with the two modifications on Mask2former, our framework
achieves SOTA performance and outperforms Mask2former by clear margins of 1.1%,
1.9%, and 1.1% single-scale mIoU with ResNet-50, Swin-T, and Swin-B backbones
on the ADE20K validation set, respectively. Source code is available at
https://github.com/ziplab/FASeg
|
[
{
"version": "v1",
"created": "Mon, 4 Apr 2022 05:16:41 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Nov 2022 02:56:14 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 02:42:17 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"He",
"Haoyu",
""
],
[
"Cai",
"Jianfei",
""
],
[
"Pan",
"Zizheng",
""
],
[
"Liu",
"Jing",
""
],
[
"Zhang",
"Jing",
""
],
[
"Tao",
"Dacheng",
""
],
[
"Zhuang",
"Bohan",
""
]
] |
TITLE: Dynamic Focus-aware Positional Queries for Semantic Segmentation
ABSTRACT: The DETR-like segmentors have underpinned the most recent breakthroughs in
semantic segmentation, which end-to-end train a set of queries representing the
class prototypes or target segments. Recently, masked attention is proposed to
restrict each query to only attend to the foreground regions predicted by the
preceding decoder block for easier optimization. Although promising, it relies
on the learnable parameterized positional queries which tend to encode the
dataset statistics, leading to inaccurate localization for distinct individual
queries. In this paper, we propose a simple yet effective query design for
semantic segmentation termed Dynamic Focus-aware Positional Queries (DFPQ),
which dynamically generates positional queries conditioned on the
cross-attention scores from the preceding decoder block and the positional
encodings for the corresponding image features, simultaneously. Therefore, our
DFPQ preserves rich localization information for the target segments and
provides accurate and fine-grained positional priors. In addition, we propose
to efficiently deal with high-resolution cross-attention by only aggregating
the contextual tokens based on the low-resolution cross-attention scores to
perform local relation aggregation. Extensive experiments on ADE20K and
Cityscapes show that with the two modifications on Mask2former, our framework
achieves SOTA performance and outperforms Mask2former by clear margins of 1.1%,
1.9%, and 1.1% single-scale mIoU with ResNet-50, Swin-T, and Swin-B backbones
on the ADE20K validation set, respectively. Source code is available at
https://github.com/ziplab/FASeg
|
no_new_dataset
| 0.707632
|
2204.06521
|
Virginie Do
|
Virginie Do and Nicolas Usunier
|
Optimizing generalized Gini indices for fairness in rankings
|
Accepted to SIGIR 2022
| null | null | null |
cs.IR cs.AI cs.CY
|
http://creativecommons.org/licenses/by/4.0/
|
There is growing interest in designing recommender systems that aim at being
fair towards item producers or their least satisfied users. Inspired by the
domain of inequality measurement in economics, this paper explores the use of
generalized Gini welfare functions (GGFs) as a means to specify the normative
criterion that recommender systems should optimize for. GGFs weight individuals
depending on their ranks in the population, giving more weight to worse-off
individuals to promote equality. Depending on these weights, GGFs minimize the
Gini index of item exposure to promote equality between items, or focus on the
performance on specific quantiles of least satisfied users. GGFs for ranking
are challenging to optimize because they are non-differentiable. We resolve
this challenge by leveraging tools from non-smooth optimization and projection
operators used in differentiable sorting. We present experiments using real
datasets with up to 15k users and items, which show that our approach obtains
better trade-offs than the baselines on a variety of recommendation tasks and
fairness criteria.
|
[
{
"version": "v1",
"created": "Sat, 2 Apr 2022 09:40:07 GMT"
},
{
"version": "v2",
"created": "Thu, 14 Apr 2022 09:36:54 GMT"
},
{
"version": "v3",
"created": "Sat, 23 Apr 2022 09:36:13 GMT"
},
{
"version": "v4",
"created": "Tue, 28 Mar 2023 08:45:10 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Do",
"Virginie",
""
],
[
"Usunier",
"Nicolas",
""
]
] |
TITLE: Optimizing generalized Gini indices for fairness in rankings
ABSTRACT: There is growing interest in designing recommender systems that aim at being
fair towards item producers or their least satisfied users. Inspired by the
domain of inequality measurement in economics, this paper explores the use of
generalized Gini welfare functions (GGFs) as a means to specify the normative
criterion that recommender systems should optimize for. GGFs weight individuals
depending on their ranks in the population, giving more weight to worse-off
individuals to promote equality. Depending on these weights, GGFs minimize the
Gini index of item exposure to promote equality between items, or focus on the
performance on specific quantiles of least satisfied users. GGFs for ranking
are challenging to optimize because they are non-differentiable. We resolve
this challenge by leveraging tools from non-smooth optimization and projection
operators used in differentiable sorting. We present experiments using real
datasets with up to 15k users and items, which show that our approach obtains
better trade-offs than the baselines on a variety of recommendation tasks and
fairness criteria.
|
no_new_dataset
| 0.711205
|
2204.10997
|
Haozheng Zhang
|
Haozheng Zhang, Hubert P. H. Shum and Edmond S. L. Ho
|
Cerebral Palsy Prediction with Frequency Attention Informed Graph
Convolutional Networks
| null | null | null | null |
cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Early diagnosis and intervention are clinically considered the paramount part
of treating cerebral palsy (CP), so it is essential to design an efficient and
interpretable automatic prediction system for CP. We highlight a significant
difference between CP infants' frequency of human movement and that of the
healthy group, which improves prediction performance. However, the existing
deep learning-based methods did not use the frequency information of infants'
movement for CP prediction. This paper proposes a frequency attention informed
graph convolutional network and validates it on two consumer-grade RGB video
datasets, namely MINI-RGBD and RVI-38 datasets. Our proposed frequency
attention module aids in improving both classification performance and system
interpretability. In addition, we design a frequency-binning method that
retains the critical frequency of the human joint position data while filtering
the noise. Our prediction performance achieves state-of-the-art research on
both datasets. Our work demonstrates the effectiveness of frequency information
in supporting the prediction of CP non-intrusively and provides a way for
supporting the early diagnosis of CP in the resource-limited regions where the
clinical resources are not abundant.
|
[
{
"version": "v1",
"created": "Sat, 23 Apr 2022 05:48:17 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 16:14:46 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhang",
"Haozheng",
""
],
[
"Shum",
"Hubert P. H.",
""
],
[
"Ho",
"Edmond S. L.",
""
]
] |
TITLE: Cerebral Palsy Prediction with Frequency Attention Informed Graph
Convolutional Networks
ABSTRACT: Early diagnosis and intervention are clinically considered the paramount part
of treating cerebral palsy (CP), so it is essential to design an efficient and
interpretable automatic prediction system for CP. We highlight a significant
difference between CP infants' frequency of human movement and that of the
healthy group, which improves prediction performance. However, the existing
deep learning-based methods did not use the frequency information of infants'
movement for CP prediction. This paper proposes a frequency attention informed
graph convolutional network and validates it on two consumer-grade RGB video
datasets, namely MINI-RGBD and RVI-38 datasets. Our proposed frequency
attention module aids in improving both classification performance and system
interpretability. In addition, we design a frequency-binning method that
retains the critical frequency of the human joint position data while filtering
the noise. Our prediction performance achieves state-of-the-art research on
both datasets. Our work demonstrates the effectiveness of frequency information
in supporting the prediction of CP non-intrusively and provides a way for
supporting the early diagnosis of CP in the resource-limited regions where the
clinical resources are not abundant.
|
no_new_dataset
| 0.710446
|
2205.15530
|
Saisai Ding
|
Jun Shi, Yuanming Zhang, Zheng Li, Xiangmin Han, Saisai Ding, Jun
Wang, Shihui Ying
|
Pseudo-Data based Self-Supervised Federated Learning for Classification
of Histopathological Images
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Computer-aided diagnosis (CAD) can help pathologists improve diagnostic
accuracy together with consistency and repeatability for cancers. However, the
CAD models trained with the histopathological images only from a single center
(hospital) generally suffer from the generalization problem due to the
straining inconsistencies among different centers. In this work, we propose a
pseudo-data based self-supervised federated learning (FL) framework, named
SSL-FT-BT, to improve both the diagnostic accuracy and generalization of CAD
models. Specifically, the pseudo histopathological images are generated from
each center, which contains inherent and specific properties corresponding to
the real images in this center, but does not include the privacy information.
These pseudo images are then shared in the central server for self-supervised
learning (SSL). A multi-task SSL is then designed to fully learn both the
center-specific information and common inherent representation according to the
data characteristics. Moreover, a novel Barlow Twins based FL (FL-BT) algorithm
is proposed to improve the local training for the CAD model in each center by
conducting contrastive learning, which benefits the optimization of the global
model in the FL procedure. The experimental results on three public
histopathological image datasets indicate the effectiveness of the proposed
SSL-FL-BT on both diagnostic accuracy and generalization.
|
[
{
"version": "v1",
"created": "Tue, 31 May 2022 04:23:50 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 15:07:27 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Shi",
"Jun",
""
],
[
"Zhang",
"Yuanming",
""
],
[
"Li",
"Zheng",
""
],
[
"Han",
"Xiangmin",
""
],
[
"Ding",
"Saisai",
""
],
[
"Wang",
"Jun",
""
],
[
"Ying",
"Shihui",
""
]
] |
TITLE: Pseudo-Data based Self-Supervised Federated Learning for Classification
of Histopathological Images
ABSTRACT: Computer-aided diagnosis (CAD) can help pathologists improve diagnostic
accuracy together with consistency and repeatability for cancers. However, the
CAD models trained with the histopathological images only from a single center
(hospital) generally suffer from the generalization problem due to the
straining inconsistencies among different centers. In this work, we propose a
pseudo-data based self-supervised federated learning (FL) framework, named
SSL-FT-BT, to improve both the diagnostic accuracy and generalization of CAD
models. Specifically, the pseudo histopathological images are generated from
each center, which contains inherent and specific properties corresponding to
the real images in this center, but does not include the privacy information.
These pseudo images are then shared in the central server for self-supervised
learning (SSL). A multi-task SSL is then designed to fully learn both the
center-specific information and common inherent representation according to the
data characteristics. Moreover, a novel Barlow Twins based FL (FL-BT) algorithm
is proposed to improve the local training for the CAD model in each center by
conducting contrastive learning, which benefits the optimization of the global
model in the FL procedure. The experimental results on three public
histopathological image datasets indicate the effectiveness of the proposed
SSL-FL-BT on both diagnostic accuracy and generalization.
|
no_new_dataset
| 0.710791
|
2206.05717
|
Juncheng Wang
|
Juncheng Wang, Junyu Gao, Yuan Yuan, Qi Wang
|
Crowd Localization from Gaussian Mixture Scoped Knowledge and Scoped
Teacher
|
Accepted by IEEE TIP
| null |
10.1109/TIP.2023.3251727
| null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Crowd localization is to predict each instance head position in crowd
scenarios. Since the distance of instances being to the camera are variant,
there exists tremendous gaps among scales of instances within an image, which
is called the intrinsic scale shift. The core reason of intrinsic scale shift
being one of the most essential issues in crowd localization is that it is
ubiquitous in crowd scenes and makes scale distribution chaotic.
To this end, the paper concentrates on access to tackle the chaos of the
scale distribution incurred by intrinsic scale shift. We propose Gaussian
Mixture Scope (GMS) to regularize the chaotic scale distribution. Concretely,
the GMS utilizes a Gaussian mixture distribution to adapt to scale distribution
and decouples the mixture model into sub-normal distributions to regularize the
chaos within the sub-distributions. Then, an alignment is introduced to
regularize the chaos among sub-distributions. However, despite that GMS is
effective in regularizing the data distribution, it amounts to dislodging the
hard samples in training set, which incurs overfitting. We assert that it is
blamed on the block of transferring the latent knowledge exploited by GMS from
data to model. Therefore, a Scoped Teacher playing a role of bridge in
knowledge transform is proposed. What' s more, the consistency regularization
is also introduced to implement knowledge transform. To that effect, the
further constraints are deployed on Scoped Teacher to derive feature
consistence between teacher and student end.
With proposed GMS and Scoped Teacher implemented on five mainstream datasets
of crowd localization, the extensive experiments demonstrate the superiority of
our work. Moreover, comparing with existing crowd locators, our work achieves
state-of-the-art via F1-meansure comprehensively on five datasets.
|
[
{
"version": "v1",
"created": "Sun, 12 Jun 2022 11:07:15 GMT"
},
{
"version": "v2",
"created": "Thu, 23 Feb 2023 10:24:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Juncheng",
""
],
[
"Gao",
"Junyu",
""
],
[
"Yuan",
"Yuan",
""
],
[
"Wang",
"Qi",
""
]
] |
TITLE: Crowd Localization from Gaussian Mixture Scoped Knowledge and Scoped
Teacher
ABSTRACT: Crowd localization is to predict each instance head position in crowd
scenarios. Since the distance of instances being to the camera are variant,
there exists tremendous gaps among scales of instances within an image, which
is called the intrinsic scale shift. The core reason of intrinsic scale shift
being one of the most essential issues in crowd localization is that it is
ubiquitous in crowd scenes and makes scale distribution chaotic.
To this end, the paper concentrates on access to tackle the chaos of the
scale distribution incurred by intrinsic scale shift. We propose Gaussian
Mixture Scope (GMS) to regularize the chaotic scale distribution. Concretely,
the GMS utilizes a Gaussian mixture distribution to adapt to scale distribution
and decouples the mixture model into sub-normal distributions to regularize the
chaos within the sub-distributions. Then, an alignment is introduced to
regularize the chaos among sub-distributions. However, despite that GMS is
effective in regularizing the data distribution, it amounts to dislodging the
hard samples in training set, which incurs overfitting. We assert that it is
blamed on the block of transferring the latent knowledge exploited by GMS from
data to model. Therefore, a Scoped Teacher playing a role of bridge in
knowledge transform is proposed. What' s more, the consistency regularization
is also introduced to implement knowledge transform. To that effect, the
further constraints are deployed on Scoped Teacher to derive feature
consistence between teacher and student end.
With proposed GMS and Scoped Teacher implemented on five mainstream datasets
of crowd localization, the extensive experiments demonstrate the superiority of
our work. Moreover, comparing with existing crowd locators, our work achieves
state-of-the-art via F1-meansure comprehensively on five datasets.
|
no_new_dataset
| 0.712063
|
2207.14181
|
Valerio Carruba Prof.
|
Valerio Carruba, Safwan Aljbaae, Gabriel Carit\'a, Rita Cassia
Domingos, Bruno Martins
|
Optimization of Artificial Neural Networks models applied to the
identification of images of asteroids' resonant arguments
|
15 pages, 13 figures, 3 tables. Accepted for publication at Celestial
Mechanics and Dynamical Astronomy
| null |
10.1007/s10569-022-10110-7
| null |
astro-ph.EP cs.LG eess.IV
|
http://creativecommons.org/licenses/by/4.0/
|
The asteroidal main belt is crossed by a web of mean-motion and secular
resonances, that occur when there is a commensurability between fundamental
frequencies of the asteroids and planets. Traditionally, these objects were
identified by visual inspection of the time evolution of their resonant
argument, which is a combination of orbital elements of the asteroid and the
perturbing planet(s). Since the population of asteroids affected by these
resonances is, in some cases, of the order of several thousand, this has become
a taxing task for a human observer. Recent works used Convolutional Neural
Networks (CNN) models to perform such task automatically. In this work, we
compare the outcome of such models with those of some of the most advanced and
publicly available CNN architectures, like the VGG, Inception and ResNet. The
performance of such models is first tested and optimized for overfitting
issues, using validation sets and a series of regularization techniques like
data augmentation, dropout, and batch normalization. The three best-performing
models were then used to predict the labels of larger testing databases
containing thousands of images. The VGG model, with and without
regularizations, proved to be the most efficient method to predict labels of
large datasets. Since the Vera C. Rubin observatory is likely to discover up to
four million new asteroids in the next few years, the use of these models might
become quite valuable to identify populations of resonant minor bodies.
|
[
{
"version": "v1",
"created": "Thu, 28 Jul 2022 15:46:39 GMT"
},
{
"version": "v2",
"created": "Mon, 15 Aug 2022 12:44:37 GMT"
},
{
"version": "v3",
"created": "Fri, 11 Nov 2022 14:42:25 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Carruba",
"Valerio",
""
],
[
"Aljbaae",
"Safwan",
""
],
[
"Caritá",
"Gabriel",
""
],
[
"Domingos",
"Rita Cassia",
""
],
[
"Martins",
"Bruno",
""
]
] |
TITLE: Optimization of Artificial Neural Networks models applied to the
identification of images of asteroids' resonant arguments
ABSTRACT: The asteroidal main belt is crossed by a web of mean-motion and secular
resonances, that occur when there is a commensurability between fundamental
frequencies of the asteroids and planets. Traditionally, these objects were
identified by visual inspection of the time evolution of their resonant
argument, which is a combination of orbital elements of the asteroid and the
perturbing planet(s). Since the population of asteroids affected by these
resonances is, in some cases, of the order of several thousand, this has become
a taxing task for a human observer. Recent works used Convolutional Neural
Networks (CNN) models to perform such task automatically. In this work, we
compare the outcome of such models with those of some of the most advanced and
publicly available CNN architectures, like the VGG, Inception and ResNet. The
performance of such models is first tested and optimized for overfitting
issues, using validation sets and a series of regularization techniques like
data augmentation, dropout, and batch normalization. The three best-performing
models were then used to predict the labels of larger testing databases
containing thousands of images. The VGG model, with and without
regularizations, proved to be the most efficient method to predict labels of
large datasets. Since the Vera C. Rubin observatory is likely to discover up to
four million new asteroids in the next few years, the use of these models might
become quite valuable to identify populations of resonant minor bodies.
|
no_new_dataset
| 0.709208
|
2209.02032
|
Benjamin Billot
|
Benjamin Billot, Colin Magdamo, You Cheng, Steven E. Arnold, Sudeshna
Das, Juan. E. Iglesias
|
Robust machine learning segmentation for large-scale analysis of
heterogeneous clinical brain MRI datasets
|
under review, extension of MICCAI 2022 paper
| null |
10.1073/pnas.2216399120
| null |
eess.IV cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Every year, millions of brain MRI scans are acquired in hospitals, which is a
figure considerably larger than the size of any research dataset. Therefore,
the ability to analyse such scans could transform neuroimaging research. Yet,
their potential remains untapped, since no automated algorithm is robust enough
to cope with the high variability in clinical acquisitions (MR contrasts,
resolutions, orientations, artefacts, subject populations). Here we present
SynthSeg+, an AI segmentation suite that enables, for the first time, robust
analysis of heterogeneous clinical datasets. In addition to whole-brain
segmentation, SynthSeg+ also performs cortical parcellation, intracranial
volume estimation, and automated detection of faulty segmentations (mainly
caused by scans of very low quality). We demonstrate SynthSeg+ in seven
experiments, including an ageing study on 14,000 scans, where it accurately
replicates atrophy patterns observed on data of much higher quality. SynthSeg+
is publicly released as a ready-to-use tool to unlock the potential of
quantitative morphometry.
|
[
{
"version": "v1",
"created": "Mon, 5 Sep 2022 16:09:24 GMT"
},
{
"version": "v2",
"created": "Wed, 4 Jan 2023 21:45:36 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Billot",
"Benjamin",
""
],
[
"Magdamo",
"Colin",
""
],
[
"Cheng",
"You",
""
],
[
"Arnold",
"Steven E.",
""
],
[
"Das",
"Sudeshna",
""
],
[
"Iglesias",
"Juan. E.",
""
]
] |
TITLE: Robust machine learning segmentation for large-scale analysis of
heterogeneous clinical brain MRI datasets
ABSTRACT: Every year, millions of brain MRI scans are acquired in hospitals, which is a
figure considerably larger than the size of any research dataset. Therefore,
the ability to analyse such scans could transform neuroimaging research. Yet,
their potential remains untapped, since no automated algorithm is robust enough
to cope with the high variability in clinical acquisitions (MR contrasts,
resolutions, orientations, artefacts, subject populations). Here we present
SynthSeg+, an AI segmentation suite that enables, for the first time, robust
analysis of heterogeneous clinical datasets. In addition to whole-brain
segmentation, SynthSeg+ also performs cortical parcellation, intracranial
volume estimation, and automated detection of faulty segmentations (mainly
caused by scans of very low quality). We demonstrate SynthSeg+ in seven
experiments, including an ageing study on 14,000 scans, where it accurately
replicates atrophy patterns observed on data of much higher quality. SynthSeg+
is publicly released as a ready-to-use tool to unlock the potential of
quantitative morphometry.
|
no_new_dataset
| 0.705138
|
2209.09553
|
Reza Akbari
|
Reza Sepahvand, Reza Akbari, Behnaz Jamasb, Sattar Hashemi, Omid
Boushehrian
|
Using Word Embedding and Convolution Neural Network for Bug Triaging by
Considering Design Flaws
| null | null |
10.1016/j.scico.2023.102945
| null |
cs.SE
|
http://creativecommons.org/licenses/by/4.0/
|
Resolving bugs in the maintenance phase of software is a complicated task.
Bug assignment is one of the main tasks for resolving bugs. Some Bugs cannot be
fixed properly without making design decisions and have to be assigned to
designers, rather than programmers, to avoid emerging bad smells that may cause
subsequent bug reports. Hence, it is important to refer some bugs to the
designer to check the possible design flaws. Based on our best knowledge, there
are a few works that have considered referring bugs to designers. Hence, this
issue is considered in this work. In this paper, a dataset is created, and a
CNN-based model is proposed to predict the need for assigning a bug to a
designer by learning the peculiarities of bug reports effective in creating bad
smells in the code. The features of each bug are extracted from CNN based on
its textual features, such as a summary and description. The number of bad
samples added to it in the fixing process using the PMD tool determines the bug
tag. The summary and description of the new bug are given to the model and the
model predicts the need to refer to the designer. The accuracy of 75% (or more)
was achieved for datasets with a sufficient number of samples for deep
learning-based model training. A model is proposed to predict bug referrals to
the designer. The efficiency of the model in predicting referrals to the
designer at the time of receiving the bug report was demonstrated by testing
the model on 10 projects.
|
[
{
"version": "v1",
"created": "Tue, 20 Sep 2022 08:43:40 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Sepahvand",
"Reza",
""
],
[
"Akbari",
"Reza",
""
],
[
"Jamasb",
"Behnaz",
""
],
[
"Hashemi",
"Sattar",
""
],
[
"Boushehrian",
"Omid",
""
]
] |
TITLE: Using Word Embedding and Convolution Neural Network for Bug Triaging by
Considering Design Flaws
ABSTRACT: Resolving bugs in the maintenance phase of software is a complicated task.
Bug assignment is one of the main tasks for resolving bugs. Some Bugs cannot be
fixed properly without making design decisions and have to be assigned to
designers, rather than programmers, to avoid emerging bad smells that may cause
subsequent bug reports. Hence, it is important to refer some bugs to the
designer to check the possible design flaws. Based on our best knowledge, there
are a few works that have considered referring bugs to designers. Hence, this
issue is considered in this work. In this paper, a dataset is created, and a
CNN-based model is proposed to predict the need for assigning a bug to a
designer by learning the peculiarities of bug reports effective in creating bad
smells in the code. The features of each bug are extracted from CNN based on
its textual features, such as a summary and description. The number of bad
samples added to it in the fixing process using the PMD tool determines the bug
tag. The summary and description of the new bug are given to the model and the
model predicts the need to refer to the designer. The accuracy of 75% (or more)
was achieved for datasets with a sufficient number of samples for deep
learning-based model training. A model is proposed to predict bug referrals to
the designer. The efficiency of the model in predicting referrals to the
designer at the time of receiving the bug report was demonstrated by testing
the model on 10 projects.
|
new_dataset
| 0.716417
|
2209.09699
|
Niclas V\"odisch
|
Jos\'e Arce, Niclas V\"odisch, Daniele Cattaneo, Wolfram Burgard,
Abhinav Valada
|
PADLoC: LiDAR-Based Deep Loop Closure Detection and Registration Using
Panoptic Attention
| null |
IEEE Robotics and Automation Letters, vol. 8, no. 3, pp.
1319-1326, March 2023
|
10.1109/LRA.2023.3239312
| null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A key component of graph-based SLAM systems is the ability to detect loop
closures in a trajectory to reduce the drift accumulated over time from the
odometry. Most LiDAR-based methods achieve this goal by using only the
geometric information, disregarding the semantics of the scene. In this work,
we introduce PADLoC for joint loop closure detection and registration in
LiDAR-based SLAM frameworks. We propose a novel transformer-based head for
point cloud matching and registration, and to leverage panoptic information
during training time. In particular, we propose a novel loss function that
reframes the matching problem as a classification task for the semantic labels
and as a graph connectivity assignment for the instance labels. During
inference, PADLoC does not require panoptic annotations, making it more
versatile than other methods. Additionally, we show that using two shared
matching and registration heads with their source and target inputs swapped
increases the overall performance by enforcing forward-backward consistency. We
perform extensive evaluations of PADLoC on multiple real-world datasets
demonstrating that it achieves state-of-the-art results. The code of our work
is publicly available at http://padloc.cs.uni-freiburg.de.
|
[
{
"version": "v1",
"created": "Tue, 20 Sep 2022 13:07:49 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jan 2023 10:17:55 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 15:26:15 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Arce",
"José",
""
],
[
"Vödisch",
"Niclas",
""
],
[
"Cattaneo",
"Daniele",
""
],
[
"Burgard",
"Wolfram",
""
],
[
"Valada",
"Abhinav",
""
]
] |
TITLE: PADLoC: LiDAR-Based Deep Loop Closure Detection and Registration Using
Panoptic Attention
ABSTRACT: A key component of graph-based SLAM systems is the ability to detect loop
closures in a trajectory to reduce the drift accumulated over time from the
odometry. Most LiDAR-based methods achieve this goal by using only the
geometric information, disregarding the semantics of the scene. In this work,
we introduce PADLoC for joint loop closure detection and registration in
LiDAR-based SLAM frameworks. We propose a novel transformer-based head for
point cloud matching and registration, and to leverage panoptic information
during training time. In particular, we propose a novel loss function that
reframes the matching problem as a classification task for the semantic labels
and as a graph connectivity assignment for the instance labels. During
inference, PADLoC does not require panoptic annotations, making it more
versatile than other methods. Additionally, we show that using two shared
matching and registration heads with their source and target inputs swapped
increases the overall performance by enforcing forward-backward consistency. We
perform extensive evaluations of PADLoC on multiple real-world datasets
demonstrating that it achieves state-of-the-art results. The code of our work
is publicly available at http://padloc.cs.uni-freiburg.de.
|
no_new_dataset
| 0.710666
|
2209.14547
|
Muhammad Akbar Husnoo
|
Muhammad Akbar Husnoo, Adnan Anwar, Nasser Hosseinzadeh, Shama Naz
Islam, Abdun Naser Mahmood and Robin Doss
|
A Secure Federated Learning Framework for Residential Short Term Load
Forecasting
| null | null | null | null |
cs.CR cs.AI cs.DC cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Smart meter measurements, though critical for accurate demand forecasting,
face several drawbacks including consumers' privacy, data breach issues, to
name a few. Recent literature has explored Federated Learning (FL) as a
promising privacy-preserving machine learning alternative which enables
collaborative learning of a model without exposing private raw data for short
term load forecasting. Despite its virtue, standard FL is still vulnerable to
an intractable cyber threat known as Byzantine attack carried out by faulty
and/or malicious clients. Therefore, to improve the robustness of federated
short-term load forecasting against Byzantine threats, we develop a
state-of-the-art differentially private secured FL-based framework that ensures
the privacy of the individual smart meter's data while protect the security of
FL models and architecture. Our proposed framework leverages the idea of
gradient quantization through the Sign Stochastic Gradient Descent (SignSGD)
algorithm, where the clients only transmit the `sign' of the gradient to the
control centre after local model training. As we highlight through our
experiments involving benchmark neural networks with a set of Byzantine attack
models, our proposed approach mitigates such threats quite effectively and thus
outperforms conventional Fed-SGD models.
|
[
{
"version": "v1",
"created": "Thu, 29 Sep 2022 04:36:16 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 05:09:56 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Husnoo",
"Muhammad Akbar",
""
],
[
"Anwar",
"Adnan",
""
],
[
"Hosseinzadeh",
"Nasser",
""
],
[
"Islam",
"Shama Naz",
""
],
[
"Mahmood",
"Abdun Naser",
""
],
[
"Doss",
"Robin",
""
]
] |
TITLE: A Secure Federated Learning Framework for Residential Short Term Load
Forecasting
ABSTRACT: Smart meter measurements, though critical for accurate demand forecasting,
face several drawbacks including consumers' privacy, data breach issues, to
name a few. Recent literature has explored Federated Learning (FL) as a
promising privacy-preserving machine learning alternative which enables
collaborative learning of a model without exposing private raw data for short
term load forecasting. Despite its virtue, standard FL is still vulnerable to
an intractable cyber threat known as Byzantine attack carried out by faulty
and/or malicious clients. Therefore, to improve the robustness of federated
short-term load forecasting against Byzantine threats, we develop a
state-of-the-art differentially private secured FL-based framework that ensures
the privacy of the individual smart meter's data while protect the security of
FL models and architecture. Our proposed framework leverages the idea of
gradient quantization through the Sign Stochastic Gradient Descent (SignSGD)
algorithm, where the clients only transmit the `sign' of the gradient to the
control centre after local model training. As we highlight through our
experiments involving benchmark neural networks with a set of Byzantine attack
models, our proposed approach mitigates such threats quite effectively and thus
outperforms conventional Fed-SGD models.
|
no_new_dataset
| 0.707645
|
2210.01033
|
Bowen Dong
|
Bowen Dong, Pan Zhou, Shuicheng Yan, Wangmeng Zuo
|
LPT: Long-tailed Prompt Tuning for Image Classification
|
ICLR 2023 (poster)
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
For long-tailed classification, most works often pretrain a big model on a
large-scale dataset, and then fine-tune the whole model for adapting to
long-tailed data. Though promising, fine-tuning the whole pretrained model
tends to suffer from high cost in computation and deployment of different
models for different tasks, as well as weakened generalization ability for
overfitting to certain features of long-tailed data. To alleviate these issues,
we propose an effective Long-tailed Prompt Tuning method for long-tailed
classification. LPT introduces several trainable prompts into a frozen
pretrained model to adapt it to long-tailed data. For better effectiveness, we
divide prompts into two groups: 1) a shared prompt for the whole long-tailed
dataset to learn general features and to adapt a pretrained model into target
domain; and 2) group-specific prompts to gather group-specific features for the
samples which have similar features and also to empower the pretrained model
with discrimination ability. Then we design a two-phase training paradigm to
learn these prompts. In phase 1, we train the shared prompt via supervised
prompt tuning to adapt a pretrained model to the desired long-tailed domain. In
phase 2, we use the learnt shared prompt as query to select a small best
matched set for a group of similar samples from the group-specific prompt set
to dig the common features of these similar samples, then optimize these
prompts with dual sampling strategy and asymmetric GCL loss. By only
fine-tuning a few prompts while fixing the pretrained model, LPT can reduce
training and deployment cost by storing a few prompts, and enjoys a strong
generalization ability of the pretrained model. Experiments show that on
various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves
comparable performance than previous whole model fine-tuning methods, and is
more robust to domain-shift.
|
[
{
"version": "v1",
"created": "Mon, 3 Oct 2022 15:47:02 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 10:16:03 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Dong",
"Bowen",
""
],
[
"Zhou",
"Pan",
""
],
[
"Yan",
"Shuicheng",
""
],
[
"Zuo",
"Wangmeng",
""
]
] |
TITLE: LPT: Long-tailed Prompt Tuning for Image Classification
ABSTRACT: For long-tailed classification, most works often pretrain a big model on a
large-scale dataset, and then fine-tune the whole model for adapting to
long-tailed data. Though promising, fine-tuning the whole pretrained model
tends to suffer from high cost in computation and deployment of different
models for different tasks, as well as weakened generalization ability for
overfitting to certain features of long-tailed data. To alleviate these issues,
we propose an effective Long-tailed Prompt Tuning method for long-tailed
classification. LPT introduces several trainable prompts into a frozen
pretrained model to adapt it to long-tailed data. For better effectiveness, we
divide prompts into two groups: 1) a shared prompt for the whole long-tailed
dataset to learn general features and to adapt a pretrained model into target
domain; and 2) group-specific prompts to gather group-specific features for the
samples which have similar features and also to empower the pretrained model
with discrimination ability. Then we design a two-phase training paradigm to
learn these prompts. In phase 1, we train the shared prompt via supervised
prompt tuning to adapt a pretrained model to the desired long-tailed domain. In
phase 2, we use the learnt shared prompt as query to select a small best
matched set for a group of similar samples from the group-specific prompt set
to dig the common features of these similar samples, then optimize these
prompts with dual sampling strategy and asymmetric GCL loss. By only
fine-tuning a few prompts while fixing the pretrained model, LPT can reduce
training and deployment cost by storing a few prompts, and enjoys a strong
generalization ability of the pretrained model. Experiments show that on
various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves
comparable performance than previous whole model fine-tuning methods, and is
more robust to domain-shift.
|
no_new_dataset
| 0.710409
|
2210.01612
|
Ruoyu Wang
|
Ruoyu Wang, Zehao Yu and Shenghua Gao
|
PlaneDepth: Self-supervised Depth Estimation via Orthogonal Planes
|
Accepted by CVPR 2023. Code and models are available at:
https://github.com/svip-lab/PlaneDepth
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Multiple near frontal-parallel planes based depth representation demonstrated
impressive results in self-supervised monocular depth estimation (MDE).
Whereas, such a representation would cause the discontinuity of the ground as
it is perpendicular to the frontal-parallel planes, which is detrimental to the
identification of drivable space in autonomous driving. In this paper, we
propose the PlaneDepth, a novel orthogonal planes based presentation, including
vertical planes and ground planes. PlaneDepth estimates the depth distribution
using a Laplacian Mixture Model based on orthogonal planes for an input image.
These planes are used to synthesize a reference view to provide the
self-supervision signal. Further, we find that the widely used resizing and
cropping data augmentation breaks the orthogonality assumptions, leading to
inferior plane predictions. We address this problem by explicitly constructing
the resizing cropping transformation to rectify the predefined planes and
predicted camera pose. Moreover, we propose an augmented self-distillation loss
supervised with a bilateral occlusion mask to boost the robustness of
orthogonal planes representation for occlusions. Thanks to our orthogonal
planes representation, we can extract the ground plane in an unsupervised
manner, which is important for autonomous driving. Extensive experiments on the
KITTI dataset demonstrate the effectiveness and efficiency of our method. The
code is available at https://github.com/svip-lab/PlaneDepth.
|
[
{
"version": "v1",
"created": "Tue, 4 Oct 2022 13:51:59 GMT"
},
{
"version": "v2",
"created": "Thu, 23 Mar 2023 10:01:33 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 05:06:59 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Ruoyu",
""
],
[
"Yu",
"Zehao",
""
],
[
"Gao",
"Shenghua",
""
]
] |
TITLE: PlaneDepth: Self-supervised Depth Estimation via Orthogonal Planes
ABSTRACT: Multiple near frontal-parallel planes based depth representation demonstrated
impressive results in self-supervised monocular depth estimation (MDE).
Whereas, such a representation would cause the discontinuity of the ground as
it is perpendicular to the frontal-parallel planes, which is detrimental to the
identification of drivable space in autonomous driving. In this paper, we
propose the PlaneDepth, a novel orthogonal planes based presentation, including
vertical planes and ground planes. PlaneDepth estimates the depth distribution
using a Laplacian Mixture Model based on orthogonal planes for an input image.
These planes are used to synthesize a reference view to provide the
self-supervision signal. Further, we find that the widely used resizing and
cropping data augmentation breaks the orthogonality assumptions, leading to
inferior plane predictions. We address this problem by explicitly constructing
the resizing cropping transformation to rectify the predefined planes and
predicted camera pose. Moreover, we propose an augmented self-distillation loss
supervised with a bilateral occlusion mask to boost the robustness of
orthogonal planes representation for occlusions. Thanks to our orthogonal
planes representation, we can extract the ground plane in an unsupervised
manner, which is important for autonomous driving. Extensive experiments on the
KITTI dataset demonstrate the effectiveness and efficiency of our method. The
code is available at https://github.com/svip-lab/PlaneDepth.
|
no_new_dataset
| 0.712076
|
2210.06207
|
Abdullatif K\"oksal
|
Abdullatif K\"oksal, Silvia Severini, Hinrich Sch\"utze
|
SilverAlign: MT-Based Silver Data Algorithm For Evaluating Word
Alignment
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Word alignments are essential for a variety of NLP tasks. Therefore, choosing
the best approaches for their creation is crucial. However, the scarce
availability of gold evaluation data makes the choice difficult. We propose
SilverAlign, a new method to automatically create silver data for the
evaluation of word aligners by exploiting machine translation and minimal
pairs. We show that performance on our silver data correlates well with gold
benchmarks for 9 language pairs, making our approach a valid resource for
evaluation of different domains and languages when gold data are not available.
This addresses the important scenario of missing gold data alignments for
low-resource languages.
|
[
{
"version": "v1",
"created": "Wed, 12 Oct 2022 13:48:59 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 22:00:44 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Köksal",
"Abdullatif",
""
],
[
"Severini",
"Silvia",
""
],
[
"Schütze",
"Hinrich",
""
]
] |
TITLE: SilverAlign: MT-Based Silver Data Algorithm For Evaluating Word
Alignment
ABSTRACT: Word alignments are essential for a variety of NLP tasks. Therefore, choosing
the best approaches for their creation is crucial. However, the scarce
availability of gold evaluation data makes the choice difficult. We propose
SilverAlign, a new method to automatically create silver data for the
evaluation of word aligners by exploiting machine translation and minimal
pairs. We show that performance on our silver data correlates well with gold
benchmarks for 9 language pairs, making our approach a valid resource for
evaluation of different domains and languages when gold data are not available.
This addresses the important scenario of missing gold data alignments for
low-resource languages.
|
no_new_dataset
| 0.708408
|
2210.08383
|
Shiro Kuriwaki
|
Christopher T. Kenny, Shiro Kuriwaki, Cory McCartan, Evan T. R.
Rosenman, Tyler Simko, Kosuke Imai
|
Comment: The Essential Role of Policy Evaluation for the 2020 Census
Disclosure Avoidance System
|
Version accepted to Harvard Data Science Review
|
Harvard Data Science Review, (Special Issue 2, 2023)
|
10.1162/99608f92.abc2c765
| null |
stat.AP cs.CY
|
http://creativecommons.org/licenses/by/4.0/
|
In "Differential Perspectives: Epistemic Disconnects Surrounding the US
Census Bureau's Use of Differential Privacy," boyd and Sarathy argue that
empirical evaluations of the Census Disclosure Avoidance System (DAS),
including our published analysis, failed to recognize how the benchmark data
against which the 2020 DAS was evaluated is never a ground truth of population
counts. In this commentary, we explain why policy evaluation, which was the
main goal of our analysis, is still meaningful without access to a perfect
ground truth. We also point out that our evaluation leveraged features specific
to the decennial Census and redistricting data, such as block-level population
invariance under swapping and voter file racial identification, better
approximating a comparison with the ground truth. Lastly, we show that accurate
statistical predictions of individual race based on the Bayesian Improved
Surname Geocoding, while not a violation of differential privacy, substantially
increases the disclosure risk of private information the Census Bureau sought
to protect. We conclude by arguing that policy makers must confront a key
trade-off between data utility and privacy protection, and an epistemic
disconnect alone is insufficient to explain disagreements between policy
choices.
|
[
{
"version": "v1",
"created": "Sat, 15 Oct 2022 21:41:54 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Kenny",
"Christopher T.",
""
],
[
"Kuriwaki",
"Shiro",
""
],
[
"McCartan",
"Cory",
""
],
[
"Rosenman",
"Evan T. R.",
""
],
[
"Simko",
"Tyler",
""
],
[
"Imai",
"Kosuke",
""
]
] |
TITLE: Comment: The Essential Role of Policy Evaluation for the 2020 Census
Disclosure Avoidance System
ABSTRACT: In "Differential Perspectives: Epistemic Disconnects Surrounding the US
Census Bureau's Use of Differential Privacy," boyd and Sarathy argue that
empirical evaluations of the Census Disclosure Avoidance System (DAS),
including our published analysis, failed to recognize how the benchmark data
against which the 2020 DAS was evaluated is never a ground truth of population
counts. In this commentary, we explain why policy evaluation, which was the
main goal of our analysis, is still meaningful without access to a perfect
ground truth. We also point out that our evaluation leveraged features specific
to the decennial Census and redistricting data, such as block-level population
invariance under swapping and voter file racial identification, better
approximating a comparison with the ground truth. Lastly, we show that accurate
statistical predictions of individual race based on the Bayesian Improved
Surname Geocoding, while not a violation of differential privacy, substantially
increases the disclosure risk of private information the Census Bureau sought
to protect. We conclude by arguing that policy makers must confront a key
trade-off between data utility and privacy protection, and an epistemic
disconnect alone is insufficient to explain disagreements between policy
choices.
|
no_new_dataset
| 0.710201
|
2211.01778
|
Hyungtae Lee
|
Yi-Ting Shen, Hyungtae Lee, Heesung Kwon, Shuvra Shikhar Bhattacharyya
|
Progressive Transformation Learning for Leveraging Virtual Images in
Training
|
CVPR 2023 (Selected as Highlight)
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
To effectively interrogate UAV-based images for detecting objects of
interest, such as humans, it is essential to acquire large-scale UAV-based
datasets that include human instances with various poses captured from widely
varying viewing angles. As a viable alternative to laborious and costly data
curation, we introduce Progressive Transformation Learning (PTL), which
gradually augments a training dataset by adding transformed virtual images with
enhanced realism. Generally, a virtual2real transformation generator in the
conditional GAN framework suffers from quality degradation when a large domain
gap exists between real and virtual images. To deal with the domain gap, PTL
takes a novel approach that progressively iterates the following three steps:
1) select a subset from a pool of virtual images according to the domain gap,
2) transform the selected virtual images to enhance realism, and 3) add the
transformed virtual images to the training set while removing them from the
pool. In PTL, accurately quantifying the domain gap is critical. To do that, we
theoretically demonstrate that the feature representation space of a given
object detector can be modeled as a multivariate Gaussian distribution from
which the Mahalanobis distance between a virtual object and the Gaussian
distribution of each object category in the representation space can be readily
computed. Experiments show that PTL results in a substantial performance
increase over the baseline, especially in the small data and the cross-domain
regime.
|
[
{
"version": "v1",
"created": "Thu, 3 Nov 2022 13:04:15 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 19:21:27 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Shen",
"Yi-Ting",
""
],
[
"Lee",
"Hyungtae",
""
],
[
"Kwon",
"Heesung",
""
],
[
"Bhattacharyya",
"Shuvra Shikhar",
""
]
] |
TITLE: Progressive Transformation Learning for Leveraging Virtual Images in
Training
ABSTRACT: To effectively interrogate UAV-based images for detecting objects of
interest, such as humans, it is essential to acquire large-scale UAV-based
datasets that include human instances with various poses captured from widely
varying viewing angles. As a viable alternative to laborious and costly data
curation, we introduce Progressive Transformation Learning (PTL), which
gradually augments a training dataset by adding transformed virtual images with
enhanced realism. Generally, a virtual2real transformation generator in the
conditional GAN framework suffers from quality degradation when a large domain
gap exists between real and virtual images. To deal with the domain gap, PTL
takes a novel approach that progressively iterates the following three steps:
1) select a subset from a pool of virtual images according to the domain gap,
2) transform the selected virtual images to enhance realism, and 3) add the
transformed virtual images to the training set while removing them from the
pool. In PTL, accurately quantifying the domain gap is critical. To do that, we
theoretically demonstrate that the feature representation space of a given
object detector can be modeled as a multivariate Gaussian distribution from
which the Mahalanobis distance between a virtual object and the Gaussian
distribution of each object category in the representation space can be readily
computed. Experiments show that PTL results in a substantial performance
increase over the baseline, especially in the small data and the cross-domain
regime.
|
no_new_dataset
| 0.710176
|
2211.03128
|
Giuseppe Vietri
|
Travis Dick, Cynthia Dwork, Michael Kearns, Terrance Liu, Aaron Roth,
Giuseppe Vietri, Zhiwei Steven Wu
|
Confidence-Ranked Reconstruction of Census Microdata from Published
Statistics
| null | null |
10.1073/pnas.2218605120
| null |
cs.CY cs.CR cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
A reconstruction attack on a private dataset $D$ takes as input some publicly
accessible information about the dataset and produces a list of candidate
elements of $D$. We introduce a new class of data reconstruction attacks based
on randomized methods for non-convex optimization. We empirically demonstrate
that our attacks can not only reconstruct full rows of $D$ from aggregate query
statistics $Q(D)\in \mathbb{R}^m$, but can do so in a way that reliably ranks
reconstructed rows by their odds of appearing in the private data, providing a
signature that could be used for prioritizing reconstructed rows for further
actions such as identify theft or hate crime. We also design a sequence of
baselines for evaluating reconstruction attacks. Our attacks significantly
outperform those that are based only on access to a public distribution or
population from which the private dataset $D$ was sampled, demonstrating that
they are exploiting information in the aggregate statistics $Q(D)$, and not
simply the overall structure of the distribution. In other words, the queries
$Q(D)$ are permitting reconstruction of elements of this dataset, not the
distribution from which $D$ was drawn. These findings are established both on
2010 U.S. decennial Census data and queries and Census-derived American
Community Survey datasets. Taken together, our methods and experiments
illustrate the risks in releasing numerically precise aggregate statistics of a
large dataset, and provide further motivation for the careful application of
provably private techniques such as differential privacy.
|
[
{
"version": "v1",
"created": "Sun, 6 Nov 2022 14:08:43 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Feb 2023 17:32:02 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Dick",
"Travis",
""
],
[
"Dwork",
"Cynthia",
""
],
[
"Kearns",
"Michael",
""
],
[
"Liu",
"Terrance",
""
],
[
"Roth",
"Aaron",
""
],
[
"Vietri",
"Giuseppe",
""
],
[
"Wu",
"Zhiwei Steven",
""
]
] |
TITLE: Confidence-Ranked Reconstruction of Census Microdata from Published
Statistics
ABSTRACT: A reconstruction attack on a private dataset $D$ takes as input some publicly
accessible information about the dataset and produces a list of candidate
elements of $D$. We introduce a new class of data reconstruction attacks based
on randomized methods for non-convex optimization. We empirically demonstrate
that our attacks can not only reconstruct full rows of $D$ from aggregate query
statistics $Q(D)\in \mathbb{R}^m$, but can do so in a way that reliably ranks
reconstructed rows by their odds of appearing in the private data, providing a
signature that could be used for prioritizing reconstructed rows for further
actions such as identify theft or hate crime. We also design a sequence of
baselines for evaluating reconstruction attacks. Our attacks significantly
outperform those that are based only on access to a public distribution or
population from which the private dataset $D$ was sampled, demonstrating that
they are exploiting information in the aggregate statistics $Q(D)$, and not
simply the overall structure of the distribution. In other words, the queries
$Q(D)$ are permitting reconstruction of elements of this dataset, not the
distribution from which $D$ was drawn. These findings are established both on
2010 U.S. decennial Census data and queries and Census-derived American
Community Survey datasets. Taken together, our methods and experiments
illustrate the risks in releasing numerically precise aggregate statistics of a
large dataset, and provide further motivation for the careful application of
provably private techniques such as differential privacy.
|
no_new_dataset
| 0.702492
|
2211.11592
|
Nando Metzger
|
Nando Metzger, Rodrigo Caye Daudt, Konrad Schindler
|
Guided Depth Super-Resolution by Deep Anisotropic Diffusion
|
8 main pages, Accepted to CVPR2023
| null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Performing super-resolution of a depth image using the guidance from an RGB
image is a problem that concerns several fields, such as robotics, medical
imaging, and remote sensing. While deep learning methods have achieved good
results in this problem, recent work highlighted the value of combining modern
methods with more formal frameworks. In this work, we propose a novel approach
which combines guided anisotropic diffusion with a deep convolutional network
and advances the state of the art for guided depth super-resolution. The edge
transferring/enhancing properties of the diffusion are boosted by the
contextual reasoning capabilities of modern networks, and a strict adjustment
step guarantees perfect adherence to the source image. We achieve unprecedented
results in three commonly used benchmarks for guided depth super-resolution.
The performance gain compared to other methods is the largest at larger scales,
such as x32 scaling. Code
(https://github.com/prs-eth/Diffusion-Super-Resolution) for the proposed method
is available to promote reproducibility of our results.
|
[
{
"version": "v1",
"created": "Mon, 21 Nov 2022 15:48:13 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Nov 2022 09:58:10 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 11:31:08 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Metzger",
"Nando",
""
],
[
"Daudt",
"Rodrigo Caye",
""
],
[
"Schindler",
"Konrad",
""
]
] |
TITLE: Guided Depth Super-Resolution by Deep Anisotropic Diffusion
ABSTRACT: Performing super-resolution of a depth image using the guidance from an RGB
image is a problem that concerns several fields, such as robotics, medical
imaging, and remote sensing. While deep learning methods have achieved good
results in this problem, recent work highlighted the value of combining modern
methods with more formal frameworks. In this work, we propose a novel approach
which combines guided anisotropic diffusion with a deep convolutional network
and advances the state of the art for guided depth super-resolution. The edge
transferring/enhancing properties of the diffusion are boosted by the
contextual reasoning capabilities of modern networks, and a strict adjustment
step guarantees perfect adherence to the source image. We achieve unprecedented
results in three commonly used benchmarks for guided depth super-resolution.
The performance gain compared to other methods is the largest at larger scales,
such as x32 scaling. Code
(https://github.com/prs-eth/Diffusion-Super-Resolution) for the proposed method
is available to promote reproducibility of our results.
|
no_new_dataset
| 0.710226
|
2211.11734
|
Karthik Shetty
|
Karthik Shetty, Annette Birkhold, Srikrishna Jaganathan, Norbert
Strobel, Markus Kowarschik, Andreas Maier, Bernhard Egger
|
PLIKS: A Pseudo-Linear Inverse Kinematic Solver for 3D Human Body
Estimation
|
CVPR2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We introduce PLIKS (Pseudo-Linear Inverse Kinematic Solver) for
reconstruction of a 3D mesh of the human body from a single 2D image. Current
techniques directly regress the shape, pose, and translation of a parametric
model from an input image through a non-linear mapping with minimal flexibility
to any external influences. We approach the task as a model-in-the-loop
optimization problem. PLIKS is built on a linearized formulation of the
parametric SMPL model. Using PLIKS, we can analytically reconstruct the human
model via 2D pixel-aligned vertices. This enables us with the flexibility to
use accurate camera calibration information when available. PLIKS offers an
easy way to introduce additional constraints such as shape and translation. We
present quantitative evaluations which confirm that PLIKS achieves more
accurate reconstruction with greater than 10% improvement compared to other
state-of-the-art methods with respect to the standard 3D human pose and shape
benchmarks while also obtaining a reconstruction error improvement of 12.9 mm
on the newer AGORA dataset.
|
[
{
"version": "v1",
"created": "Mon, 21 Nov 2022 18:54:12 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 23:50:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Shetty",
"Karthik",
""
],
[
"Birkhold",
"Annette",
""
],
[
"Jaganathan",
"Srikrishna",
""
],
[
"Strobel",
"Norbert",
""
],
[
"Kowarschik",
"Markus",
""
],
[
"Maier",
"Andreas",
""
],
[
"Egger",
"Bernhard",
""
]
] |
TITLE: PLIKS: A Pseudo-Linear Inverse Kinematic Solver for 3D Human Body
Estimation
ABSTRACT: We introduce PLIKS (Pseudo-Linear Inverse Kinematic Solver) for
reconstruction of a 3D mesh of the human body from a single 2D image. Current
techniques directly regress the shape, pose, and translation of a parametric
model from an input image through a non-linear mapping with minimal flexibility
to any external influences. We approach the task as a model-in-the-loop
optimization problem. PLIKS is built on a linearized formulation of the
parametric SMPL model. Using PLIKS, we can analytically reconstruct the human
model via 2D pixel-aligned vertices. This enables us with the flexibility to
use accurate camera calibration information when available. PLIKS offers an
easy way to introduce additional constraints such as shape and translation. We
present quantitative evaluations which confirm that PLIKS achieves more
accurate reconstruction with greater than 10% improvement compared to other
state-of-the-art methods with respect to the standard 3D human pose and shape
benchmarks while also obtaining a reconstruction error improvement of 12.9 mm
on the newer AGORA dataset.
|
no_new_dataset
| 0.708055
|
2211.12562
|
Radu Alexandru Rosu
|
Radu Alexandru Rosu and Sven Behnke
|
PermutoSDF: Fast Multi-View Reconstruction with Implicit Surfaces using
Permutohedral Lattices
|
Accepted to CVPR 2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Neural radiance-density field methods have become increasingly popular for
the task of novel-view rendering. Their recent extension to hash-based
positional encoding ensures fast training and inference with visually pleasing
results. However, density-based methods struggle with recovering accurate
surface geometry. Hybrid methods alleviate this issue by optimizing the density
based on an underlying SDF. However, current SDF methods are overly smooth and
miss fine geometric details. In this work, we combine the strengths of these
two lines of work in a novel hash-based implicit surface representation. We
propose improvements to the two areas by replacing the voxel hash encoding with
a permutohedral lattice which optimizes faster, especially for higher
dimensions. We additionally propose a regularization scheme which is crucial
for recovering high-frequency geometric detail. We evaluate our method on
multiple datasets and show that we can recover geometric detail at the level of
pores and wrinkles while using only RGB images for supervision. Furthermore,
using sphere tracing we can render novel views at 30 fps on an RTX 3090. Code
is publicly available at: https://radualexandru.github.io/permuto_sdf
|
[
{
"version": "v1",
"created": "Tue, 22 Nov 2022 20:27:44 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 15:12:29 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Rosu",
"Radu Alexandru",
""
],
[
"Behnke",
"Sven",
""
]
] |
TITLE: PermutoSDF: Fast Multi-View Reconstruction with Implicit Surfaces using
Permutohedral Lattices
ABSTRACT: Neural radiance-density field methods have become increasingly popular for
the task of novel-view rendering. Their recent extension to hash-based
positional encoding ensures fast training and inference with visually pleasing
results. However, density-based methods struggle with recovering accurate
surface geometry. Hybrid methods alleviate this issue by optimizing the density
based on an underlying SDF. However, current SDF methods are overly smooth and
miss fine geometric details. In this work, we combine the strengths of these
two lines of work in a novel hash-based implicit surface representation. We
propose improvements to the two areas by replacing the voxel hash encoding with
a permutohedral lattice which optimizes faster, especially for higher
dimensions. We additionally propose a regularization scheme which is crucial
for recovering high-frequency geometric detail. We evaluate our method on
multiple datasets and show that we can recover geometric detail at the level of
pores and wrinkles while using only RGB images for supervision. Furthermore,
using sphere tracing we can render novel views at 30 fps on an RTX 3090. Code
is publicly available at: https://radualexandru.github.io/permuto_sdf
|
no_new_dataset
| 0.710603
|
2211.15658
|
Yuanwen Yue
|
Yuanwen Yue, Theodora Kontogianni, Konrad Schindler, Francis Engelmann
|
Connecting the Dots: Floorplan Reconstruction Using Two-Level Queries
|
CVPR 2023 camera-ready. Project page:
https://ywyue.github.io/RoomFormer
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
We address 2D floorplan reconstruction from 3D scans. Existing approaches
typically employ heuristically designed multi-stage pipelines. Instead, we
formulate floorplan reconstruction as a single-stage structured prediction
task: find a variable-size set of polygons, which in turn are variable-length
sequences of ordered vertices. To solve it we develop a novel Transformer
architecture that generates polygons of multiple rooms in parallel, in a
holistic manner without hand-crafted intermediate stages. The model features
two-level queries for polygons and corners, and includes polygon matching to
make the network end-to-end trainable. Our method achieves a new
state-of-the-art for two challenging datasets, Structured3D and SceneCAD, along
with significantly faster inference than previous methods. Moreover, it can
readily be extended to predict additional information, i.e., semantic room
types and architectural elements like doors and windows. Our code and models
are available at: https://github.com/ywyue/RoomFormer.
|
[
{
"version": "v1",
"created": "Mon, 28 Nov 2022 18:59:09 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 02:20:16 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Yue",
"Yuanwen",
""
],
[
"Kontogianni",
"Theodora",
""
],
[
"Schindler",
"Konrad",
""
],
[
"Engelmann",
"Francis",
""
]
] |
TITLE: Connecting the Dots: Floorplan Reconstruction Using Two-Level Queries
ABSTRACT: We address 2D floorplan reconstruction from 3D scans. Existing approaches
typically employ heuristically designed multi-stage pipelines. Instead, we
formulate floorplan reconstruction as a single-stage structured prediction
task: find a variable-size set of polygons, which in turn are variable-length
sequences of ordered vertices. To solve it we develop a novel Transformer
architecture that generates polygons of multiple rooms in parallel, in a
holistic manner without hand-crafted intermediate stages. The model features
two-level queries for polygons and corners, and includes polygon matching to
make the network end-to-end trainable. Our method achieves a new
state-of-the-art for two challenging datasets, Structured3D and SceneCAD, along
with significantly faster inference than previous methods. Moreover, it can
readily be extended to predict additional information, i.e., semantic room
types and architectural elements like doors and windows. Our code and models
are available at: https://github.com/ywyue/RoomFormer.
|
no_new_dataset
| 0.710069
|
2211.16764
|
Andong Li
|
Andong Li, Guochen Yu, Chengshi Zheng, Wenzhe Liu, and Xiaodong Li
|
A General Unfolding Speech Enhancement Method Motivated by Taylor's
Theorem
|
Submitted to TASLP, revised version, 17 pages
| null | null | null |
cs.SD eess.AS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
While deep neural networks have facilitated significant advancements in the
field of speech enhancement, most existing methods are developed following
either empirical or relatively blind criteria, lacking adequate guidelines in
pipeline design. Inspired by Taylor's theorem, we propose a general unfolding
framework for both single- and multi-channel speech enhancement tasks.
Concretely, we formulate the complex spectrum recovery into the spectral
magnitude mapping in the neighborhood space of the noisy mixture, in which an
unknown sparse term is introduced and applied for phase modification in
advance. Based on that, the mapping function is decomposed into the
superimposition of the 0th-order and high-order polynomials in Taylor's series,
where the former coarsely removes the interference in the magnitude domain and
the latter progressively complements the remaining spectral detail in the
complex spectrum domain. In addition, we study the relation between adjacent
order terms and reveal that each high-order term can be recursively estimated
with its lower-order term, and each high-order term is then proposed to
evaluate using a surrogate function with trainable weights so that the whole
system can be trained in an end-to-end manner. Given that the proposed
framework is devised based on Taylor's theorem, it possesses improved internal
flexibility. Extensive experiments are conducted on WSJ0-SI84, DNS-Challenge,
Voicebank+Demand, spatialized Librispeech, and L3DAS22 multi-channel speech
enhancement challenge datasets. Quantitative results show that the proposed
approach yields competitive performance over existing top-performing approaches
in terms of multiple objective metrics.
|
[
{
"version": "v1",
"created": "Wed, 30 Nov 2022 06:05:59 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 04:33:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Li",
"Andong",
""
],
[
"Yu",
"Guochen",
""
],
[
"Zheng",
"Chengshi",
""
],
[
"Liu",
"Wenzhe",
""
],
[
"Li",
"Xiaodong",
""
]
] |
TITLE: A General Unfolding Speech Enhancement Method Motivated by Taylor's
Theorem
ABSTRACT: While deep neural networks have facilitated significant advancements in the
field of speech enhancement, most existing methods are developed following
either empirical or relatively blind criteria, lacking adequate guidelines in
pipeline design. Inspired by Taylor's theorem, we propose a general unfolding
framework for both single- and multi-channel speech enhancement tasks.
Concretely, we formulate the complex spectrum recovery into the spectral
magnitude mapping in the neighborhood space of the noisy mixture, in which an
unknown sparse term is introduced and applied for phase modification in
advance. Based on that, the mapping function is decomposed into the
superimposition of the 0th-order and high-order polynomials in Taylor's series,
where the former coarsely removes the interference in the magnitude domain and
the latter progressively complements the remaining spectral detail in the
complex spectrum domain. In addition, we study the relation between adjacent
order terms and reveal that each high-order term can be recursively estimated
with its lower-order term, and each high-order term is then proposed to
evaluate using a surrogate function with trainable weights so that the whole
system can be trained in an end-to-end manner. Given that the proposed
framework is devised based on Taylor's theorem, it possesses improved internal
flexibility. Extensive experiments are conducted on WSJ0-SI84, DNS-Challenge,
Voicebank+Demand, spatialized Librispeech, and L3DAS22 multi-channel speech
enhancement challenge datasets. Quantitative results show that the proposed
approach yields competitive performance over existing top-performing approaches
in terms of multiple objective metrics.
|
no_new_dataset
| 0.709422
|
2212.02573
|
Zhipeng Du
|
Zhipeng Du, Jiankang Deng, Miaojing Shi
|
Domain-General Crowd Counting in Unseen Scenarios
|
Accepted to AAAI 2023 as Oral Presentation
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Domain shift across crowd data severely hinders crowd counting models to
generalize to unseen scenarios. Although domain adaptive crowd counting
approaches close this gap to a certain extent, they are still dependent on the
target domain data to adapt (e.g. finetune) their models to the specific
domain. In this paper, we aim to train a model based on a single source domain
which can generalize well on any unseen domain. This falls into the realm of
domain generalization that remains unexplored in crowd counting. We first
introduce a dynamic sub-domain division scheme which divides the source domain
into multiple sub-domains such that we can initiate a meta-learning framework
for domain generalization. The sub-domain division is dynamically refined
during the meta-learning. Next, in order to disentangle domain-invariant
information from domain-specific information in image features, we design the
domain-invariant and -specific crowd memory modules to re-encode image
features. Two types of losses, i.e. feature reconstruction and orthogonal
losses, are devised to enable this disentanglement. Extensive experiments on
several standard crowd counting benchmarks i.e. SHA, SHB, QNRF, and NWPU, show
the strong generalizability of our method.
|
[
{
"version": "v1",
"created": "Mon, 5 Dec 2022 19:52:28 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 10:05:20 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Du",
"Zhipeng",
""
],
[
"Deng",
"Jiankang",
""
],
[
"Shi",
"Miaojing",
""
]
] |
TITLE: Domain-General Crowd Counting in Unseen Scenarios
ABSTRACT: Domain shift across crowd data severely hinders crowd counting models to
generalize to unseen scenarios. Although domain adaptive crowd counting
approaches close this gap to a certain extent, they are still dependent on the
target domain data to adapt (e.g. finetune) their models to the specific
domain. In this paper, we aim to train a model based on a single source domain
which can generalize well on any unseen domain. This falls into the realm of
domain generalization that remains unexplored in crowd counting. We first
introduce a dynamic sub-domain division scheme which divides the source domain
into multiple sub-domains such that we can initiate a meta-learning framework
for domain generalization. The sub-domain division is dynamically refined
during the meta-learning. Next, in order to disentangle domain-invariant
information from domain-specific information in image features, we design the
domain-invariant and -specific crowd memory modules to re-encode image
features. Two types of losses, i.e. feature reconstruction and orthogonal
losses, are devised to enable this disentanglement. Extensive experiments on
several standard crowd counting benchmarks i.e. SHA, SHB, QNRF, and NWPU, show
the strong generalizability of our method.
|
no_new_dataset
| 0.709849
|
2212.06200
|
Pavel Tokmakov
|
Pavel Tokmakov, Jie Li, Adrien Gaidon
|
Breaking the "Object" in Video Object Segmentation
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
The appearance of an object can be fleeting when it transforms. As eggs are
broken or paper is torn, their color, shape and texture can change
dramatically, preserving virtually nothing of the original except for the
identity itself. Yet, this important phenomenon is largely absent from existing
video object segmentation (VOS) benchmarks. In this work, we close the gap by
collecting a new dataset for Video Object Segmentation under Transformations
(VOST). It consists of more than 700 high-resolution videos, captured in
diverse environments, which are 21 seconds long on average and densely labeled
with instance masks. A careful, multi-step approach is adopted to ensure that
these videos focus on complex object transformations, capturing their full
temporal extent. We then extensively evaluate state-of-the-art VOS methods and
make a number of important discoveries. In particular, we show that existing
methods struggle when applied to this novel task and that their main limitation
lies in over-reliance on static appearance cues. This motivates us to propose a
few modifications for the top-performing baseline that improve its capabilities
by better modeling spatio-temporal information. But more broadly, the hope is
to stimulate discussion on learning more robust video object representations.
|
[
{
"version": "v1",
"created": "Mon, 12 Dec 2022 19:22:17 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 16:51:28 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Tokmakov",
"Pavel",
""
],
[
"Li",
"Jie",
""
],
[
"Gaidon",
"Adrien",
""
]
] |
TITLE: Breaking the "Object" in Video Object Segmentation
ABSTRACT: The appearance of an object can be fleeting when it transforms. As eggs are
broken or paper is torn, their color, shape and texture can change
dramatically, preserving virtually nothing of the original except for the
identity itself. Yet, this important phenomenon is largely absent from existing
video object segmentation (VOS) benchmarks. In this work, we close the gap by
collecting a new dataset for Video Object Segmentation under Transformations
(VOST). It consists of more than 700 high-resolution videos, captured in
diverse environments, which are 21 seconds long on average and densely labeled
with instance masks. A careful, multi-step approach is adopted to ensure that
these videos focus on complex object transformations, capturing their full
temporal extent. We then extensively evaluate state-of-the-art VOS methods and
make a number of important discoveries. In particular, we show that existing
methods struggle when applied to this novel task and that their main limitation
lies in over-reliance on static appearance cues. This motivates us to propose a
few modifications for the top-performing baseline that improve its capabilities
by better modeling spatio-temporal information. But more broadly, the hope is
to stimulate discussion on learning more robust video object representations.
|
new_dataset
| 0.714404
|
2212.07766
|
Remi Pautrat
|
R\'emi Pautrat, Daniel Barath, Viktor Larsson, Martin R. Oswald, Marc
Pollefeys
|
DeepLSD: Line Segment Detection and Refinement with Deep Image Gradients
|
Accepted at CVPR 2023
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Line segments are ubiquitous in our human-made world and are increasingly
used in vision tasks. They are complementary to feature points thanks to their
spatial extent and the structural information they provide. Traditional line
detectors based on the image gradient are extremely fast and accurate, but lack
robustness in noisy images and challenging conditions. Their learned
counterparts are more repeatable and can handle challenging images, but at the
cost of a lower accuracy and a bias towards wireframe lines. We propose to
combine traditional and learned approaches to get the best of both worlds: an
accurate and robust line detector that can be trained in the wild without
ground truth lines. Our new line segment detector, DeepLSD, processes images
with a deep network to generate a line attraction field, before converting it
to a surrogate image gradient magnitude and angle, which is then fed to any
existing handcrafted line detector. Additionally, we propose a new optimization
tool to refine line segments based on the attraction field and vanishing
points. This refinement improves the accuracy of current deep detectors by a
large margin. We demonstrate the performance of our method on low-level line
detection metrics, as well as on several downstream tasks using multiple
challenging datasets. The source code and models are available at
https://github.com/cvg/DeepLSD.
|
[
{
"version": "v1",
"created": "Thu, 15 Dec 2022 12:36:49 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 11:04:53 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 13:59:47 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Pautrat",
"Rémi",
""
],
[
"Barath",
"Daniel",
""
],
[
"Larsson",
"Viktor",
""
],
[
"Oswald",
"Martin R.",
""
],
[
"Pollefeys",
"Marc",
""
]
] |
TITLE: DeepLSD: Line Segment Detection and Refinement with Deep Image Gradients
ABSTRACT: Line segments are ubiquitous in our human-made world and are increasingly
used in vision tasks. They are complementary to feature points thanks to their
spatial extent and the structural information they provide. Traditional line
detectors based on the image gradient are extremely fast and accurate, but lack
robustness in noisy images and challenging conditions. Their learned
counterparts are more repeatable and can handle challenging images, but at the
cost of a lower accuracy and a bias towards wireframe lines. We propose to
combine traditional and learned approaches to get the best of both worlds: an
accurate and robust line detector that can be trained in the wild without
ground truth lines. Our new line segment detector, DeepLSD, processes images
with a deep network to generate a line attraction field, before converting it
to a surrogate image gradient magnitude and angle, which is then fed to any
existing handcrafted line detector. Additionally, we propose a new optimization
tool to refine line segments based on the attraction field and vanishing
points. This refinement improves the accuracy of current deep detectors by a
large margin. We demonstrate the performance of our method on low-level line
detection metrics, as well as on several downstream tasks using multiple
challenging datasets. The source code and models are available at
https://github.com/cvg/DeepLSD.
|
no_new_dataset
| 0.705189
|
2212.08420
|
Mert B\"ulent Sar{\i}y{\i}ld{\i}z
|
Mert Bulent Sariyildiz, Karteek Alahari, Diane Larlus, Yannis
Kalantidis
|
Fake it till you make it: Learning transferable representations from
synthetic ImageNet clones
|
Accepted to CVPR 2023
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent image generation models such as Stable Diffusion have exhibited an
impressive ability to generate fairly realistic images starting from a simple
text prompt. Could such models render real images obsolete for training image
prediction models? In this paper, we answer part of this provocative question
by investigating the need for real images when training models for ImageNet
classification. Provided only with the class names that have been used to build
the dataset, we explore the ability of Stable Diffusion to generate synthetic
clones of ImageNet and measure how useful these are for training classification
models from scratch. We show that with minimal and class-agnostic prompt
engineering, ImageNet clones are able to close a large part of the gap between
models produced by synthetic images and models trained with real images, for
the several standard classification benchmarks that we consider in this study.
More importantly, we show that models trained on synthetic images exhibit
strong generalization properties and perform on par with models trained on real
data for transfer. Project page: https://europe.naverlabs.com/imagenet-sd/
|
[
{
"version": "v1",
"created": "Fri, 16 Dec 2022 11:44:01 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 08:11:53 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Sariyildiz",
"Mert Bulent",
""
],
[
"Alahari",
"Karteek",
""
],
[
"Larlus",
"Diane",
""
],
[
"Kalantidis",
"Yannis",
""
]
] |
TITLE: Fake it till you make it: Learning transferable representations from
synthetic ImageNet clones
ABSTRACT: Recent image generation models such as Stable Diffusion have exhibited an
impressive ability to generate fairly realistic images starting from a simple
text prompt. Could such models render real images obsolete for training image
prediction models? In this paper, we answer part of this provocative question
by investigating the need for real images when training models for ImageNet
classification. Provided only with the class names that have been used to build
the dataset, we explore the ability of Stable Diffusion to generate synthetic
clones of ImageNet and measure how useful these are for training classification
models from scratch. We show that with minimal and class-agnostic prompt
engineering, ImageNet clones are able to close a large part of the gap between
models produced by synthetic images and models trained with real images, for
the several standard classification benchmarks that we consider in this study.
More importantly, we show that models trained on synthetic images exhibit
strong generalization properties and perform on par with models trained on real
data for transfer. Project page: https://europe.naverlabs.com/imagenet-sd/
|
no_new_dataset
| 0.687237
|
2301.00184
|
Wenhao Wu
|
Wenhao Wu, Haipeng Luo, Bo Fang, Jingdong Wang, Wanli Ouyang
|
Cap4Video: What Can Auxiliary Captions Do for Text-Video Retrieval?
|
Accepted by CVPR 2023. Selected as a Highlight (Top 2.5% of ALL
submissions)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Most existing text-video retrieval methods focus on cross-modal matching
between the visual content of videos and textual query sentences. However, in
real-world scenarios, online videos are often accompanied by relevant text
information such as titles, tags, and even subtitles, which can be utilized to
match textual queries. This insight has motivated us to propose a novel
approach to text-video retrieval, where we directly generate associated
captions from videos using zero-shot video captioning with knowledge from
web-scale pre-trained models (e.g., CLIP and GPT-2). Given the generated
captions, a natural question arises: what benefits do they bring to text-video
retrieval? To answer this, we introduce Cap4Video, a new framework that
leverages captions in three ways: i) Input data: video-caption pairs can
augment the training data. ii) Intermediate feature interaction: we perform
cross-modal feature interaction between the video and caption to produce
enhanced video representations. iii) Output score: the Query-Caption matching
branch can complement the original Query-Video matching branch for text-video
retrieval. We conduct comprehensive ablation studies to demonstrate the
effectiveness of our approach. Without any post-processing, Cap4Video achieves
state-of-the-art performance on four standard text-video retrieval benchmarks:
MSR-VTT (51.4%), VATEX (66.6%), MSVD (51.8%), and DiDeMo (52.0%). The code is
available at https://github.com/whwu95/Cap4Video .
|
[
{
"version": "v1",
"created": "Sat, 31 Dec 2022 11:50:32 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 09:39:45 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 07:09:59 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wu",
"Wenhao",
""
],
[
"Luo",
"Haipeng",
""
],
[
"Fang",
"Bo",
""
],
[
"Wang",
"Jingdong",
""
],
[
"Ouyang",
"Wanli",
""
]
] |
TITLE: Cap4Video: What Can Auxiliary Captions Do for Text-Video Retrieval?
ABSTRACT: Most existing text-video retrieval methods focus on cross-modal matching
between the visual content of videos and textual query sentences. However, in
real-world scenarios, online videos are often accompanied by relevant text
information such as titles, tags, and even subtitles, which can be utilized to
match textual queries. This insight has motivated us to propose a novel
approach to text-video retrieval, where we directly generate associated
captions from videos using zero-shot video captioning with knowledge from
web-scale pre-trained models (e.g., CLIP and GPT-2). Given the generated
captions, a natural question arises: what benefits do they bring to text-video
retrieval? To answer this, we introduce Cap4Video, a new framework that
leverages captions in three ways: i) Input data: video-caption pairs can
augment the training data. ii) Intermediate feature interaction: we perform
cross-modal feature interaction between the video and caption to produce
enhanced video representations. iii) Output score: the Query-Caption matching
branch can complement the original Query-Video matching branch for text-video
retrieval. We conduct comprehensive ablation studies to demonstrate the
effectiveness of our approach. Without any post-processing, Cap4Video achieves
state-of-the-art performance on four standard text-video retrieval benchmarks:
MSR-VTT (51.4%), VATEX (66.6%), MSVD (51.8%), and DiDeMo (52.0%). The code is
available at https://github.com/whwu95/Cap4Video .
|
no_new_dataset
| 0.711243
|
2301.02371
|
Shaofei Huang
|
Shaofei Huang, Zhenwei Shen, Zehao Huang, Zi-han Ding, Jiao Dai,
Jizhong Han, Naiyan Wang, Si Liu
|
Anchor3DLane: Learning to Regress 3D Anchors for Monocular 3D Lane
Detection
|
Accepted by CVPR 2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Monocular 3D lane detection is a challenging task due to its lack of depth
information. A popular solution is to first transform the front-viewed (FV)
images or features into the bird-eye-view (BEV) space with inverse perspective
mapping (IPM) and detect lanes from BEV features. However, the reliance of IPM
on flat ground assumption and loss of context information make it inaccurate to
restore 3D information from BEV representations. An attempt has been made to
get rid of BEV and predict 3D lanes from FV representations directly, while it
still underperforms other BEV-based methods given its lack of structured
representation for 3D lanes. In this paper, we define 3D lane anchors in the 3D
space and propose a BEV-free method named Anchor3DLane to predict 3D lanes
directly from FV representations. 3D lane anchors are projected to the FV
features to extract their features which contain both good structural and
context information to make accurate predictions. In addition, we also develop
a global optimization method that makes use of the equal-width property between
lanes to reduce the lateral error of predictions. Extensive experiments on
three popular 3D lane detection benchmarks show that our Anchor3DLane
outperforms previous BEV-based methods and achieves state-of-the-art
performances. The code is available at:
https://github.com/tusen-ai/Anchor3DLane.
|
[
{
"version": "v1",
"created": "Fri, 6 Jan 2023 04:35:31 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 04:28:30 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Huang",
"Shaofei",
""
],
[
"Shen",
"Zhenwei",
""
],
[
"Huang",
"Zehao",
""
],
[
"Ding",
"Zi-han",
""
],
[
"Dai",
"Jiao",
""
],
[
"Han",
"Jizhong",
""
],
[
"Wang",
"Naiyan",
""
],
[
"Liu",
"Si",
""
]
] |
TITLE: Anchor3DLane: Learning to Regress 3D Anchors for Monocular 3D Lane
Detection
ABSTRACT: Monocular 3D lane detection is a challenging task due to its lack of depth
information. A popular solution is to first transform the front-viewed (FV)
images or features into the bird-eye-view (BEV) space with inverse perspective
mapping (IPM) and detect lanes from BEV features. However, the reliance of IPM
on flat ground assumption and loss of context information make it inaccurate to
restore 3D information from BEV representations. An attempt has been made to
get rid of BEV and predict 3D lanes from FV representations directly, while it
still underperforms other BEV-based methods given its lack of structured
representation for 3D lanes. In this paper, we define 3D lane anchors in the 3D
space and propose a BEV-free method named Anchor3DLane to predict 3D lanes
directly from FV representations. 3D lane anchors are projected to the FV
features to extract their features which contain both good structural and
context information to make accurate predictions. In addition, we also develop
a global optimization method that makes use of the equal-width property between
lanes to reduce the lateral error of predictions. Extensive experiments on
three popular 3D lane detection benchmarks show that our Anchor3DLane
outperforms previous BEV-based methods and achieves state-of-the-art
performances. The code is available at:
https://github.com/tusen-ai/Anchor3DLane.
|
no_new_dataset
| 0.709629
|
2301.03829
|
Kaiping Zheng
|
Kaiping Zheng, Thao Nguyen, Jesslyn Hwei Sing Chong, Charlene Enhui
Goh, Melanie Herschel, Hee Hoon Lee, Changshuo Liu, Beng Chin Ooi, Wei Wang
and James Yip
|
From Plate to Prevention: A Dietary Nutrient-aided Platform for Health
Promotion in Singapore
| null | null | null | null |
cs.LG cs.AI cs.CV cs.DB cs.MM
|
http://creativecommons.org/licenses/by/4.0/
|
Singapore has been striving to improve the provision of healthcare services
to her people. In this course, the government has taken note of the deficiency
in regulating and supervising people's nutrient intake, which is identified as
a contributing factor to the development of chronic diseases. Consequently,
this issue has garnered significant attention. In this paper, we share our
experience in addressing this issue and attaining medical-grade nutrient intake
information to benefit Singaporeans in different aspects. To this end, we
develop the FoodSG platform to incubate diverse healthcare-oriented
applications as a service in Singapore, taking into account their shared
requirements. We further identify the profound meaning of localized food
datasets and systematically clean and curate a localized Singaporean food
dataset FoodSG-233. To overcome the hurdle in recognition performance brought
by Singaporean multifarious food dishes, we propose to integrate supervised
contrastive learning into our food recognition model FoodSG-SCL for the
intrinsic capability to mine hard positive/negative samples and therefore boost
the accuracy. Through a comprehensive evaluation, we present performance
results of the proposed model and insights on food-related healthcare
applications. The FoodSG-233 dataset has been released in
https://foodlg.comp.nus.edu.sg/.
|
[
{
"version": "v1",
"created": "Tue, 10 Jan 2023 07:51:36 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 15:54:39 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zheng",
"Kaiping",
""
],
[
"Nguyen",
"Thao",
""
],
[
"Chong",
"Jesslyn Hwei Sing",
""
],
[
"Goh",
"Charlene Enhui",
""
],
[
"Herschel",
"Melanie",
""
],
[
"Lee",
"Hee Hoon",
""
],
[
"Liu",
"Changshuo",
""
],
[
"Ooi",
"Beng Chin",
""
],
[
"Wang",
"Wei",
""
],
[
"Yip",
"James",
""
]
] |
TITLE: From Plate to Prevention: A Dietary Nutrient-aided Platform for Health
Promotion in Singapore
ABSTRACT: Singapore has been striving to improve the provision of healthcare services
to her people. In this course, the government has taken note of the deficiency
in regulating and supervising people's nutrient intake, which is identified as
a contributing factor to the development of chronic diseases. Consequently,
this issue has garnered significant attention. In this paper, we share our
experience in addressing this issue and attaining medical-grade nutrient intake
information to benefit Singaporeans in different aspects. To this end, we
develop the FoodSG platform to incubate diverse healthcare-oriented
applications as a service in Singapore, taking into account their shared
requirements. We further identify the profound meaning of localized food
datasets and systematically clean and curate a localized Singaporean food
dataset FoodSG-233. To overcome the hurdle in recognition performance brought
by Singaporean multifarious food dishes, we propose to integrate supervised
contrastive learning into our food recognition model FoodSG-SCL for the
intrinsic capability to mine hard positive/negative samples and therefore boost
the accuracy. Through a comprehensive evaluation, we present performance
results of the proposed model and insights on food-related healthcare
applications. The FoodSG-233 dataset has been released in
https://foodlg.comp.nus.edu.sg/.
|
new_dataset
| 0.716262
|
2301.03949
|
Mengyi Zhao
|
Mengyi Zhao, Mengyuan Liu, Bin Ren, Shuling Dai, and Nicu Sebe
|
Modiff: Action-Conditioned 3D Motion Generation with Denoising Diffusion
Probabilistic Models
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Diffusion-based generative models have recently emerged as powerful solutions
for high-quality synthesis in multiple domains. Leveraging the bidirectional
Markov chains, diffusion probabilistic models generate samples by inferring the
reversed Markov chain based on the learned distribution mapping at the forward
diffusion process. In this work, we propose Modiff, a conditional paradigm that
benefits from the denoising diffusion probabilistic model (DDPM) to tackle the
problem of realistic and diverse action-conditioned 3D skeleton-based motion
generation. We are a pioneering attempt that uses DDPM to synthesize a variable
number of motion sequences conditioned on a categorical action. We evaluate our
approach on the large-scale NTU RGB+D dataset and show improvements over
state-of-the-art motion generation methods.
|
[
{
"version": "v1",
"created": "Tue, 10 Jan 2023 13:15:42 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 08:26:30 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhao",
"Mengyi",
""
],
[
"Liu",
"Mengyuan",
""
],
[
"Ren",
"Bin",
""
],
[
"Dai",
"Shuling",
""
],
[
"Sebe",
"Nicu",
""
]
] |
TITLE: Modiff: Action-Conditioned 3D Motion Generation with Denoising Diffusion
Probabilistic Models
ABSTRACT: Diffusion-based generative models have recently emerged as powerful solutions
for high-quality synthesis in multiple domains. Leveraging the bidirectional
Markov chains, diffusion probabilistic models generate samples by inferring the
reversed Markov chain based on the learned distribution mapping at the forward
diffusion process. In this work, we propose Modiff, a conditional paradigm that
benefits from the denoising diffusion probabilistic model (DDPM) to tackle the
problem of realistic and diverse action-conditioned 3D skeleton-based motion
generation. We are a pioneering attempt that uses DDPM to synthesize a variable
number of motion sequences conditioned on a categorical action. We evaluate our
approach on the large-scale NTU RGB+D dataset and show improvements over
state-of-the-art motion generation methods.
|
no_new_dataset
| 0.711437
|
2301.09919
|
Nico Daheim
|
Jakub Macina, Nico Daheim, Lingzhi Wang, Tanmay Sinha, Manu Kapur,
Iryna Gurevych, Mrinmaya Sachan
|
Opportunities and Challenges in Neural Dialog Tutoring
|
EACL 2023 (main conference, camera-ready)
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Designing dialog tutors has been challenging as it involves modeling the
diverse and complex pedagogical strategies employed by human tutors. Although
there have been significant recent advances in neural conversational systems
using large language models (LLMs) and growth in available dialog corpora,
dialog tutoring has largely remained unaffected by these advances. In this
paper, we rigorously analyze various generative language models on two dialog
tutoring datasets for language learning using automatic and human evaluations
to understand the new opportunities brought by these advances as well as the
challenges we must overcome to build models that would be usable in real
educational settings. We find that although current approaches can model
tutoring in constrained learning scenarios when the number of concepts to be
taught and possible teacher strategies are small, they perform poorly in less
constrained scenarios. Our human quality evaluation shows that both models and
ground-truth annotations exhibit low performance in terms of equitable
tutoring, which measures learning opportunities for students and how engaging
the dialog is. To understand the behavior of our models in a real tutoring
setting, we conduct a user study using expert annotators and find a
significantly large number of model reasoning errors in 45% of conversations.
Finally, we connect our findings to outline future work.
|
[
{
"version": "v1",
"created": "Tue, 24 Jan 2023 11:00:17 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 19:13:35 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Macina",
"Jakub",
""
],
[
"Daheim",
"Nico",
""
],
[
"Wang",
"Lingzhi",
""
],
[
"Sinha",
"Tanmay",
""
],
[
"Kapur",
"Manu",
""
],
[
"Gurevych",
"Iryna",
""
],
[
"Sachan",
"Mrinmaya",
""
]
] |
TITLE: Opportunities and Challenges in Neural Dialog Tutoring
ABSTRACT: Designing dialog tutors has been challenging as it involves modeling the
diverse and complex pedagogical strategies employed by human tutors. Although
there have been significant recent advances in neural conversational systems
using large language models (LLMs) and growth in available dialog corpora,
dialog tutoring has largely remained unaffected by these advances. In this
paper, we rigorously analyze various generative language models on two dialog
tutoring datasets for language learning using automatic and human evaluations
to understand the new opportunities brought by these advances as well as the
challenges we must overcome to build models that would be usable in real
educational settings. We find that although current approaches can model
tutoring in constrained learning scenarios when the number of concepts to be
taught and possible teacher strategies are small, they perform poorly in less
constrained scenarios. Our human quality evaluation shows that both models and
ground-truth annotations exhibit low performance in terms of equitable
tutoring, which measures learning opportunities for students and how engaging
the dialog is. To understand the behavior of our models in a real tutoring
setting, we conduct a user study using expert annotators and find a
significantly large number of model reasoning errors in 45% of conversations.
Finally, we connect our findings to outline future work.
|
no_new_dataset
| 0.708036
|
2302.05075
|
Weichao Zhao
|
Weichao Zhao, Hezhen Hu, Wengang Zhou, Jiaxin Shi, Houqiang Li
|
BEST: BERT Pre-Training for Sign Language Recognition with Coupling
Tokenization
|
Accepted by AAAI 2023 (Oral)
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
In this work, we are dedicated to leveraging the BERT pre-training success
and modeling the domain-specific statistics to fertilize the sign language
recognition~(SLR) model. Considering the dominance of hand and body in sign
language expression, we organize them as pose triplet units and feed them into
the Transformer backbone in a frame-wise manner. Pre-training is performed via
reconstructing the masked triplet unit from the corrupted input sequence, which
learns the hierarchical correlation context cues among internal and external
triplet units. Notably, different from the highly semantic word token in BERT,
the pose unit is a low-level signal originally located in continuous space,
which prevents the direct adoption of the BERT cross-entropy objective. To this
end, we bridge this semantic gap via coupling tokenization of the triplet unit.
It adaptively extracts the discrete pseudo label from the pose triplet unit,
which represents the semantic gesture/body state. After pre-training, we
fine-tune the pre-trained encoder on the downstream SLR task, jointly with the
newly added task-specific layer. Extensive experiments are conducted to
validate the effectiveness of our proposed method, achieving new
state-of-the-art performance on all four benchmarks with a notable gain.
|
[
{
"version": "v1",
"created": "Fri, 10 Feb 2023 06:23:44 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Feb 2023 13:49:45 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 02:02:04 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhao",
"Weichao",
""
],
[
"Hu",
"Hezhen",
""
],
[
"Zhou",
"Wengang",
""
],
[
"Shi",
"Jiaxin",
""
],
[
"Li",
"Houqiang",
""
]
] |
TITLE: BEST: BERT Pre-Training for Sign Language Recognition with Coupling
Tokenization
ABSTRACT: In this work, we are dedicated to leveraging the BERT pre-training success
and modeling the domain-specific statistics to fertilize the sign language
recognition~(SLR) model. Considering the dominance of hand and body in sign
language expression, we organize them as pose triplet units and feed them into
the Transformer backbone in a frame-wise manner. Pre-training is performed via
reconstructing the masked triplet unit from the corrupted input sequence, which
learns the hierarchical correlation context cues among internal and external
triplet units. Notably, different from the highly semantic word token in BERT,
the pose unit is a low-level signal originally located in continuous space,
which prevents the direct adoption of the BERT cross-entropy objective. To this
end, we bridge this semantic gap via coupling tokenization of the triplet unit.
It adaptively extracts the discrete pseudo label from the pose triplet unit,
which represents the semantic gesture/body state. After pre-training, we
fine-tune the pre-trained encoder on the downstream SLR task, jointly with the
newly added task-specific layer. Extensive experiments are conducted to
validate the effectiveness of our proposed method, achieving new
state-of-the-art performance on all four benchmarks with a notable gain.
|
no_new_dataset
| 0.709265
|
2302.07387
|
Hui Ding
|
Jiang Liu, Hui Ding, Zhaowei Cai, Yuting Zhang, Ravi Kumar Satzoda,
Vijay Mahadevan, R. Manmatha
|
PolyFormer: Referring Image Segmentation as Sequential Polygon
Generation
|
CVPR 2023. Project Page: https://polyformer.github.io/
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
In this work, instead of directly predicting the pixel-level segmentation
masks, the problem of referring image segmentation is formulated as sequential
polygon generation, and the predicted polygons can be later converted into
segmentation masks. This is enabled by a new sequence-to-sequence framework,
Polygon Transformer (PolyFormer), which takes a sequence of image patches and
text query tokens as input, and outputs a sequence of polygon vertices
autoregressively. For more accurate geometric localization, we propose a
regression-based decoder, which predicts the precise floating-point coordinates
directly, without any coordinate quantization error. In the experiments,
PolyFormer outperforms the prior art by a clear margin, e.g., 5.40% and 4.52%
absolute improvements on the challenging RefCOCO+ and RefCOCOg datasets. It
also shows strong generalization ability when evaluated on the referring video
segmentation task without fine-tuning, e.g., achieving competitive 61.5% J&F on
the Ref-DAVIS17 dataset.
|
[
{
"version": "v1",
"created": "Tue, 14 Feb 2023 23:00:25 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 23:22:31 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Liu",
"Jiang",
""
],
[
"Ding",
"Hui",
""
],
[
"Cai",
"Zhaowei",
""
],
[
"Zhang",
"Yuting",
""
],
[
"Satzoda",
"Ravi Kumar",
""
],
[
"Mahadevan",
"Vijay",
""
],
[
"Manmatha",
"R.",
""
]
] |
TITLE: PolyFormer: Referring Image Segmentation as Sequential Polygon
Generation
ABSTRACT: In this work, instead of directly predicting the pixel-level segmentation
masks, the problem of referring image segmentation is formulated as sequential
polygon generation, and the predicted polygons can be later converted into
segmentation masks. This is enabled by a new sequence-to-sequence framework,
Polygon Transformer (PolyFormer), which takes a sequence of image patches and
text query tokens as input, and outputs a sequence of polygon vertices
autoregressively. For more accurate geometric localization, we propose a
regression-based decoder, which predicts the precise floating-point coordinates
directly, without any coordinate quantization error. In the experiments,
PolyFormer outperforms the prior art by a clear margin, e.g., 5.40% and 4.52%
absolute improvements on the challenging RefCOCO+ and RefCOCOg datasets. It
also shows strong generalization ability when evaluated on the referring video
segmentation task without fine-tuning, e.g., achieving competitive 61.5% J&F on
the Ref-DAVIS17 dataset.
|
no_new_dataset
| 0.712595
|
2302.11014
|
Sayak Kundu
|
Chung-Kuan Cheng, Andrew B. Kahng, Sayak Kundu, Yucheng Wang, Zhiang
Wang
|
Assessment of Reinforcement Learning for Macro Placement
|
There are eight pages and one page for reference. It includes five
figures and seven tables. This paper has been invited to ISPD 2023
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
We provide open, transparent implementation and assessment of Google Brain's
deep reinforcement learning approach to macro placement and its Circuit
Training (CT) implementation in GitHub. We implement in open source key
"blackbox" elements of CT, and clarify discrepancies between CT and Nature
paper. New testcases on open enablements are developed and released. We assess
CT alongside multiple alternative macro placers, with all evaluation flows and
related scripts public in GitHub. Our experiments also encompass academic
mixed-size placement benchmarks, as well as ablation and stability studies. We
comment on the impact of Nature and CT, as well as directions for future
research.
|
[
{
"version": "v1",
"created": "Tue, 21 Feb 2023 21:26:28 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 18:01:40 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Cheng",
"Chung-Kuan",
""
],
[
"Kahng",
"Andrew B.",
""
],
[
"Kundu",
"Sayak",
""
],
[
"Wang",
"Yucheng",
""
],
[
"Wang",
"Zhiang",
""
]
] |
TITLE: Assessment of Reinforcement Learning for Macro Placement
ABSTRACT: We provide open, transparent implementation and assessment of Google Brain's
deep reinforcement learning approach to macro placement and its Circuit
Training (CT) implementation in GitHub. We implement in open source key
"blackbox" elements of CT, and clarify discrepancies between CT and Nature
paper. New testcases on open enablements are developed and released. We assess
CT alongside multiple alternative macro placers, with all evaluation flows and
related scripts public in GitHub. Our experiments also encompass academic
mixed-size placement benchmarks, as well as ablation and stability studies. We
comment on the impact of Nature and CT, as well as directions for future
research.
|
no_new_dataset
| 0.710377
|
2302.14434
|
Biwen Lei
|
Biwen Lei, Jianqiang Ren, Mengyang Feng, Miaomiao Cui, Xuansong Xie
|
A Hierarchical Representation Network for Accurate and Detailed Face
Reconstruction from In-The-Wild Images
|
Accepted by CVPR2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Limited by the nature of the low-dimensional representational capacity of
3DMM, most of the 3DMM-based face reconstruction (FR) methods fail to recover
high-frequency facial details, such as wrinkles, dimples, etc. Some attempt to
solve the problem by introducing detail maps or non-linear operations, however,
the results are still not vivid. To this end, we in this paper present a novel
hierarchical representation network (HRN) to achieve accurate and detailed face
reconstruction from a single image. Specifically, we implement the geometry
disentanglement and introduce the hierarchical representation to fulfill
detailed face modeling. Meanwhile, 3D priors of facial details are incorporated
to enhance the accuracy and authenticity of the reconstruction results. We also
propose a de-retouching module to achieve better decoupling of the geometry and
appearance. It is noteworthy that our framework can be extended to a multi-view
fashion by considering detail consistency of different views. Extensive
experiments on two single-view and two multi-view FR benchmarks demonstrate
that our method outperforms the existing methods in both reconstruction
accuracy and visual effects. Finally, we introduce a high-quality 3D face
dataset FaceHD-100 to boost the research of high-fidelity face reconstruction.
The project homepage is at https://younglbw.github.io/HRN-homepage/.
|
[
{
"version": "v1",
"created": "Tue, 28 Feb 2023 09:24:36 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 06:27:14 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Lei",
"Biwen",
""
],
[
"Ren",
"Jianqiang",
""
],
[
"Feng",
"Mengyang",
""
],
[
"Cui",
"Miaomiao",
""
],
[
"Xie",
"Xuansong",
""
]
] |
TITLE: A Hierarchical Representation Network for Accurate and Detailed Face
Reconstruction from In-The-Wild Images
ABSTRACT: Limited by the nature of the low-dimensional representational capacity of
3DMM, most of the 3DMM-based face reconstruction (FR) methods fail to recover
high-frequency facial details, such as wrinkles, dimples, etc. Some attempt to
solve the problem by introducing detail maps or non-linear operations, however,
the results are still not vivid. To this end, we in this paper present a novel
hierarchical representation network (HRN) to achieve accurate and detailed face
reconstruction from a single image. Specifically, we implement the geometry
disentanglement and introduce the hierarchical representation to fulfill
detailed face modeling. Meanwhile, 3D priors of facial details are incorporated
to enhance the accuracy and authenticity of the reconstruction results. We also
propose a de-retouching module to achieve better decoupling of the geometry and
appearance. It is noteworthy that our framework can be extended to a multi-view
fashion by considering detail consistency of different views. Extensive
experiments on two single-view and two multi-view FR benchmarks demonstrate
that our method outperforms the existing methods in both reconstruction
accuracy and visual effects. Finally, we introduce a high-quality 3D face
dataset FaceHD-100 to boost the research of high-fidelity face reconstruction.
The project homepage is at https://younglbw.github.io/HRN-homepage/.
|
no_new_dataset
| 0.710057
|
2303.01979
|
Sangmin Hong
|
Sangmin Hong, Mohsen Yavartanoo, Reyhaneh Neshatavar, Kyoung Mu Lee
|
ACL-SPC: Adaptive Closed-Loop system for Self-Supervised Point Cloud
Completion
|
Published at CVPR 2023
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Point cloud completion addresses filling in the missing parts of a partial
point cloud obtained from depth sensors and generating a complete point cloud.
Although there has been steep progress in the supervised methods on the
synthetic point cloud completion task, it is hardly applicable in real-world
scenarios due to the domain gap between the synthetic and real-world datasets
or the requirement of prior information. To overcome these limitations, we
propose a novel self-supervised framework ACL-SPC for point cloud completion to
train and test on the same data. ACL-SPC takes a single partial input and
attempts to output the complete point cloud using an adaptive closed-loop (ACL)
system that enforces the output same for the variation of an input. We evaluate
our proposed ACL-SPC on various datasets to prove that it can successfully
learn to complete a partial point cloud as the first self-supervised scheme.
Results show that our method is comparable with unsupervised methods and
achieves superior performance on the real-world dataset compared to the
supervised methods trained on the synthetic dataset. Extensive experiments
justify the necessity of self-supervised learning and the effectiveness of our
proposed method for the real-world point cloud completion task. The code is
publicly available from https://github.com/Sangminhong/ACL-SPC_PyTorch
|
[
{
"version": "v1",
"created": "Thu, 2 Mar 2023 08:02:45 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Mar 2023 08:26:32 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 12:44:30 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Hong",
"Sangmin",
""
],
[
"Yavartanoo",
"Mohsen",
""
],
[
"Neshatavar",
"Reyhaneh",
""
],
[
"Lee",
"Kyoung Mu",
""
]
] |
TITLE: ACL-SPC: Adaptive Closed-Loop system for Self-Supervised Point Cloud
Completion
ABSTRACT: Point cloud completion addresses filling in the missing parts of a partial
point cloud obtained from depth sensors and generating a complete point cloud.
Although there has been steep progress in the supervised methods on the
synthetic point cloud completion task, it is hardly applicable in real-world
scenarios due to the domain gap between the synthetic and real-world datasets
or the requirement of prior information. To overcome these limitations, we
propose a novel self-supervised framework ACL-SPC for point cloud completion to
train and test on the same data. ACL-SPC takes a single partial input and
attempts to output the complete point cloud using an adaptive closed-loop (ACL)
system that enforces the output same for the variation of an input. We evaluate
our proposed ACL-SPC on various datasets to prove that it can successfully
learn to complete a partial point cloud as the first self-supervised scheme.
Results show that our method is comparable with unsupervised methods and
achieves superior performance on the real-world dataset compared to the
supervised methods trained on the synthetic dataset. Extensive experiments
justify the necessity of self-supervised learning and the effectiveness of our
proposed method for the real-world point cloud completion task. The code is
publicly available from https://github.com/Sangminhong/ACL-SPC_PyTorch
|
no_new_dataset
| 0.711249
|
2303.03729
|
Huanyu Zhou
|
Huanyu Zhou, Qingjie Liu, Yunhong Wang
|
Learning Discriminative Representations for Skeleton Based Action
Recognition
|
Accepted by CVPR2023. 10 pages, 5 figures, 5 tables
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Human action recognition aims at classifying the category of human action
from a segment of a video. Recently, people have dived into designing GCN-based
models to extract features from skeletons for performing this task, because
skeleton representations are much more efficient and robust than other
modalities such as RGB frames. However, when employing the skeleton data, some
important clues like related items are also discarded. It results in some
ambiguous actions that are hard to be distinguished and tend to be
misclassified. To alleviate this problem, we propose an auxiliary feature
refinement head (FR Head), which consists of spatial-temporal decoupling and
contrastive feature refinement, to obtain discriminative representations of
skeletons. Ambiguous samples are dynamically discovered and calibrated in the
feature space. Furthermore, FR Head could be imposed on different stages of
GCNs to build a multi-level refinement for stronger supervision. Extensive
experiments are conducted on NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.
Our proposed models obtain competitive results from state-of-the-art methods
and can help to discriminate those ambiguous samples. Codes are available at
https://github.com/zhysora/FR-Head.
|
[
{
"version": "v1",
"created": "Tue, 7 Mar 2023 08:37:48 GMT"
},
{
"version": "v2",
"created": "Tue, 14 Mar 2023 13:07:09 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 03:17:05 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhou",
"Huanyu",
""
],
[
"Liu",
"Qingjie",
""
],
[
"Wang",
"Yunhong",
""
]
] |
TITLE: Learning Discriminative Representations for Skeleton Based Action
Recognition
ABSTRACT: Human action recognition aims at classifying the category of human action
from a segment of a video. Recently, people have dived into designing GCN-based
models to extract features from skeletons for performing this task, because
skeleton representations are much more efficient and robust than other
modalities such as RGB frames. However, when employing the skeleton data, some
important clues like related items are also discarded. It results in some
ambiguous actions that are hard to be distinguished and tend to be
misclassified. To alleviate this problem, we propose an auxiliary feature
refinement head (FR Head), which consists of spatial-temporal decoupling and
contrastive feature refinement, to obtain discriminative representations of
skeletons. Ambiguous samples are dynamically discovered and calibrated in the
feature space. Furthermore, FR Head could be imposed on different stages of
GCNs to build a multi-level refinement for stronger supervision. Extensive
experiments are conducted on NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.
Our proposed models obtain competitive results from state-of-the-art methods
and can help to discriminate those ambiguous samples. Codes are available at
https://github.com/zhysora/FR-Head.
|
no_new_dataset
| 0.709667
|
2303.05371
|
Anchit Gupta
|
Anchit Gupta, Wenhan Xiong, Yixin Nie, Ian Jones, Barlas O\u{g}uz
|
3DGen: Triplane Latent Diffusion for Textured Mesh Generation
| null | null | null | null |
cs.CV cs.GR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Latent diffusion models for image generation have crossed a quality threshold
which enabled them to achieve mass adoption. Recently, a series of works have
made advancements towards replicating this success in the 3D domain,
introducing techniques such as point cloud VAE, triplane representation, neural
implicit surfaces and differentiable rendering based training. We take another
step along this direction, combining these developments in a two-step pipeline
consisting of 1) a triplane VAE which can learn latent representations of
textured meshes and 2) a conditional diffusion model which generates the
triplane features. For the first time this architecture allows conditional and
unconditional generation of high quality textured or untextured 3D meshes
across multiple diverse categories in a few seconds on a single GPU. It
outperforms previous work substantially on image-conditioned and unconditional
generation on mesh quality as well as texture generation. Furthermore, we
demonstrate the scalability of our model to large datasets for increased
quality and diversity. We will release our code and trained models.
|
[
{
"version": "v1",
"created": "Thu, 9 Mar 2023 16:18:14 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 18:04:20 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Gupta",
"Anchit",
""
],
[
"Xiong",
"Wenhan",
""
],
[
"Nie",
"Yixin",
""
],
[
"Jones",
"Ian",
""
],
[
"Oğuz",
"Barlas",
""
]
] |
TITLE: 3DGen: Triplane Latent Diffusion for Textured Mesh Generation
ABSTRACT: Latent diffusion models for image generation have crossed a quality threshold
which enabled them to achieve mass adoption. Recently, a series of works have
made advancements towards replicating this success in the 3D domain,
introducing techniques such as point cloud VAE, triplane representation, neural
implicit surfaces and differentiable rendering based training. We take another
step along this direction, combining these developments in a two-step pipeline
consisting of 1) a triplane VAE which can learn latent representations of
textured meshes and 2) a conditional diffusion model which generates the
triplane features. For the first time this architecture allows conditional and
unconditional generation of high quality textured or untextured 3D meshes
across multiple diverse categories in a few seconds on a single GPU. It
outperforms previous work substantially on image-conditioned and unconditional
generation on mesh quality as well as texture generation. Furthermore, we
demonstrate the scalability of our model to large datasets for increased
quality and diversity. We will release our code and trained models.
|
no_new_dataset
| 0.709416
|
2303.06552
|
Michael Rotman
|
Michael Rotman, Lior Wolf
|
Energy Regularized RNNs for Solving Non-Stationary Bandit Problems
| null | null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We consider a Multi-Armed Bandit problem in which the rewards are
non-stationary and are dependent on past actions and potentially on past
contexts. At the heart of our method, we employ a recurrent neural network,
which models these sequences. In order to balance between exploration and
exploitation, we present an energy minimization term that prevents the neural
network from becoming too confident in support of a certain action. This term
provably limits the gap between the maximal and minimal probabilities assigned
by the network. In a diverse set of experiments, we demonstrate that our method
is at least as effective as methods suggested to solve the sub-problem of
Rotting Bandits, and can solve intuitive extensions of various benchmark
problems. We share our implementation at
https://github.com/rotmanmi/Energy-Regularized-RNN.
|
[
{
"version": "v1",
"created": "Sun, 12 Mar 2023 03:32:43 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 15:20:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Rotman",
"Michael",
""
],
[
"Wolf",
"Lior",
""
]
] |
TITLE: Energy Regularized RNNs for Solving Non-Stationary Bandit Problems
ABSTRACT: We consider a Multi-Armed Bandit problem in which the rewards are
non-stationary and are dependent on past actions and potentially on past
contexts. At the heart of our method, we employ a recurrent neural network,
which models these sequences. In order to balance between exploration and
exploitation, we present an energy minimization term that prevents the neural
network from becoming too confident in support of a certain action. This term
provably limits the gap between the maximal and minimal probabilities assigned
by the network. In a diverse set of experiments, we demonstrate that our method
is at least as effective as methods suggested to solve the sub-problem of
Rotting Bandits, and can solve intuitive extensions of various benchmark
problems. We share our implementation at
https://github.com/rotmanmi/Energy-Regularized-RNN.
|
no_new_dataset
| 0.709233
|
2303.10720
|
Junjiao Tian
|
Junjiao Tian, Xiaoliang Dai, Chih-Yao Ma, Zecheng He, Yen-Cheng Liu,
Zsolt Kira
|
Trainable Projected Gradient Method for Robust Fine-tuning
|
Accepted to CVPR2023
|
Conference on Computer Vision and Pattern Recognition 2023
| null | null |
cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Recent studies on transfer learning have shown that selectively fine-tuning a
subset of layers or customizing different learning rates for each layer can
greatly improve robustness to out-of-distribution (OOD) data and retain
generalization capability in the pre-trained models. However, most of these
methods employ manually crafted heuristics or expensive hyper-parameter
searches, which prevent them from scaling up to large datasets and neural
networks. To solve this problem, we propose Trainable Projected Gradient Method
(TPGM) to automatically learn the constraint imposed for each layer for a
fine-grained fine-tuning regularization. This is motivated by formulating
fine-tuning as a bi-level constrained optimization problem. Specifically, TPGM
maintains a set of projection radii, i.e., distance constraints between the
fine-tuned model and the pre-trained model, for each layer, and enforces them
through weight projections. To learn the constraints, we propose a bi-level
optimization to automatically learn the best set of projection radii in an
end-to-end manner. Theoretically, we show that the bi-level optimization
formulation could explain the regularization capability of TPGM. Empirically,
with little hyper-parameter search cost, TPGM outperforms existing fine-tuning
methods in OOD performance while matching the best in-distribution (ID)
performance. For example, when fine-tuned on DomainNet-Real and ImageNet,
compared to vanilla fine-tuning, TPGM shows $22\%$ and $10\%$ relative OOD
improvement respectively on their sketch counterparts. Code is available at
\url{https://github.com/PotatoTian/TPGM}.
|
[
{
"version": "v1",
"created": "Sun, 19 Mar 2023 17:30:44 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 15:04:36 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Tian",
"Junjiao",
""
],
[
"Dai",
"Xiaoliang",
""
],
[
"Ma",
"Chih-Yao",
""
],
[
"He",
"Zecheng",
""
],
[
"Liu",
"Yen-Cheng",
""
],
[
"Kira",
"Zsolt",
""
]
] |
TITLE: Trainable Projected Gradient Method for Robust Fine-tuning
ABSTRACT: Recent studies on transfer learning have shown that selectively fine-tuning a
subset of layers or customizing different learning rates for each layer can
greatly improve robustness to out-of-distribution (OOD) data and retain
generalization capability in the pre-trained models. However, most of these
methods employ manually crafted heuristics or expensive hyper-parameter
searches, which prevent them from scaling up to large datasets and neural
networks. To solve this problem, we propose Trainable Projected Gradient Method
(TPGM) to automatically learn the constraint imposed for each layer for a
fine-grained fine-tuning regularization. This is motivated by formulating
fine-tuning as a bi-level constrained optimization problem. Specifically, TPGM
maintains a set of projection radii, i.e., distance constraints between the
fine-tuned model and the pre-trained model, for each layer, and enforces them
through weight projections. To learn the constraints, we propose a bi-level
optimization to automatically learn the best set of projection radii in an
end-to-end manner. Theoretically, we show that the bi-level optimization
formulation could explain the regularization capability of TPGM. Empirically,
with little hyper-parameter search cost, TPGM outperforms existing fine-tuning
methods in OOD performance while matching the best in-distribution (ID)
performance. For example, when fine-tuned on DomainNet-Real and ImageNet,
compared to vanilla fine-tuning, TPGM shows $22\%$ and $10\%$ relative OOD
improvement respectively on their sketch counterparts. Code is available at
\url{https://github.com/PotatoTian/TPGM}.
|
no_new_dataset
| 0.70724
|
2303.10820
|
Shogo Sato
|
Shogo Sato, Yasuhiro Yao, Taiga Yoshida, Takuhiro Kaneko, Shingo Ando,
Jun Shimamura
|
Unsupervised Intrinsic Image Decomposition with LiDAR Intensity
|
Accepted by CVPR2023, Dataset link :
(https://github.com/ntthilab-cv/NTT-intrinsic-dataset)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Intrinsic image decomposition (IID) is the task that decomposes a natural
image into albedo and shade. While IID is typically solved through supervised
learning methods, it is not ideal due to the difficulty in observing ground
truth albedo and shade in general scenes. Conversely, unsupervised learning
methods are currently underperforming supervised learning methods since there
are no criteria for solving the ill-posed problems. Recently, light detection
and ranging (LiDAR) is widely used due to its ability to make highly precise
distance measurements. Thus, we have focused on the utilization of LiDAR,
especially LiDAR intensity, to address this issue. In this paper, we propose
unsupervised intrinsic image decomposition with LiDAR intensity (IID-LI). Since
the conventional unsupervised learning methods consist of image-to-image
transformations, simply inputting LiDAR intensity is not an effective approach.
Therefore, we design an intensity consistency loss that computes the error
between LiDAR intensity and gray-scaled albedo to provide a criterion for the
ill-posed problem. In addition, LiDAR intensity is difficult to handle due to
its sparsity and occlusion, hence, a LiDAR intensity densification module is
proposed. We verified the estimating quality using our own dataset, which
include RGB images, LiDAR intensity and human judged annotations. As a result,
we achieved an estimation accuracy that outperforms conventional unsupervised
learning methods. Dataset link :
(https://github.com/ntthilab-cv/NTT-intrinsic-dataset).
|
[
{
"version": "v1",
"created": "Mon, 20 Mar 2023 01:26:16 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 07:38:31 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Sato",
"Shogo",
""
],
[
"Yao",
"Yasuhiro",
""
],
[
"Yoshida",
"Taiga",
""
],
[
"Kaneko",
"Takuhiro",
""
],
[
"Ando",
"Shingo",
""
],
[
"Shimamura",
"Jun",
""
]
] |
TITLE: Unsupervised Intrinsic Image Decomposition with LiDAR Intensity
ABSTRACT: Intrinsic image decomposition (IID) is the task that decomposes a natural
image into albedo and shade. While IID is typically solved through supervised
learning methods, it is not ideal due to the difficulty in observing ground
truth albedo and shade in general scenes. Conversely, unsupervised learning
methods are currently underperforming supervised learning methods since there
are no criteria for solving the ill-posed problems. Recently, light detection
and ranging (LiDAR) is widely used due to its ability to make highly precise
distance measurements. Thus, we have focused on the utilization of LiDAR,
especially LiDAR intensity, to address this issue. In this paper, we propose
unsupervised intrinsic image decomposition with LiDAR intensity (IID-LI). Since
the conventional unsupervised learning methods consist of image-to-image
transformations, simply inputting LiDAR intensity is not an effective approach.
Therefore, we design an intensity consistency loss that computes the error
between LiDAR intensity and gray-scaled albedo to provide a criterion for the
ill-posed problem. In addition, LiDAR intensity is difficult to handle due to
its sparsity and occlusion, hence, a LiDAR intensity densification module is
proposed. We verified the estimating quality using our own dataset, which
include RGB images, LiDAR intensity and human judged annotations. As a result,
we achieved an estimation accuracy that outperforms conventional unsupervised
learning methods. Dataset link :
(https://github.com/ntthilab-cv/NTT-intrinsic-dataset).
|
no_new_dataset
| 0.706634
|
2303.11203
|
Li Li
|
Li Li, Hubert P. H. Shum, Toby P. Breckon
|
Less is More: Reducing Task and Model Complexity for 3D Point Cloud
Semantic Segmentation
|
Accepted by CVPR 2023; 11 pages, 8 figures; Code at
https://github.com/l1997i/lim3d
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Whilst the availability of 3D LiDAR point cloud data has significantly grown
in recent years, annotation remains expensive and time-consuming, leading to a
demand for semi-supervised semantic segmentation methods with application
domains such as autonomous driving. Existing work very often employs relatively
large segmentation backbone networks to improve segmentation accuracy, at the
expense of computational costs. In addition, many use uniform sampling to
reduce ground truth data requirements for learning needed, often resulting in
sub-optimal performance. To address these issues, we propose a new pipeline
that employs a smaller architecture, requiring fewer ground-truth annotations
to achieve superior segmentation accuracy compared to contemporary approaches.
This is facilitated via a novel Sparse Depthwise Separable Convolution module
that significantly reduces the network parameter count while retaining overall
task performance. To effectively sub-sample our training data, we propose a new
Spatio-Temporal Redundant Frame Downsampling (ST-RFD) method that leverages
knowledge of sensor motion within the environment to extract a more diverse
subset of training data frame samples. To leverage the use of limited annotated
data samples, we further propose a soft pseudo-label method informed by LiDAR
reflectivity. Our method outperforms contemporary semi-supervised work in terms
of mIoU, using less labeled data, on the SemanticKITTI (59.5@5%) and
ScribbleKITTI (58.1@5%) benchmark datasets, based on a 2.3x reduction in model
parameters and 641x fewer multiply-add operations whilst also demonstrating
significant performance improvement on limited training data (i.e., Less is
More).
|
[
{
"version": "v1",
"created": "Mon, 20 Mar 2023 15:36:10 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 14:56:30 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Li",
"Li",
""
],
[
"Shum",
"Hubert P. H.",
""
],
[
"Breckon",
"Toby P.",
""
]
] |
TITLE: Less is More: Reducing Task and Model Complexity for 3D Point Cloud
Semantic Segmentation
ABSTRACT: Whilst the availability of 3D LiDAR point cloud data has significantly grown
in recent years, annotation remains expensive and time-consuming, leading to a
demand for semi-supervised semantic segmentation methods with application
domains such as autonomous driving. Existing work very often employs relatively
large segmentation backbone networks to improve segmentation accuracy, at the
expense of computational costs. In addition, many use uniform sampling to
reduce ground truth data requirements for learning needed, often resulting in
sub-optimal performance. To address these issues, we propose a new pipeline
that employs a smaller architecture, requiring fewer ground-truth annotations
to achieve superior segmentation accuracy compared to contemporary approaches.
This is facilitated via a novel Sparse Depthwise Separable Convolution module
that significantly reduces the network parameter count while retaining overall
task performance. To effectively sub-sample our training data, we propose a new
Spatio-Temporal Redundant Frame Downsampling (ST-RFD) method that leverages
knowledge of sensor motion within the environment to extract a more diverse
subset of training data frame samples. To leverage the use of limited annotated
data samples, we further propose a soft pseudo-label method informed by LiDAR
reflectivity. Our method outperforms contemporary semi-supervised work in terms
of mIoU, using less labeled data, on the SemanticKITTI (59.5@5%) and
ScribbleKITTI (58.1@5%) benchmark datasets, based on a 2.3x reduction in model
parameters and 641x fewer multiply-add operations whilst also demonstrating
significant performance improvement on limited training data (i.e., Less is
More).
|
no_new_dataset
| 0.710641
|
2303.12313
|
Haipeng Zhou
|
Haipeng Zhou, Lei Zhu, Yuyin Zhou
|
Distribution Aligned Diffusion and Prototype-guided network for
Unsupervised Domain Adaptive Segmentation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Diffusion Probabilistic Model (DPM) has emerged as a highly effective
generative model in the field of computer vision. Its intermediate latent
vectors offer rich semantic information, making it an attractive option for
various downstream tasks such as segmentation and detection. In order to
explore its potential further, we have taken a step forward and considered a
more complex scenario in the medical image domain, specifically, under an
unsupervised adaptation condition. To this end, we propose a Diffusion-based
and Prototype-guided network (DP-Net) for unsupervised domain adaptive
segmentation. Concretely, our DP-Net consists of two stages: 1) Distribution
Aligned Diffusion (DADiff), which involves training a domain discriminator to
minimize the difference between the intermediate features generated by the DPM,
thereby aligning the inter-domain distribution; and 2) Prototype-guided
Consistency Learning (PCL), which utilizes feature centroids as prototypes and
applies a prototype-guided loss to ensure that the segmentor learns consistent
content from both source and target domains. Our approach is evaluated on
fundus datasets through a series of experiments, which demonstrate that the
performance of the proposed method is reliable and outperforms state-of-the-art
methods. Our work presents a promising direction for using DPM in complex
medical image scenarios, opening up new possibilities for further research in
medical imaging.
|
[
{
"version": "v1",
"created": "Wed, 22 Mar 2023 05:03:14 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 04:59:00 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 02:59:38 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhou",
"Haipeng",
""
],
[
"Zhu",
"Lei",
""
],
[
"Zhou",
"Yuyin",
""
]
] |
TITLE: Distribution Aligned Diffusion and Prototype-guided network for
Unsupervised Domain Adaptive Segmentation
ABSTRACT: The Diffusion Probabilistic Model (DPM) has emerged as a highly effective
generative model in the field of computer vision. Its intermediate latent
vectors offer rich semantic information, making it an attractive option for
various downstream tasks such as segmentation and detection. In order to
explore its potential further, we have taken a step forward and considered a
more complex scenario in the medical image domain, specifically, under an
unsupervised adaptation condition. To this end, we propose a Diffusion-based
and Prototype-guided network (DP-Net) for unsupervised domain adaptive
segmentation. Concretely, our DP-Net consists of two stages: 1) Distribution
Aligned Diffusion (DADiff), which involves training a domain discriminator to
minimize the difference between the intermediate features generated by the DPM,
thereby aligning the inter-domain distribution; and 2) Prototype-guided
Consistency Learning (PCL), which utilizes feature centroids as prototypes and
applies a prototype-guided loss to ensure that the segmentor learns consistent
content from both source and target domains. Our approach is evaluated on
fundus datasets through a series of experiments, which demonstrate that the
performance of the proposed method is reliable and outperforms state-of-the-art
methods. Our work presents a promising direction for using DPM in complex
medical image scenarios, opening up new possibilities for further research in
medical imaging.
|
no_new_dataset
| 0.709988
|
2303.12888
|
Yuxuan Hu
|
Yuxuan Hu, Albert Lui, Mark Goldstein, Mukund Sudarshan, Andrea
Tinsay, Cindy Tsui, Samuel Maidman, John Medamana, Neil Jethani, Aahlad Puli,
Vuthy Nguy, Yindalon Aphinyanaphongs, Nicholas Kiefer, Nathaniel Smilowitz,
James Horowitz, Tania Ahuja, Glenn I Fishman, Judith Hochman, Stuart Katz,
Samuel Bernard, Rajesh Ranganath
|
A dynamic risk score for early prediction of cardiogenic shock using
machine learning
| null | null | null | null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Myocardial infarction and heart failure are major cardiovascular diseases
that affect millions of people in the US. The morbidity and mortality are
highest among patients who develop cardiogenic shock. Early recognition of
cardiogenic shock is critical. Prompt implementation of treatment measures can
prevent the deleterious spiral of ischemia, low blood pressure, and reduced
cardiac output due to cardiogenic shock. However, early identification of
cardiogenic shock has been challenging due to human providers' inability to
process the enormous amount of data in the cardiac intensive care unit (ICU)
and lack of an effective risk stratification tool. We developed a deep
learning-based risk stratification tool, called CShock, for patients admitted
into the cardiac ICU with acute decompensated heart failure and/or myocardial
infarction to predict onset of cardiogenic shock. To develop and validate
CShock, we annotated cardiac ICU datasets with physician adjudicated outcomes.
CShock achieved an area under the receiver operator characteristic curve
(AUROC) of 0.820, which substantially outperformed CardShock (AUROC 0.519), a
well-established risk score for cardiogenic shock prognosis. CShock was
externally validated in an independent patient cohort and achieved an AUROC of
0.800, demonstrating its generalizability in other cardiac ICUs.
|
[
{
"version": "v1",
"created": "Wed, 22 Mar 2023 20:05:22 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 12:08:42 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Hu",
"Yuxuan",
""
],
[
"Lui",
"Albert",
""
],
[
"Goldstein",
"Mark",
""
],
[
"Sudarshan",
"Mukund",
""
],
[
"Tinsay",
"Andrea",
""
],
[
"Tsui",
"Cindy",
""
],
[
"Maidman",
"Samuel",
""
],
[
"Medamana",
"John",
""
],
[
"Jethani",
"Neil",
""
],
[
"Puli",
"Aahlad",
""
],
[
"Nguy",
"Vuthy",
""
],
[
"Aphinyanaphongs",
"Yindalon",
""
],
[
"Kiefer",
"Nicholas",
""
],
[
"Smilowitz",
"Nathaniel",
""
],
[
"Horowitz",
"James",
""
],
[
"Ahuja",
"Tania",
""
],
[
"Fishman",
"Glenn I",
""
],
[
"Hochman",
"Judith",
""
],
[
"Katz",
"Stuart",
""
],
[
"Bernard",
"Samuel",
""
],
[
"Ranganath",
"Rajesh",
""
]
] |
TITLE: A dynamic risk score for early prediction of cardiogenic shock using
machine learning
ABSTRACT: Myocardial infarction and heart failure are major cardiovascular diseases
that affect millions of people in the US. The morbidity and mortality are
highest among patients who develop cardiogenic shock. Early recognition of
cardiogenic shock is critical. Prompt implementation of treatment measures can
prevent the deleterious spiral of ischemia, low blood pressure, and reduced
cardiac output due to cardiogenic shock. However, early identification of
cardiogenic shock has been challenging due to human providers' inability to
process the enormous amount of data in the cardiac intensive care unit (ICU)
and lack of an effective risk stratification tool. We developed a deep
learning-based risk stratification tool, called CShock, for patients admitted
into the cardiac ICU with acute decompensated heart failure and/or myocardial
infarction to predict onset of cardiogenic shock. To develop and validate
CShock, we annotated cardiac ICU datasets with physician adjudicated outcomes.
CShock achieved an area under the receiver operator characteristic curve
(AUROC) of 0.820, which substantially outperformed CardShock (AUROC 0.519), a
well-established risk score for cardiogenic shock prognosis. CShock was
externally validated in an independent patient cohort and achieved an AUROC of
0.800, demonstrating its generalizability in other cardiac ICUs.
|
no_new_dataset
| 0.70724
|
2303.12999
|
Chaoqun You
|
Chaoqun You, Kun Guo, Gang Feng, Peng Yang, Tony Q. S. Quek
|
Automated Federated Learning in Mobile Edge Networks -- Fast Adaptation
and Convergence
| null | null |
10.1109/JIOT.2023.3262664
| null |
cs.LG cs.DC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Federated Learning (FL) can be used in mobile edge networks to train machine
learning models in a distributed manner. Recently, FL has been interpreted
within a Model-Agnostic Meta-Learning (MAML) framework, which brings FL
significant advantages in fast adaptation and convergence over heterogeneous
datasets. However, existing research simply combines MAML and FL without
explicitly addressing how much benefit MAML brings to FL and how to maximize
such benefit over mobile edge networks. In this paper, we quantify the benefit
from two aspects: optimizing FL hyperparameters (i.e., sampled data size and
the number of communication rounds) and resource allocation (i.e., transmit
power) in mobile edge networks. Specifically, we formulate the MAML-based FL
design as an overall learning time minimization problem, under the constraints
of model accuracy and energy consumption. Facilitated by the convergence
analysis of MAML-based FL, we decompose the formulated problem and then solve
it using analytical solutions and the coordinate descent method. With the
obtained FL hyperparameters and resource allocation, we design a MAML-based FL
algorithm, called Automated Federated Learning (AutoFL), that is able to
conduct fast adaptation and convergence. Extensive experimental results verify
that AutoFL outperforms other benchmark algorithms regarding the learning time
and convergence performance.
|
[
{
"version": "v1",
"created": "Thu, 23 Mar 2023 02:42:10 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"You",
"Chaoqun",
""
],
[
"Guo",
"Kun",
""
],
[
"Feng",
"Gang",
""
],
[
"Yang",
"Peng",
""
],
[
"Quek",
"Tony Q. S.",
""
]
] |
TITLE: Automated Federated Learning in Mobile Edge Networks -- Fast Adaptation
and Convergence
ABSTRACT: Federated Learning (FL) can be used in mobile edge networks to train machine
learning models in a distributed manner. Recently, FL has been interpreted
within a Model-Agnostic Meta-Learning (MAML) framework, which brings FL
significant advantages in fast adaptation and convergence over heterogeneous
datasets. However, existing research simply combines MAML and FL without
explicitly addressing how much benefit MAML brings to FL and how to maximize
such benefit over mobile edge networks. In this paper, we quantify the benefit
from two aspects: optimizing FL hyperparameters (i.e., sampled data size and
the number of communication rounds) and resource allocation (i.e., transmit
power) in mobile edge networks. Specifically, we formulate the MAML-based FL
design as an overall learning time minimization problem, under the constraints
of model accuracy and energy consumption. Facilitated by the convergence
analysis of MAML-based FL, we decompose the formulated problem and then solve
it using analytical solutions and the coordinate descent method. With the
obtained FL hyperparameters and resource allocation, we design a MAML-based FL
algorithm, called Automated Federated Learning (AutoFL), that is able to
conduct fast adaptation and convergence. Extensive experimental results verify
that AutoFL outperforms other benchmark algorithms regarding the learning time
and convergence performance.
|
no_new_dataset
| 0.709265
|
2303.13284
|
Debayan Banerjee
|
Debayan Banerjee, Pranav Ajit Nair, Ricardo Usbeck, Chris Biemann
|
GETT-QA: Graph Embedding based T2T Transformer for Knowledge Graph
Question Answering
|
16 pages single column format accepted at ESWC 2023 research track
| null | null | null |
cs.CL cs.DB cs.IR
|
http://creativecommons.org/licenses/by/4.0/
|
In this work, we present an end-to-end Knowledge Graph Question Answering
(KGQA) system named GETT-QA. GETT-QA uses T5, a popular text-to-text
pre-trained language model. The model takes a question in natural language as
input and produces a simpler form of the intended SPARQL query. In the simpler
form, the model does not directly produce entity and relation IDs. Instead, it
produces corresponding entity and relation labels. The labels are grounded to
KG entity and relation IDs in a subsequent step. To further improve the
results, we instruct the model to produce a truncated version of the KG
embedding for each entity. The truncated KG embedding enables a finer search
for disambiguation purposes. We find that T5 is able to learn the truncated KG
embeddings without any change of loss function, improving KGQA performance. As
a result, we report strong results for LC-QuAD 2.0 and SimpleQuestions-Wikidata
datasets on end-to-end KGQA over Wikidata.
|
[
{
"version": "v1",
"created": "Thu, 23 Mar 2023 14:06:26 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Mar 2023 14:59:15 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Mar 2023 09:48:50 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Banerjee",
"Debayan",
""
],
[
"Nair",
"Pranav Ajit",
""
],
[
"Usbeck",
"Ricardo",
""
],
[
"Biemann",
"Chris",
""
]
] |
TITLE: GETT-QA: Graph Embedding based T2T Transformer for Knowledge Graph
Question Answering
ABSTRACT: In this work, we present an end-to-end Knowledge Graph Question Answering
(KGQA) system named GETT-QA. GETT-QA uses T5, a popular text-to-text
pre-trained language model. The model takes a question in natural language as
input and produces a simpler form of the intended SPARQL query. In the simpler
form, the model does not directly produce entity and relation IDs. Instead, it
produces corresponding entity and relation labels. The labels are grounded to
KG entity and relation IDs in a subsequent step. To further improve the
results, we instruct the model to produce a truncated version of the KG
embedding for each entity. The truncated KG embedding enables a finer search
for disambiguation purposes. We find that T5 is able to learn the truncated KG
embeddings without any change of loss function, improving KGQA performance. As
a result, we report strong results for LC-QuAD 2.0 and SimpleQuestions-Wikidata
datasets on end-to-end KGQA over Wikidata.
|
no_new_dataset
| 0.70978
|
2303.13489
|
Tongzhou Mu
|
Tongzhou Mu, Hao Su
|
Boosting Reinforcement Learning and Planning with Demonstrations: A
Survey
| null | null | null | null |
cs.LG cs.AI cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Although reinforcement learning has seen tremendous success recently, this
kind of trial-and-error learning can be impractical or inefficient in complex
environments. The use of demonstrations, on the other hand, enables agents to
benefit from expert knowledge rather than having to discover the best action to
take through exploration. In this survey, we discuss the advantages of using
demonstrations in sequential decision making, various ways to apply
demonstrations in learning-based decision making paradigms (for example,
reinforcement learning and planning in the learned models), and how to collect
the demonstrations in various scenarios. Additionally, we exemplify a practical
pipeline for generating and utilizing demonstrations in the recently proposed
ManiSkill robot learning benchmark.
|
[
{
"version": "v1",
"created": "Thu, 23 Mar 2023 17:53:44 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 19:25:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Mu",
"Tongzhou",
""
],
[
"Su",
"Hao",
""
]
] |
TITLE: Boosting Reinforcement Learning and Planning with Demonstrations: A
Survey
ABSTRACT: Although reinforcement learning has seen tremendous success recently, this
kind of trial-and-error learning can be impractical or inefficient in complex
environments. The use of demonstrations, on the other hand, enables agents to
benefit from expert knowledge rather than having to discover the best action to
take through exploration. In this survey, we discuss the advantages of using
demonstrations in sequential decision making, various ways to apply
demonstrations in learning-based decision making paradigms (for example,
reinforcement learning and planning in the learned models), and how to collect
the demonstrations in various scenarios. Additionally, we exemplify a practical
pipeline for generating and utilizing demonstrations in the recently proposed
ManiSkill robot learning benchmark.
|
no_new_dataset
| 0.712245
|
2303.14167
|
Yuanbo Yang
|
Yuanbo Yang, Yifei Yang, Hanlei Guo, Rong Xiong, Yue Wang, Yiyi Liao
|
UrbanGIRAFFE: Representing Urban Scenes as Compositional Generative
Neural Feature Fields
|
Project page: https://lv3d.github.io/urbanGIRAFFE
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Generating photorealistic images with controllable camera pose and scene
contents is essential for many applications including AR/VR and simulation.
Despite the fact that rapid progress has been made in 3D-aware generative
models, most existing methods focus on object-centric images and are not
applicable to generating urban scenes for free camera viewpoint control and
scene editing. To address this challenging task, we propose UrbanGIRAFFE, which
uses a coarse 3D panoptic prior, including the layout distribution of
uncountable stuff and countable objects, to guide a 3D-aware generative model.
Our model is compositional and controllable as it breaks down the scene into
stuff, objects, and sky. Using stuff prior in the form of semantic voxel grids,
we build a conditioned stuff generator that effectively incorporates the coarse
semantic and geometry information. The object layout prior further allows us to
learn an object generator from cluttered scenes. With proper loss functions,
our approach facilitates photorealistic 3D-aware image synthesis with diverse
controllability, including large camera movement, stuff editing, and object
manipulation. We validate the effectiveness of our model on both synthetic and
real-world datasets, including the challenging KITTI-360 dataset.
|
[
{
"version": "v1",
"created": "Fri, 24 Mar 2023 17:28:07 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 02:00:39 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Yang",
"Yuanbo",
""
],
[
"Yang",
"Yifei",
""
],
[
"Guo",
"Hanlei",
""
],
[
"Xiong",
"Rong",
""
],
[
"Wang",
"Yue",
""
],
[
"Liao",
"Yiyi",
""
]
] |
TITLE: UrbanGIRAFFE: Representing Urban Scenes as Compositional Generative
Neural Feature Fields
ABSTRACT: Generating photorealistic images with controllable camera pose and scene
contents is essential for many applications including AR/VR and simulation.
Despite the fact that rapid progress has been made in 3D-aware generative
models, most existing methods focus on object-centric images and are not
applicable to generating urban scenes for free camera viewpoint control and
scene editing. To address this challenging task, we propose UrbanGIRAFFE, which
uses a coarse 3D panoptic prior, including the layout distribution of
uncountable stuff and countable objects, to guide a 3D-aware generative model.
Our model is compositional and controllable as it breaks down the scene into
stuff, objects, and sky. Using stuff prior in the form of semantic voxel grids,
we build a conditioned stuff generator that effectively incorporates the coarse
semantic and geometry information. The object layout prior further allows us to
learn an object generator from cluttered scenes. With proper loss functions,
our approach facilitates photorealistic 3D-aware image synthesis with diverse
controllability, including large camera movement, stuff editing, and object
manipulation. We validate the effectiveness of our model on both synthetic and
real-world datasets, including the challenging KITTI-360 dataset.
|
no_new_dataset
| 0.709057
|
2303.15101
|
Zongrui Li
|
Zongrui Li, Qian Zheng, Boxin Shi, Gang Pan, Xudong Jiang
|
DANI-Net: Uncalibrated Photometric Stereo by Differentiable Shadow
Handling, Anisotropic Reflectance Modeling, and Neural Inverse Rendering
|
Accepted by CVPR 2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Uncalibrated photometric stereo (UPS) is challenging due to the inherent
ambiguity brought by the unknown light. Although the ambiguity is alleviated on
non-Lambertian objects, the problem is still difficult to solve for more
general objects with complex shapes introducing irregular shadows and general
materials with complex reflectance like anisotropic reflectance. To exploit
cues from shadow and reflectance to solve UPS and improve performance on
general materials, we propose DANI-Net, an inverse rendering framework with
differentiable shadow handling and anisotropic reflectance modeling. Unlike
most previous methods that use non-differentiable shadow maps and assume
isotropic material, our network benefits from cues of shadow and anisotropic
reflectance through two differentiable paths. Experiments on multiple
real-world datasets demonstrate our superior and robust performance.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 11:10:54 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 07:14:47 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Li",
"Zongrui",
""
],
[
"Zheng",
"Qian",
""
],
[
"Shi",
"Boxin",
""
],
[
"Pan",
"Gang",
""
],
[
"Jiang",
"Xudong",
""
]
] |
TITLE: DANI-Net: Uncalibrated Photometric Stereo by Differentiable Shadow
Handling, Anisotropic Reflectance Modeling, and Neural Inverse Rendering
ABSTRACT: Uncalibrated photometric stereo (UPS) is challenging due to the inherent
ambiguity brought by the unknown light. Although the ambiguity is alleviated on
non-Lambertian objects, the problem is still difficult to solve for more
general objects with complex shapes introducing irregular shadows and general
materials with complex reflectance like anisotropic reflectance. To exploit
cues from shadow and reflectance to solve UPS and improve performance on
general materials, we propose DANI-Net, an inverse rendering framework with
differentiable shadow handling and anisotropic reflectance modeling. Unlike
most previous methods that use non-differentiable shadow maps and assume
isotropic material, our network benefits from cues of shadow and anisotropic
reflectance through two differentiable paths. Experiments on multiple
real-world datasets demonstrate our superior and robust performance.
|
no_new_dataset
| 0.711857
|
2303.15140
|
Zhikang Liu
|
Zhikang Liu, Yiming Zhou, Yuansheng Xu, Zilei Wang
|
SimpleNet: A Simple Network for Image Anomaly Detection and Localization
|
Accepted to CVPR 2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a simple and application-friendly network (called SimpleNet) for
detecting and localizing anomalies. SimpleNet consists of four components: (1)
a pre-trained Feature Extractor that generates local features, (2) a shallow
Feature Adapter that transfers local features towards target domain, (3) a
simple Anomaly Feature Generator that counterfeits anomaly features by adding
Gaussian noise to normal features, and (4) a binary Anomaly Discriminator that
distinguishes anomaly features from normal features. During inference, the
Anomaly Feature Generator would be discarded. Our approach is based on three
intuitions. First, transforming pre-trained features to target-oriented
features helps avoid domain bias. Second, generating synthetic anomalies in
feature space is more effective, as defects may not have much commonality in
the image space. Third, a simple discriminator is much efficient and practical.
In spite of simplicity, SimpleNet outperforms previous methods quantitatively
and qualitatively. On the MVTec AD benchmark, SimpleNet achieves an anomaly
detection AUROC of 99.6%, reducing the error by 55.5% compared to the next best
performing model. Furthermore, SimpleNet is faster than existing methods, with
a high frame rate of 77 FPS on a 3080ti GPU. Additionally, SimpleNet
demonstrates significant improvements in performance on the One-Class Novelty
Detection task. Code: https://github.com/DonaldRR/SimpleNet.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 12:18:52 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Mar 2023 04:02:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Liu",
"Zhikang",
""
],
[
"Zhou",
"Yiming",
""
],
[
"Xu",
"Yuansheng",
""
],
[
"Wang",
"Zilei",
""
]
] |
TITLE: SimpleNet: A Simple Network for Image Anomaly Detection and Localization
ABSTRACT: We propose a simple and application-friendly network (called SimpleNet) for
detecting and localizing anomalies. SimpleNet consists of four components: (1)
a pre-trained Feature Extractor that generates local features, (2) a shallow
Feature Adapter that transfers local features towards target domain, (3) a
simple Anomaly Feature Generator that counterfeits anomaly features by adding
Gaussian noise to normal features, and (4) a binary Anomaly Discriminator that
distinguishes anomaly features from normal features. During inference, the
Anomaly Feature Generator would be discarded. Our approach is based on three
intuitions. First, transforming pre-trained features to target-oriented
features helps avoid domain bias. Second, generating synthetic anomalies in
feature space is more effective, as defects may not have much commonality in
the image space. Third, a simple discriminator is much efficient and practical.
In spite of simplicity, SimpleNet outperforms previous methods quantitatively
and qualitatively. On the MVTec AD benchmark, SimpleNet achieves an anomaly
detection AUROC of 99.6%, reducing the error by 55.5% compared to the next best
performing model. Furthermore, SimpleNet is faster than existing methods, with
a high frame rate of 77 FPS on a 3080ti GPU. Additionally, SimpleNet
demonstrates significant improvements in performance on the One-Class Novelty
Detection task. Code: https://github.com/DonaldRR/SimpleNet.
|
no_new_dataset
| 0.709806
|
2303.15467
|
Jun Cen
|
Jun Cen, Shiwei Zhang, Xiang Wang, Yixuan Pei, Zhiwu Qing, Yingya
Zhang, Qifeng Chen
|
Enlarging Instance-specific and Class-specific Information for Open-set
Action Recognition
|
To appear at CVPR2023
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Open-set action recognition is to reject unknown human action cases which are
out of the distribution of the training set. Existing methods mainly focus on
learning better uncertainty scores but dismiss the importance of feature
representations. We find that features with richer semantic diversity can
significantly improve the open-set performance under the same uncertainty
scores. In this paper, we begin with analyzing the feature representation
behavior in the open-set action recognition (OSAR) problem based on the
information bottleneck (IB) theory, and propose to enlarge the
instance-specific (IS) and class-specific (CS) information contained in the
feature for better performance. To this end, a novel Prototypical Similarity
Learning (PSL) framework is proposed to keep the instance variance within the
same class to retain more IS information. Besides, we notice that unknown
samples sharing similar appearances to known samples are easily misclassified
as known classes. To alleviate this issue, video shuffling is further
introduced in our PSL to learn distinct temporal information between original
and shuffled samples, which we find enlarges the CS information. Extensive
experiments demonstrate that the proposed PSL can significantly boost both the
open-set and closed-set performance and achieves state-of-the-art results on
multiple benchmarks. Code is available at https://github.com/Jun-CEN/PSL.
|
[
{
"version": "v1",
"created": "Sat, 25 Mar 2023 04:07:36 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Cen",
"Jun",
""
],
[
"Zhang",
"Shiwei",
""
],
[
"Wang",
"Xiang",
""
],
[
"Pei",
"Yixuan",
""
],
[
"Qing",
"Zhiwu",
""
],
[
"Zhang",
"Yingya",
""
],
[
"Chen",
"Qifeng",
""
]
] |
TITLE: Enlarging Instance-specific and Class-specific Information for Open-set
Action Recognition
ABSTRACT: Open-set action recognition is to reject unknown human action cases which are
out of the distribution of the training set. Existing methods mainly focus on
learning better uncertainty scores but dismiss the importance of feature
representations. We find that features with richer semantic diversity can
significantly improve the open-set performance under the same uncertainty
scores. In this paper, we begin with analyzing the feature representation
behavior in the open-set action recognition (OSAR) problem based on the
information bottleneck (IB) theory, and propose to enlarge the
instance-specific (IS) and class-specific (CS) information contained in the
feature for better performance. To this end, a novel Prototypical Similarity
Learning (PSL) framework is proposed to keep the instance variance within the
same class to retain more IS information. Besides, we notice that unknown
samples sharing similar appearances to known samples are easily misclassified
as known classes. To alleviate this issue, video shuffling is further
introduced in our PSL to learn distinct temporal information between original
and shuffled samples, which we find enlarges the CS information. Extensive
experiments demonstrate that the proposed PSL can significantly boost both the
open-set and closed-set performance and achieves state-of-the-art results on
multiple benchmarks. Code is available at https://github.com/Jun-CEN/PSL.
|
no_new_dataset
| 0.7082
|
2303.15486
|
Rongyu Zhang
|
Rongyu Zhang, Xiaowei Chi, Guiliang Liu, Wenyi Zhang, Yuan Du, Fangxin
Wang
|
Unimodal Training-Multimodal Prediction: Cross-modal Federated Learning
with Hierarchical Aggregation
|
10 pages,5 figures
| null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Multimodal learning has seen great success mining data features from multiple
modalities with remarkable model performance improvement. Meanwhile, federated
learning (FL) addresses the data sharing problem, enabling privacy-preserved
collaborative training to provide sufficient precious data. Great potential,
therefore, arises with the confluence of them, known as multimodal federated
learning. However, limitation lies in the predominant approaches as they often
assume that each local dataset records samples from all modalities. In this
paper, we aim to bridge this gap by proposing an Unimodal Training - Multimodal
Prediction (UTMP) framework under the context of multimodal federated learning.
We design HA-Fedformer, a novel transformer-based model that empowers unimodal
training with only a unimodal dataset at the client and multimodal testing by
aggregating multiple clients' knowledge for better accuracy. The key advantages
are twofold. Firstly, to alleviate the impact of data non-IID, we develop an
uncertainty-aware aggregation method for the local encoders with layer-wise
Markov Chain Monte Carlo sampling. Secondly, to overcome the challenge of
unaligned language sequence, we implement a cross-modal decoder aggregation to
capture the hidden signal correlation between decoders trained by data from
different modalities. Our experiments on popular sentiment analysis benchmarks,
CMU-MOSI and CMU-MOSEI, demonstrate that HA-Fedformer significantly outperforms
state-of-the-art multimodal models under the UTMP federated learning
frameworks, with 15%-20% improvement on most attributes.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 07:07:33 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhang",
"Rongyu",
""
],
[
"Chi",
"Xiaowei",
""
],
[
"Liu",
"Guiliang",
""
],
[
"Zhang",
"Wenyi",
""
],
[
"Du",
"Yuan",
""
],
[
"Wang",
"Fangxin",
""
]
] |
TITLE: Unimodal Training-Multimodal Prediction: Cross-modal Federated Learning
with Hierarchical Aggregation
ABSTRACT: Multimodal learning has seen great success mining data features from multiple
modalities with remarkable model performance improvement. Meanwhile, federated
learning (FL) addresses the data sharing problem, enabling privacy-preserved
collaborative training to provide sufficient precious data. Great potential,
therefore, arises with the confluence of them, known as multimodal federated
learning. However, limitation lies in the predominant approaches as they often
assume that each local dataset records samples from all modalities. In this
paper, we aim to bridge this gap by proposing an Unimodal Training - Multimodal
Prediction (UTMP) framework under the context of multimodal federated learning.
We design HA-Fedformer, a novel transformer-based model that empowers unimodal
training with only a unimodal dataset at the client and multimodal testing by
aggregating multiple clients' knowledge for better accuracy. The key advantages
are twofold. Firstly, to alleviate the impact of data non-IID, we develop an
uncertainty-aware aggregation method for the local encoders with layer-wise
Markov Chain Monte Carlo sampling. Secondly, to overcome the challenge of
unaligned language sequence, we implement a cross-modal decoder aggregation to
capture the hidden signal correlation between decoders trained by data from
different modalities. Our experiments on popular sentiment analysis benchmarks,
CMU-MOSI and CMU-MOSEI, demonstrate that HA-Fedformer significantly outperforms
state-of-the-art multimodal models under the UTMP federated learning
frameworks, with 15%-20% improvement on most attributes.
|
no_new_dataset
| 0.709252
|
2303.15491
|
Robin Georg Claus Maack
|
Robin G. C. Maack, Jonas Lukasczyk, Julien Tierny, Hans Hagen, Ross
Maciejewski, Christoph Garth
|
Parallel Computation of Piecewise Linear Morse-Smale Segmentations
|
Journal: IEEE Transactions on Visualization and Computer Graphics /
Submitted: 22-Jun-2022 / Accepted: 13-Mar-2023
| null |
10.1109/TVCG.2023.3261981
| null |
cs.GR
|
http://creativecommons.org/licenses/by-sa/4.0/
|
This paper presents a well-scaling parallel algorithm for the computation of
Morse-Smale (MS) segmentations, including the region separators and region
boundaries. The segmentation of the domain into ascending and descending
manifolds, solely defined on the vertices, improves the computational time
using path compression and fully segments the border region. Region boundaries
and region separators are generated using a multi-label marching tetrahedra
algorithm. This enables a fast and simple solution to find optimal parameter
settings in preliminary exploration steps by generating an MS complex preview.
It also poses a rapid option to generate a fast visual representation of the
region geometries for immediate utilization. Two experiments demonstrate the
performance of our approach with speedups of over an order of magnitude in
comparison to two publicly available implementations. The example section shows
the similarity to the MS complex, the useability of the approach, and the
benefits of this method with respect to the presented datasets. We provide our
implementation with the paper.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 13:33:07 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Maack",
"Robin G. C.",
""
],
[
"Lukasczyk",
"Jonas",
""
],
[
"Tierny",
"Julien",
""
],
[
"Hagen",
"Hans",
""
],
[
"Maciejewski",
"Ross",
""
],
[
"Garth",
"Christoph",
""
]
] |
TITLE: Parallel Computation of Piecewise Linear Morse-Smale Segmentations
ABSTRACT: This paper presents a well-scaling parallel algorithm for the computation of
Morse-Smale (MS) segmentations, including the region separators and region
boundaries. The segmentation of the domain into ascending and descending
manifolds, solely defined on the vertices, improves the computational time
using path compression and fully segments the border region. Region boundaries
and region separators are generated using a multi-label marching tetrahedra
algorithm. This enables a fast and simple solution to find optimal parameter
settings in preliminary exploration steps by generating an MS complex preview.
It also poses a rapid option to generate a fast visual representation of the
region geometries for immediate utilization. Two experiments demonstrate the
performance of our approach with speedups of over an order of magnitude in
comparison to two publicly available implementations. The example section shows
the similarity to the MS complex, the useability of the approach, and the
benefits of this method with respect to the presented datasets. We provide our
implementation with the paper.
|
no_new_dataset
| 0.708603
|
2303.15494
|
Sichao Fu
|
Wenhao Qiu, Sichao Fu, Jingyi Zhang, Chengxiang Lei, Qinmu Peng
|
Semantic-visual Guided Transformer for Few-shot Class-incremental
Learning
|
Accepted by IEEE International Conference on Multimedia and Expo
(ICME 2023)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Few-shot class-incremental learning (FSCIL) has recently attracted extensive
attention in various areas. Existing FSCIL methods highly depend on the
robustness of the feature backbone pre-trained on base classes. In recent
years, different Transformer variants have obtained significant processes in
the feature representation learning of massive fields. Nevertheless, the
progress of the Transformer in FSCIL scenarios has not achieved the potential
promised in other fields so far. In this paper, we develop a semantic-visual
guided Transformer (SV-T) to enhance the feature extracting capacity of the
pre-trained feature backbone on incremental classes. Specifically, we first
utilize the visual (image) labels provided by the base classes to supervise the
optimization of the Transformer. And then, a text encoder is introduced to
automatically generate the corresponding semantic (text) labels for each image
from the base classes. Finally, the constructed semantic labels are further
applied to the Transformer for guiding its hyperparameters updating. Our SV-T
can take full advantage of more supervision information from base classes and
further enhance the training robustness of the feature backbone. More
importantly, our SV-T is an independent method, which can directly apply to the
existing FSCIL architectures for acquiring embeddings of various incremental
classes. Extensive experiments on three benchmarks, two FSCIL architectures,
and two Transformer variants show that our proposed SV-T obtains a significant
improvement in comparison to the existing state-of-the-art FSCIL methods.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 15:06:49 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Qiu",
"Wenhao",
""
],
[
"Fu",
"Sichao",
""
],
[
"Zhang",
"Jingyi",
""
],
[
"Lei",
"Chengxiang",
""
],
[
"Peng",
"Qinmu",
""
]
] |
TITLE: Semantic-visual Guided Transformer for Few-shot Class-incremental
Learning
ABSTRACT: Few-shot class-incremental learning (FSCIL) has recently attracted extensive
attention in various areas. Existing FSCIL methods highly depend on the
robustness of the feature backbone pre-trained on base classes. In recent
years, different Transformer variants have obtained significant processes in
the feature representation learning of massive fields. Nevertheless, the
progress of the Transformer in FSCIL scenarios has not achieved the potential
promised in other fields so far. In this paper, we develop a semantic-visual
guided Transformer (SV-T) to enhance the feature extracting capacity of the
pre-trained feature backbone on incremental classes. Specifically, we first
utilize the visual (image) labels provided by the base classes to supervise the
optimization of the Transformer. And then, a text encoder is introduced to
automatically generate the corresponding semantic (text) labels for each image
from the base classes. Finally, the constructed semantic labels are further
applied to the Transformer for guiding its hyperparameters updating. Our SV-T
can take full advantage of more supervision information from base classes and
further enhance the training robustness of the feature backbone. More
importantly, our SV-T is an independent method, which can directly apply to the
existing FSCIL architectures for acquiring embeddings of various incremental
classes. Extensive experiments on three benchmarks, two FSCIL architectures,
and two Transformer variants show that our proposed SV-T obtains a significant
improvement in comparison to the existing state-of-the-art FSCIL methods.
|
no_new_dataset
| 0.710415
|
2303.15528
|
Vishal Vinod
|
K. Ram Prabhakar, Vishal Vinod, Nihar Ranjan Sahoo, R. Venkatesh Babu
|
Few-Shot Domain Adaptation for Low Light RAW Image Enhancement
|
BMVC 2021 Best Student Paper Award (Runner-Up). Project Page:
https://val.cds.iisc.ac.in/HDR/BMVC21/index.html
|
32nd British Machine Vision Conference 2021, BMVC 2021, 327
| null | null |
cs.CV eess.IV
|
http://creativecommons.org/licenses/by/4.0/
|
Enhancing practical low light raw images is a difficult task due to severe
noise and color distortions from short exposure time and limited illumination.
Despite the success of existing Convolutional Neural Network (CNN) based
methods, their performance is not adaptable to different camera domains. In
addition, such methods also require large datasets with short-exposure and
corresponding long-exposure ground truth raw images for each camera domain,
which is tedious to compile. To address this issue, we present a novel few-shot
domain adaptation method to utilize the existing source camera labeled data
with few labeled samples from the target camera to improve the target domain's
enhancement quality in extreme low-light imaging. Our experiments show that
only ten or fewer labeled samples from the target camera domain are sufficient
to achieve similar or better enhancement performance than training a model with
a large labeled target camera dataset. To support research in this direction,
we also present a new low-light raw image dataset captured with a Nikon camera,
comprising short-exposure and their corresponding long-exposure ground truth
images.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 18:10:52 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Prabhakar",
"K. Ram",
""
],
[
"Vinod",
"Vishal",
""
],
[
"Sahoo",
"Nihar Ranjan",
""
],
[
"Babu",
"R. Venkatesh",
""
]
] |
TITLE: Few-Shot Domain Adaptation for Low Light RAW Image Enhancement
ABSTRACT: Enhancing practical low light raw images is a difficult task due to severe
noise and color distortions from short exposure time and limited illumination.
Despite the success of existing Convolutional Neural Network (CNN) based
methods, their performance is not adaptable to different camera domains. In
addition, such methods also require large datasets with short-exposure and
corresponding long-exposure ground truth raw images for each camera domain,
which is tedious to compile. To address this issue, we present a novel few-shot
domain adaptation method to utilize the existing source camera labeled data
with few labeled samples from the target camera to improve the target domain's
enhancement quality in extreme low-light imaging. Our experiments show that
only ten or fewer labeled samples from the target camera domain are sufficient
to achieve similar or better enhancement performance than training a model with
a large labeled target camera dataset. To support research in this direction,
we also present a new low-light raw image dataset captured with a Nikon camera,
comprising short-exposure and their corresponding long-exposure ground truth
images.
|
new_dataset
| 0.713188
|
2303.15555
|
Zhipeng Bao
|
Zhipeng Bao, Pavel Tokmakov, Yu-Xiong Wang, Adrien Gaidon, and Martial
Hebert
|
Object Discovery from Motion-Guided Tokens
| null |
CVPR 2023
| null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Object discovery -- separating objects from the background without manual
labels -- is a fundamental open challenge in computer vision. Previous methods
struggle to go beyond clustering of low-level cues, whether handcrafted (e.g.,
color, texture) or learned (e.g., from auto-encoders). In this work, we augment
the auto-encoder representation learning framework with two key components:
motion-guidance and mid-level feature tokenization. Although both have been
separately investigated, we introduce a new transformer decoder showing that
their benefits can compound thanks to motion-guided vector quantization. We
show that our architecture effectively leverages the synergy between motion and
tokenization, improving upon the state of the art on both synthetic and real
datasets. Our approach enables the emergence of interpretable object-specific
mid-level features, demonstrating the benefits of motion-guidance (no labeling)
and quantization (interpretability, memory efficiency).
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 19:14:00 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Bao",
"Zhipeng",
""
],
[
"Tokmakov",
"Pavel",
""
],
[
"Wang",
"Yu-Xiong",
""
],
[
"Gaidon",
"Adrien",
""
],
[
"Hebert",
"Martial",
""
]
] |
TITLE: Object Discovery from Motion-Guided Tokens
ABSTRACT: Object discovery -- separating objects from the background without manual
labels -- is a fundamental open challenge in computer vision. Previous methods
struggle to go beyond clustering of low-level cues, whether handcrafted (e.g.,
color, texture) or learned (e.g., from auto-encoders). In this work, we augment
the auto-encoder representation learning framework with two key components:
motion-guidance and mid-level feature tokenization. Although both have been
separately investigated, we introduce a new transformer decoder showing that
their benefits can compound thanks to motion-guided vector quantization. We
show that our architecture effectively leverages the synergy between motion and
tokenization, improving upon the state of the art on both synthetic and real
datasets. Our approach enables the emergence of interpretable object-specific
mid-level features, demonstrating the benefits of motion-guidance (no labeling)
and quantization (interpretability, memory efficiency).
|
no_new_dataset
| 0.7023
|
2303.15569
|
Xiaowei Yu
|
Xiaowei Yu, Lu Zhang, Haixing Dai, Yanjun Lyu, Lin Zhao, Zihao Wu,
David Liu, Tianming Liu and Dajiang Zhu
|
Core-Periphery Principle Guided Redesign of Self-Attention in
Transformers
|
Core-periphery, functional brain networks, ViT
| null | null | null |
cs.LG q-bio.NC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Designing more efficient, reliable, and explainable neural network
architectures is critical to studies that are based on artificial intelligence
(AI) techniques. Previous studies, by post-hoc analysis, have found that the
best-performing ANNs surprisingly resemble biological neural networks (BNN),
which indicates that ANNs and BNNs may share some common principles to achieve
optimal performance in either machine learning or cognitive/behavior tasks.
Inspired by this phenomenon, we proactively instill organizational principles
of BNNs to guide the redesign of ANNs. We leverage the Core-Periphery (CP)
organization, which is widely found in human brain networks, to guide the
information communication mechanism in the self-attention of vision transformer
(ViT) and name this novel framework as CP-ViT. In CP-ViT, the attention
operation between nodes is defined by a sparse graph with a Core-Periphery
structure (CP graph), where the core nodes are redesigned and reorganized to
play an integrative role and serve as a center for other periphery nodes to
exchange information. We evaluated the proposed CP-ViT on multiple public
datasets, including medical image datasets (INbreast) and natural image
datasets. Interestingly, by incorporating the BNN-derived principle (CP
structure) into the redesign of ViT, our CP-ViT outperforms other
state-of-the-art ANNs. In general, our work advances the state of the art in
three aspects: 1) This work provides novel insights for brain-inspired AI: we
can utilize the principles found in BNNs to guide and improve our ANN
architecture design; 2) We show that there exist sweet spots of CP graphs that
lead to CP-ViTs with significantly improved performance; and 3) The core nodes
in CP-ViT correspond to task-related meaningful and important image patches,
which can significantly enhance the interpretability of the trained deep model.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 19:36:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Yu",
"Xiaowei",
""
],
[
"Zhang",
"Lu",
""
],
[
"Dai",
"Haixing",
""
],
[
"Lyu",
"Yanjun",
""
],
[
"Zhao",
"Lin",
""
],
[
"Wu",
"Zihao",
""
],
[
"Liu",
"David",
""
],
[
"Liu",
"Tianming",
""
],
[
"Zhu",
"Dajiang",
""
]
] |
TITLE: Core-Periphery Principle Guided Redesign of Self-Attention in
Transformers
ABSTRACT: Designing more efficient, reliable, and explainable neural network
architectures is critical to studies that are based on artificial intelligence
(AI) techniques. Previous studies, by post-hoc analysis, have found that the
best-performing ANNs surprisingly resemble biological neural networks (BNN),
which indicates that ANNs and BNNs may share some common principles to achieve
optimal performance in either machine learning or cognitive/behavior tasks.
Inspired by this phenomenon, we proactively instill organizational principles
of BNNs to guide the redesign of ANNs. We leverage the Core-Periphery (CP)
organization, which is widely found in human brain networks, to guide the
information communication mechanism in the self-attention of vision transformer
(ViT) and name this novel framework as CP-ViT. In CP-ViT, the attention
operation between nodes is defined by a sparse graph with a Core-Periphery
structure (CP graph), where the core nodes are redesigned and reorganized to
play an integrative role and serve as a center for other periphery nodes to
exchange information. We evaluated the proposed CP-ViT on multiple public
datasets, including medical image datasets (INbreast) and natural image
datasets. Interestingly, by incorporating the BNN-derived principle (CP
structure) into the redesign of ViT, our CP-ViT outperforms other
state-of-the-art ANNs. In general, our work advances the state of the art in
three aspects: 1) This work provides novel insights for brain-inspired AI: we
can utilize the principles found in BNNs to guide and improve our ANN
architecture design; 2) We show that there exist sweet spots of CP graphs that
lead to CP-ViTs with significantly improved performance; and 3) The core nodes
in CP-ViT correspond to task-related meaningful and important image patches,
which can significantly enhance the interpretability of the trained deep model.
|
no_new_dataset
| 0.709422
|
2303.15570
|
Christian Remi Wewer
|
Christian Remi Wewer and Alexandros Iosifidis
|
Online Non-Destructive Moisture Content Estimation of Filter Media
During Drying Using Artificial Neural Networks
|
11 pages, 9 figures
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Moisture content (MC) estimation is important in the manufacturing process of
drying bulky filter media products as it is the prerequisite for drying
optimization. In this study, a dataset collected by performing 161 drying
industrial experiments is described and a methodology for MC estimation in an
non-destructive and online manner during industrial drying is presented. An
artificial neural network (ANN) based method is compared to state-of-the-art MC
estimation methods reported in the literature. Results of model fitting and
training show that a three-layer Perceptron achieves the lowest error.
Experimental results show that ANNs combined with oven settings data, drying
time and product temperature can be used to reliably estimate the MC of bulky
filter media products.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 19:37:53 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wewer",
"Christian Remi",
""
],
[
"Iosifidis",
"Alexandros",
""
]
] |
TITLE: Online Non-Destructive Moisture Content Estimation of Filter Media
During Drying Using Artificial Neural Networks
ABSTRACT: Moisture content (MC) estimation is important in the manufacturing process of
drying bulky filter media products as it is the prerequisite for drying
optimization. In this study, a dataset collected by performing 161 drying
industrial experiments is described and a methodology for MC estimation in an
non-destructive and online manner during industrial drying is presented. An
artificial neural network (ANN) based method is compared to state-of-the-art MC
estimation methods reported in the literature. Results of model fitting and
training show that a three-layer Perceptron achieves the lowest error.
Experimental results show that ANNs combined with oven settings data, drying
time and product temperature can be used to reliably estimate the MC of bulky
filter media products.
|
no_new_dataset
| 0.711249
|
2303.15571
|
Ruyi Ding
|
Ruyi Ding, Cheng Gongye, Siyue Wang, Aidong Ding, Yunsi Fei
|
EMShepherd: Detecting Adversarial Samples via Side-channel Leakage
| null | null | null | null |
cs.CR cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Deep Neural Networks (DNN) are vulnerable to adversarial perturbations-small
changes crafted deliberately on the input to mislead the model for wrong
predictions. Adversarial attacks have disastrous consequences for deep
learning-empowered critical applications. Existing defense and detection
techniques both require extensive knowledge of the model, testing inputs, and
even execution details. They are not viable for general deep learning
implementations where the model internal is unknown, a common 'black-box'
scenario for model users. Inspired by the fact that electromagnetic (EM)
emanations of a model inference are dependent on both operations and data and
may contain footprints of different input classes, we propose a framework,
EMShepherd, to capture EM traces of model execution, perform processing on
traces and exploit them for adversarial detection. Only benign samples and
their EM traces are used to train the adversarial detector: a set of EM
classifiers and class-specific unsupervised anomaly detectors. When the victim
model system is under attack by an adversarial example, the model execution
will be different from executions for the known classes, and the EM trace will
be different. We demonstrate that our air-gapped EMShepherd can effectively
detect different adversarial attacks on a commonly used FPGA deep learning
accelerator for both Fashion MNIST and CIFAR-10 datasets. It achieves a 100%
detection rate on most types of adversarial samples, which is comparable to the
state-of-the-art 'white-box' software-based detectors.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 19:38:55 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Ding",
"Ruyi",
""
],
[
"Gongye",
"Cheng",
""
],
[
"Wang",
"Siyue",
""
],
[
"Ding",
"Aidong",
""
],
[
"Fei",
"Yunsi",
""
]
] |
TITLE: EMShepherd: Detecting Adversarial Samples via Side-channel Leakage
ABSTRACT: Deep Neural Networks (DNN) are vulnerable to adversarial perturbations-small
changes crafted deliberately on the input to mislead the model for wrong
predictions. Adversarial attacks have disastrous consequences for deep
learning-empowered critical applications. Existing defense and detection
techniques both require extensive knowledge of the model, testing inputs, and
even execution details. They are not viable for general deep learning
implementations where the model internal is unknown, a common 'black-box'
scenario for model users. Inspired by the fact that electromagnetic (EM)
emanations of a model inference are dependent on both operations and data and
may contain footprints of different input classes, we propose a framework,
EMShepherd, to capture EM traces of model execution, perform processing on
traces and exploit them for adversarial detection. Only benign samples and
their EM traces are used to train the adversarial detector: a set of EM
classifiers and class-specific unsupervised anomaly detectors. When the victim
model system is under attack by an adversarial example, the model execution
will be different from executions for the known classes, and the EM trace will
be different. We demonstrate that our air-gapped EMShepherd can effectively
detect different adversarial attacks on a commonly used FPGA deep learning
accelerator for both Fashion MNIST and CIFAR-10 datasets. It achieves a 100%
detection rate on most types of adversarial samples, which is comparable to the
state-of-the-art 'white-box' software-based detectors.
|
no_new_dataset
| 0.709208
|
2303.15583
|
Chin Tseng
|
Chin Tseng, Ghulam Jilani Quadri, Zeyu Wang, Danielle Albers Szafir
|
Measuring Categorical Perception in Color-Coded Scatterplots
|
The paper has been accepted to the ACM CHI 2023. 14 pages, 7 figures
| null |
10.1145/3544548.3581416
| null |
cs.HC
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Scatterplots commonly use color to encode categorical data. However, as
datasets increase in size and complexity, the efficacy of these channels may
vary. Designers lack insight into how robust different design choices are to
variations in category numbers. This paper presents a crowdsourced experiment
measuring how the number of categories and choice of color encodings used in
multiclass scatterplots influences the viewers' abilities to analyze data
across classes. Participants estimated relative means in a series of
scatterplots with 2 to 10 categories encoded using ten color palettes drawn
from popular design tools. Our results show that the number of categories and
color discriminability within a color palette notably impact people's
perception of categorical data in scatterplots and that the judgments become
harder as the number of categories grows. We examine existing palette design
heuristics in light of our results to help designers make robust color choices
informed by the parameters of their data.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 20:26:46 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Tseng",
"Chin",
""
],
[
"Quadri",
"Ghulam Jilani",
""
],
[
"Wang",
"Zeyu",
""
],
[
"Szafir",
"Danielle Albers",
""
]
] |
TITLE: Measuring Categorical Perception in Color-Coded Scatterplots
ABSTRACT: Scatterplots commonly use color to encode categorical data. However, as
datasets increase in size and complexity, the efficacy of these channels may
vary. Designers lack insight into how robust different design choices are to
variations in category numbers. This paper presents a crowdsourced experiment
measuring how the number of categories and choice of color encodings used in
multiclass scatterplots influences the viewers' abilities to analyze data
across classes. Participants estimated relative means in a series of
scatterplots with 2 to 10 categories encoded using ten color palettes drawn
from popular design tools. Our results show that the number of categories and
color discriminability within a color palette notably impact people's
perception of categorical data in scatterplots and that the judgments become
harder as the number of categories grows. We examine existing palette design
heuristics in light of our results to help designers make robust color choices
informed by the parameters of their data.
|
no_new_dataset
| 0.704986
|
2303.15591
|
Rajshekhar Das
|
Rajshekhar Das, Yonatan Dukler, Avinash Ravichandran, Ashwin
Swaminathan
|
Learning Expressive Prompting With Residuals for Vision Transformers
|
Accepted at CVPR (2023)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Prompt learning is an efficient approach to adapt transformers by inserting
learnable set of parameters into the input and intermediate representations of
a pre-trained model. In this work, we present Expressive Prompts with Residuals
(EXPRES) which modifies the prompt learning paradigm specifically for effective
adaptation of vision transformers (ViT). Out method constructs downstream
representations via learnable ``output'' tokens, that are akin to the learned
class tokens of the ViT. Further for better steering of the downstream
representation processed by the frozen transformer, we introduce residual
learnable tokens that are added to the output of various computations. We apply
EXPRES for image classification, few shot learning, and semantic segmentation,
and show our method is capable of achieving state of the art prompt tuning on
3/3 categories of the VTAB benchmark. In addition to strong performance, we
observe that our approach is an order of magnitude more prompt efficient than
existing visual prompting baselines. We analytically show the computational
benefits of our approach over weight space adaptation techniques like
finetuning. Lastly we systematically corroborate the architectural design of
our method via a series of ablation experiments.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 20:47:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Das",
"Rajshekhar",
""
],
[
"Dukler",
"Yonatan",
""
],
[
"Ravichandran",
"Avinash",
""
],
[
"Swaminathan",
"Ashwin",
""
]
] |
TITLE: Learning Expressive Prompting With Residuals for Vision Transformers
ABSTRACT: Prompt learning is an efficient approach to adapt transformers by inserting
learnable set of parameters into the input and intermediate representations of
a pre-trained model. In this work, we present Expressive Prompts with Residuals
(EXPRES) which modifies the prompt learning paradigm specifically for effective
adaptation of vision transformers (ViT). Out method constructs downstream
representations via learnable ``output'' tokens, that are akin to the learned
class tokens of the ViT. Further for better steering of the downstream
representation processed by the frozen transformer, we introduce residual
learnable tokens that are added to the output of various computations. We apply
EXPRES for image classification, few shot learning, and semantic segmentation,
and show our method is capable of achieving state of the art prompt tuning on
3/3 categories of the VTAB benchmark. In addition to strong performance, we
observe that our approach is an order of magnitude more prompt efficient than
existing visual prompting baselines. We analytically show the computational
benefits of our approach over weight space adaptation techniques like
finetuning. Lastly we systematically corroborate the architectural design of
our method via a series of ablation experiments.
|
no_new_dataset
| 0.710007
|
2303.15619
|
Bruce Xu
|
Muhammed Shahir Abdurrahman, Hashem Elezabi, Bruce Changlong Xu
|
Typhoon: Towards an Effective Task-Specific Masking Strategy for
Pre-trained Language Models
|
9 pages, 2 figures
| null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Through exploiting a high level of parallelism enabled by graphics processing
units, transformer architectures have enabled tremendous strides forward in the
field of natural language processing. In a traditional masked language model,
special MASK tokens are used to prompt our model to gather contextual
information from surrounding words to restore originally hidden information. In
this paper, we explore a task-specific masking framework for pre-trained large
language models that enables superior performance on particular downstream
tasks on the datasets in the GLUE benchmark. We develop our own masking
algorithm, Typhoon, based on token input gradients, and compare this with other
standard baselines. We find that Typhoon offers performance competitive with
whole-word masking on the MRPC dataset. Our implementation can be found in a
public Github Repository.
|
[
{
"version": "v1",
"created": "Mon, 27 Mar 2023 22:27:23 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Abdurrahman",
"Muhammed Shahir",
""
],
[
"Elezabi",
"Hashem",
""
],
[
"Xu",
"Bruce Changlong",
""
]
] |
TITLE: Typhoon: Towards an Effective Task-Specific Masking Strategy for
Pre-trained Language Models
ABSTRACT: Through exploiting a high level of parallelism enabled by graphics processing
units, transformer architectures have enabled tremendous strides forward in the
field of natural language processing. In a traditional masked language model,
special MASK tokens are used to prompt our model to gather contextual
information from surrounding words to restore originally hidden information. In
this paper, we explore a task-specific masking framework for pre-trained large
language models that enables superior performance on particular downstream
tasks on the datasets in the GLUE benchmark. We develop our own masking
algorithm, Typhoon, based on token input gradients, and compare this with other
standard baselines. We find that Typhoon offers performance competitive with
whole-word masking on the MRPC dataset. Our implementation can be found in a
public Github Repository.
|
no_new_dataset
| 0.708389
|
2303.15654
|
Canyu Zhang
|
Canyu Zhang, Zhenyao Wu, Xinyi Wu, Ziyu Zhao, Song Wang
|
Few-Shot 3D Point Cloud Semantic Segmentation via Stratified
Class-Specific Attention Based Transformer Network
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
3D point cloud semantic segmentation aims to group all points into different
semantic categories, which benefits important applications such as point cloud
scene reconstruction and understanding. Existing supervised point cloud
semantic segmentation methods usually require large-scale annotated point
clouds for training and cannot handle new categories. While a few-shot learning
method was proposed recently to address these two problems, it suffers from
high computational complexity caused by graph construction and inability to
learn fine-grained relationships among points due to the use of pooling
operations. In this paper, we further address these problems by developing a
new multi-layer transformer network for few-shot point cloud semantic
segmentation. In the proposed network, the query point cloud features are
aggregated based on the class-specific support features in different scales.
Without using pooling operations, our method makes full use of all pixel-level
features from the support samples. By better leveraging the support features
for few-shot learning, the proposed method achieves the new state-of-the-art
performance, with 15\% less inference time, over existing few-shot 3D point
cloud segmentation models on the S3DIS dataset and the ScanNet dataset.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 00:27:54 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zhang",
"Canyu",
""
],
[
"Wu",
"Zhenyao",
""
],
[
"Wu",
"Xinyi",
""
],
[
"Zhao",
"Ziyu",
""
],
[
"Wang",
"Song",
""
]
] |
TITLE: Few-Shot 3D Point Cloud Semantic Segmentation via Stratified
Class-Specific Attention Based Transformer Network
ABSTRACT: 3D point cloud semantic segmentation aims to group all points into different
semantic categories, which benefits important applications such as point cloud
scene reconstruction and understanding. Existing supervised point cloud
semantic segmentation methods usually require large-scale annotated point
clouds for training and cannot handle new categories. While a few-shot learning
method was proposed recently to address these two problems, it suffers from
high computational complexity caused by graph construction and inability to
learn fine-grained relationships among points due to the use of pooling
operations. In this paper, we further address these problems by developing a
new multi-layer transformer network for few-shot point cloud semantic
segmentation. In the proposed network, the query point cloud features are
aggregated based on the class-specific support features in different scales.
Without using pooling operations, our method makes full use of all pixel-level
features from the support samples. By better leveraging the support features
for few-shot learning, the proposed method achieves the new state-of-the-art
performance, with 15\% less inference time, over existing few-shot 3D point
cloud segmentation models on the S3DIS dataset and the ScanNet dataset.
|
no_new_dataset
| 0.713013
|
2303.15655
|
Jin Liu
|
Jin Liu and Jianye Chen and Chongfeng Fan and Fengyu Zhou
|
Joint embedding in Hierarchical distance and semantic representation
learning for link prediction
|
Submitted to Big Data research one year ago
| null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
The link prediction task aims to predict missing entities or relations in the
knowledge graph and is essential for the downstream application. Existing
well-known models deal with this task by mainly focusing on representing
knowledge graph triplets in the distance space or semantic space. However, they
can not fully capture the information of head and tail entities, nor even make
good use of hierarchical level information. Thus, in this paper, we propose a
novel knowledge graph embedding model for the link prediction task, namely,
HIE, which models each triplet (\textit{h}, \textit{r}, \textit{t}) into
distance measurement space and semantic measurement space, simultaneously.
Moreover, HIE is introduced into hierarchical-aware space to leverage rich
hierarchical information of entities and relations for better representation
learning. Specifically, we apply distance transformation operation on the head
entity in distance space to obtain the tail entity instead of translation-based
or rotation-based approaches. Experimental results of HIE on four real-world
datasets show that HIE outperforms several existing state-of-the-art knowledge
graph embedding methods on the link prediction task and deals with complex
relations accurately.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 00:42:29 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Liu",
"Jin",
""
],
[
"Chen",
"Jianye",
""
],
[
"Fan",
"Chongfeng",
""
],
[
"Zhou",
"Fengyu",
""
]
] |
TITLE: Joint embedding in Hierarchical distance and semantic representation
learning for link prediction
ABSTRACT: The link prediction task aims to predict missing entities or relations in the
knowledge graph and is essential for the downstream application. Existing
well-known models deal with this task by mainly focusing on representing
knowledge graph triplets in the distance space or semantic space. However, they
can not fully capture the information of head and tail entities, nor even make
good use of hierarchical level information. Thus, in this paper, we propose a
novel knowledge graph embedding model for the link prediction task, namely,
HIE, which models each triplet (\textit{h}, \textit{r}, \textit{t}) into
distance measurement space and semantic measurement space, simultaneously.
Moreover, HIE is introduced into hierarchical-aware space to leverage rich
hierarchical information of entities and relations for better representation
learning. Specifically, we apply distance transformation operation on the head
entity in distance space to obtain the tail entity instead of translation-based
or rotation-based approaches. Experimental results of HIE on four real-world
datasets show that HIE outperforms several existing state-of-the-art knowledge
graph embedding methods on the link prediction task and deals with complex
relations accurately.
|
no_new_dataset
| 0.710635
|
2303.15671
|
Suncheng Xiang
|
Qingzhong Chen, Shilun Cai, Crystal Cai, Zefang Yu, Dahong Qian,
Suncheng Xiang
|
Colo-SCRL: Self-Supervised Contrastive Representation Learning for
Colonoscopic Video Retrieval
|
Accepted by ICME 2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Colonoscopic video retrieval, which is a critical part of polyp treatment,
has great clinical significance for the prevention and treatment of colorectal
cancer. However, retrieval models trained on action recognition datasets
usually produce unsatisfactory retrieval results on colonoscopic datasets due
to the large domain gap between them. To seek a solution to this problem, we
construct a large-scale colonoscopic dataset named Colo-Pair for medical
practice. Based on this dataset, a simple yet effective training method called
Colo-SCRL is proposed for more robust representation learning. It aims to
refine general knowledge from colonoscopies through masked autoencoder-based
reconstruction and momentum contrast to improve retrieval performance. To the
best of our knowledge, this is the first attempt to employ the contrastive
learning paradigm for medical video retrieval. Empirical results show that our
method significantly outperforms current state-of-the-art methods in the
colonoscopic video retrieval task.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 01:27:23 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Chen",
"Qingzhong",
""
],
[
"Cai",
"Shilun",
""
],
[
"Cai",
"Crystal",
""
],
[
"Yu",
"Zefang",
""
],
[
"Qian",
"Dahong",
""
],
[
"Xiang",
"Suncheng",
""
]
] |
TITLE: Colo-SCRL: Self-Supervised Contrastive Representation Learning for
Colonoscopic Video Retrieval
ABSTRACT: Colonoscopic video retrieval, which is a critical part of polyp treatment,
has great clinical significance for the prevention and treatment of colorectal
cancer. However, retrieval models trained on action recognition datasets
usually produce unsatisfactory retrieval results on colonoscopic datasets due
to the large domain gap between them. To seek a solution to this problem, we
construct a large-scale colonoscopic dataset named Colo-Pair for medical
practice. Based on this dataset, a simple yet effective training method called
Colo-SCRL is proposed for more robust representation learning. It aims to
refine general knowledge from colonoscopies through masked autoencoder-based
reconstruction and momentum contrast to improve retrieval performance. To the
best of our knowledge, this is the first attempt to employ the contrastive
learning paradigm for medical video retrieval. Empirical results show that our
method significantly outperforms current state-of-the-art methods in the
colonoscopic video retrieval task.
|
new_dataset
| 0.716609
|
2303.15682
|
Sanxing Chen
|
Sanxing Chen, Hao Cheng, Xiaodong Liu, Jian Jiao, Yangfeng Ji and
Jianfeng Gao
|
Pre-training Transformers for Knowledge Graph Completion
| null | null | null | null |
cs.CL cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Learning transferable representation of knowledge graphs (KGs) is challenging
due to the heterogeneous, multi-relational nature of graph structures. Inspired
by Transformer-based pretrained language models' success on learning
transferable representation for texts, we introduce a novel inductive KG
representation model (iHT) for KG completion by large-scale pre-training. iHT
consists of a entity encoder (e.g., BERT) and a neighbor-aware relational
scoring function both parameterized by Transformers. We first pre-train iHT on
a large KG dataset, Wikidata5M. Our approach achieves new state-of-the-art
results on matched evaluations, with a relative improvement of more than 25% in
mean reciprocal rank over previous SOTA models. When further fine-tuned on
smaller KGs with either entity and relational shifts, pre-trained iHT
representations are shown to be transferable, significantly improving the
performance on FB15K-237 and WN18RR.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 02:10:37 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Chen",
"Sanxing",
""
],
[
"Cheng",
"Hao",
""
],
[
"Liu",
"Xiaodong",
""
],
[
"Jiao",
"Jian",
""
],
[
"Ji",
"Yangfeng",
""
],
[
"Gao",
"Jianfeng",
""
]
] |
TITLE: Pre-training Transformers for Knowledge Graph Completion
ABSTRACT: Learning transferable representation of knowledge graphs (KGs) is challenging
due to the heterogeneous, multi-relational nature of graph structures. Inspired
by Transformer-based pretrained language models' success on learning
transferable representation for texts, we introduce a novel inductive KG
representation model (iHT) for KG completion by large-scale pre-training. iHT
consists of a entity encoder (e.g., BERT) and a neighbor-aware relational
scoring function both parameterized by Transformers. We first pre-train iHT on
a large KG dataset, Wikidata5M. Our approach achieves new state-of-the-art
results on matched evaluations, with a relative improvement of more than 25% in
mean reciprocal rank over previous SOTA models. When further fine-tuned on
smaller KGs with either entity and relational shifts, pre-trained iHT
representations are shown to be transferable, significantly improving the
performance on FB15K-237 and WN18RR.
|
no_new_dataset
| 0.709069
|
2303.15686
|
Jingzhi Hu
|
Jingzhi Hu and Zhe Chen and Jun Luo
|
Multi-band Reconfigurable Holographic Surface Based ISAC Systems: Design
and Optimization
|
6 pages, 4 figures, IEEE ICC
| null | null | null |
eess.SP cs.SY eess.SY
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Metamaterial-based reconfigurable holographic surfaces (RHSs) have been
proposed as novel cost-efficient antenna arrays, which are promising for
improving the positioning and communication performance of integrated sensing
and communications (ISAC) systems. However, due to the high frequency
selectivity of the metamaterial elements, RHSs face challenges in supporting
ultra-wide bandwidth (UWB), which significantly limits the positioning
precision. In this paper, to avoid the physical limitations of UWB RHS while
enhancing the performance of RHS-based ISAC systems, we propose a multi-band
(MB) based ISAC system. We analyze its positioning precision and propose an
efficient algorithm to optimize the large number of variables in analog and
digital beamforming. Through comparison with benchmark results, simulation
results verify the efficiency of our proposed system and algorithm, and show
that the system achieves $42\%$ less positioning error, which reduces $82\%$
communication capacity loss.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 02:17:17 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Hu",
"Jingzhi",
""
],
[
"Chen",
"Zhe",
""
],
[
"Luo",
"Jun",
""
]
] |
TITLE: Multi-band Reconfigurable Holographic Surface Based ISAC Systems: Design
and Optimization
ABSTRACT: Metamaterial-based reconfigurable holographic surfaces (RHSs) have been
proposed as novel cost-efficient antenna arrays, which are promising for
improving the positioning and communication performance of integrated sensing
and communications (ISAC) systems. However, due to the high frequency
selectivity of the metamaterial elements, RHSs face challenges in supporting
ultra-wide bandwidth (UWB), which significantly limits the positioning
precision. In this paper, to avoid the physical limitations of UWB RHS while
enhancing the performance of RHS-based ISAC systems, we propose a multi-band
(MB) based ISAC system. We analyze its positioning precision and propose an
efficient algorithm to optimize the large number of variables in analog and
digital beamforming. Through comparison with benchmark results, simulation
results verify the efficiency of our proposed system and algorithm, and show
that the system achieves $42\%$ less positioning error, which reduces $82\%$
communication capacity loss.
|
no_new_dataset
| 0.711838
|
2303.15720
|
Zhiyong Cheng
|
Zhiyong Cheng and Sai Han and Fan Liu and Lei Zhu and Zan Gao and
Yuxin Peng
|
Multi-Behavior Recommendation with Cascading Graph Convolution Networks
|
Accepted by WWW 2023
| null |
10.1145/3543507.3583439
| null |
cs.IR
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Multi-behavior recommendation, which exploits auxiliary behaviors (e.g.,
click and cart) to help predict users' potential interactions on the target
behavior (e.g., buy), is regarded as an effective way to alleviate the data
sparsity or cold-start issues in recommendation. Multi-behaviors are often
taken in certain orders in real-world applications (e.g., click>cart>buy). In a
behavior chain, a latter behavior usually exhibits a stronger signal of user
preference than the former one does. Most existing multi-behavior models fail
to capture such dependencies in a behavior chain for embedding learning. In
this work, we propose a novel multi-behavior recommendation model with
cascading graph convolution networks (named MB-CGCN). In MB-CGCN, the
embeddings learned from one behavior are used as the input features for the
next behavior's embedding learning after a feature transformation operation. In
this way, our model explicitly utilizes the behavior dependencies in embedding
learning. Experiments on two benchmark datasets demonstrate the effectiveness
of our model on exploiting multi-behavior data. It outperforms the best
baseline by 33.7% and 35.9% on average over the two datasets in terms of
Recall@10 and NDCG@10, respectively.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 04:07:59 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Cheng",
"Zhiyong",
""
],
[
"Han",
"Sai",
""
],
[
"Liu",
"Fan",
""
],
[
"Zhu",
"Lei",
""
],
[
"Gao",
"Zan",
""
],
[
"Peng",
"Yuxin",
""
]
] |
TITLE: Multi-Behavior Recommendation with Cascading Graph Convolution Networks
ABSTRACT: Multi-behavior recommendation, which exploits auxiliary behaviors (e.g.,
click and cart) to help predict users' potential interactions on the target
behavior (e.g., buy), is regarded as an effective way to alleviate the data
sparsity or cold-start issues in recommendation. Multi-behaviors are often
taken in certain orders in real-world applications (e.g., click>cart>buy). In a
behavior chain, a latter behavior usually exhibits a stronger signal of user
preference than the former one does. Most existing multi-behavior models fail
to capture such dependencies in a behavior chain for embedding learning. In
this work, we propose a novel multi-behavior recommendation model with
cascading graph convolution networks (named MB-CGCN). In MB-CGCN, the
embeddings learned from one behavior are used as the input features for the
next behavior's embedding learning after a feature transformation operation. In
this way, our model explicitly utilizes the behavior dependencies in embedding
learning. Experiments on two benchmark datasets demonstrate the effectiveness
of our model on exploiting multi-behavior data. It outperforms the best
baseline by 33.7% and 35.9% on average over the two datasets in terms of
Recall@10 and NDCG@10, respectively.
|
no_new_dataset
| 0.711606
|
2303.15724
|
Satoshi Ikehata Mr.
|
Satoshi Ikehata
|
Scalable, Detailed and Mask-Free Universal Photometric Stereo
|
CVPR 2023 (Highlight). The source code will be available at
https://github.com/satoshi-ikehata/SDM-UniPS-CVPR2023
| null | null | null |
cs.CV cs.GR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we introduce SDM-UniPS, a groundbreaking Scalable, Detailed,
Mask-free, and Universal Photometric Stereo network. Our approach can recover
astonishingly intricate surface normal maps, rivaling the quality of 3D
scanners, even when images are captured under unknown, spatially-varying
lighting conditions in uncontrolled environments. We have extended previous
universal photometric stereo networks to extract spatial-light features,
utilizing all available information in high-resolution input images and
accounting for non-local interactions among surface points. Moreover, we
present a new synthetic training dataset that encompasses a diverse range of
shapes, materials, and illumination scenarios found in real-world scenes.
Through extensive evaluation, we demonstrate that our method not only surpasses
calibrated, lighting-specific techniques on public benchmarks, but also excels
with a significantly smaller number of input images even without object masks.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 04:18:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Ikehata",
"Satoshi",
""
]
] |
TITLE: Scalable, Detailed and Mask-Free Universal Photometric Stereo
ABSTRACT: In this paper, we introduce SDM-UniPS, a groundbreaking Scalable, Detailed,
Mask-free, and Universal Photometric Stereo network. Our approach can recover
astonishingly intricate surface normal maps, rivaling the quality of 3D
scanners, even when images are captured under unknown, spatially-varying
lighting conditions in uncontrolled environments. We have extended previous
universal photometric stereo networks to extract spatial-light features,
utilizing all available information in high-resolution input images and
accounting for non-local interactions among surface points. Moreover, we
present a new synthetic training dataset that encompasses a diverse range of
shapes, materials, and illumination scenarios found in real-world scenes.
Through extensive evaluation, we demonstrate that our method not only surpasses
calibrated, lighting-specific techniques on public benchmarks, but also excels
with a significantly smaller number of input images even without object masks.
|
new_dataset
| 0.712807
|
2303.15727
|
Bishal Lamichhane
|
Bishal Lamichhane
|
Evaluation of ChatGPT for NLP-based Mental Health Applications
| null | null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Large language models (LLM) have been successful in several natural language
understanding tasks and could be relevant for natural language processing
(NLP)-based mental health application research. In this work, we report the
performance of LLM-based ChatGPT (with gpt-3.5-turbo backend) in three
text-based mental health classification tasks: stress detection (2-class
classification), depression detection (2-class classification), and suicidality
detection (5-class classification). We obtained annotated social media posts
for the three classification tasks from public datasets. Then ChatGPT API
classified the social media posts with an input prompt for classification. We
obtained F1 scores of 0.73, 0.86, and 0.37 for stress detection, depression
detection, and suicidality detection, respectively. A baseline model that
always predicted the dominant class resulted in F1 scores of 0.35, 0.60, and
0.19. The zero-shot classification accuracy obtained with ChatGPT indicates a
potential use of language models for mental health classification tasks.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 04:47:43 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Lamichhane",
"Bishal",
""
]
] |
TITLE: Evaluation of ChatGPT for NLP-based Mental Health Applications
ABSTRACT: Large language models (LLM) have been successful in several natural language
understanding tasks and could be relevant for natural language processing
(NLP)-based mental health application research. In this work, we report the
performance of LLM-based ChatGPT (with gpt-3.5-turbo backend) in three
text-based mental health classification tasks: stress detection (2-class
classification), depression detection (2-class classification), and suicidality
detection (5-class classification). We obtained annotated social media posts
for the three classification tasks from public datasets. Then ChatGPT API
classified the social media posts with an input prompt for classification. We
obtained F1 scores of 0.73, 0.86, and 0.37 for stress detection, depression
detection, and suicidality detection, respectively. A baseline model that
always predicted the dominant class resulted in F1 scores of 0.35, 0.60, and
0.19. The zero-shot classification accuracy obtained with ChatGPT indicates a
potential use of language models for mental health classification tasks.
|
no_new_dataset
| 0.71022
|
2303.15743
|
Linfang Zheng
|
Linfang Zheng, Chen Wang, Yinghan Sun, Esha Dasgupta, Hua Chen, Ales
Leonardis, Wei Zhang, Hyung Jin Chang
|
HS-Pose: Hybrid Scope Feature Extraction for Category-level Object Pose
Estimation
|
Accepted by the 2023 IEEE/CVF Computer Vision and Pattern Recognition
Conference (CVPR)
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
In this paper, we focus on the problem of category-level object pose
estimation, which is challenging due to the large intra-category shape
variation. 3D graph convolution (3D-GC) based methods have been widely used to
extract local geometric features, but they have limitations for complex shaped
objects and are sensitive to noise. Moreover, the scale and translation
invariant properties of 3D-GC restrict the perception of an object's size and
translation information. In this paper, we propose a simple network structure,
the HS-layer, which extends 3D-GC to extract hybrid scope latent features from
point cloud data for category-level object pose estimation tasks. The proposed
HS-layer: 1) is able to perceive local-global geometric structure and global
information, 2) is robust to noise, and 3) can encode size and translation
information. Our experiments show that the simple replacement of the 3D-GC
layer with the proposed HS-layer on the baseline method (GPV-Pose) achieves a
significant improvement, with the performance increased by 14.5% on 5d2cm
metric and 10.3% on IoU75. Our method outperforms the state-of-the-art methods
by a large margin (8.3% on 5d2cm, 6.9% on IoU75) on the REAL275 dataset and
runs in real-time (50 FPS).
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 05:36:42 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zheng",
"Linfang",
""
],
[
"Wang",
"Chen",
""
],
[
"Sun",
"Yinghan",
""
],
[
"Dasgupta",
"Esha",
""
],
[
"Chen",
"Hua",
""
],
[
"Leonardis",
"Ales",
""
],
[
"Zhang",
"Wei",
""
],
[
"Chang",
"Hyung Jin",
""
]
] |
TITLE: HS-Pose: Hybrid Scope Feature Extraction for Category-level Object Pose
Estimation
ABSTRACT: In this paper, we focus on the problem of category-level object pose
estimation, which is challenging due to the large intra-category shape
variation. 3D graph convolution (3D-GC) based methods have been widely used to
extract local geometric features, but they have limitations for complex shaped
objects and are sensitive to noise. Moreover, the scale and translation
invariant properties of 3D-GC restrict the perception of an object's size and
translation information. In this paper, we propose a simple network structure,
the HS-layer, which extends 3D-GC to extract hybrid scope latent features from
point cloud data for category-level object pose estimation tasks. The proposed
HS-layer: 1) is able to perceive local-global geometric structure and global
information, 2) is robust to noise, and 3) can encode size and translation
information. Our experiments show that the simple replacement of the 3D-GC
layer with the proposed HS-layer on the baseline method (GPV-Pose) achieves a
significant improvement, with the performance increased by 14.5% on 5d2cm
metric and 10.3% on IoU75. Our method outperforms the state-of-the-art methods
by a large margin (8.3% on 5d2cm, 6.9% on IoU75) on the REAL275 dataset and
runs in real-time (50 FPS).
|
no_new_dataset
| 0.711638
|
2303.15745
|
Arunav Gupta
|
Arunav Gupta, Rohit Mishra, William Luu, Mehdi Bouassami
|
On Feature Scaling of Recursive Feature Machines
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
In this technical report, we explore the behavior of Recursive Feature
Machines (RFMs), a type of novel kernel machine that recursively learns
features via the average gradient outer product, through a series of
experiments on regression datasets. When successively adding random noise
features to a dataset, we observe intriguing patterns in the Mean Squared Error
(MSE) curves with the test MSE exhibiting a decrease-increase-decrease pattern.
This behavior is consistent across different dataset sizes, noise parameters,
and target functions. Interestingly, the observed MSE curves show similarities
to the "double descent" phenomenon observed in deep neural networks, hinting at
new connection between RFMs and neural network behavior. This report lays the
groundwork for future research into this peculiar behavior.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 06:00:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Gupta",
"Arunav",
""
],
[
"Mishra",
"Rohit",
""
],
[
"Luu",
"William",
""
],
[
"Bouassami",
"Mehdi",
""
]
] |
TITLE: On Feature Scaling of Recursive Feature Machines
ABSTRACT: In this technical report, we explore the behavior of Recursive Feature
Machines (RFMs), a type of novel kernel machine that recursively learns
features via the average gradient outer product, through a series of
experiments on regression datasets. When successively adding random noise
features to a dataset, we observe intriguing patterns in the Mean Squared Error
(MSE) curves with the test MSE exhibiting a decrease-increase-decrease pattern.
This behavior is consistent across different dataset sizes, noise parameters,
and target functions. Interestingly, the observed MSE curves show similarities
to the "double descent" phenomenon observed in deep neural networks, hinting at
new connection between RFMs and neural network behavior. This report lays the
groundwork for future research into this peculiar behavior.
|
no_new_dataset
| 0.712238
|
2303.15763
|
Hao Xu
|
Hao Xu, Shuang Song, Ze Mao
|
Characterizing the Performance of Emerging Deep Learning, Graph, and
High Performance Computing Workloads Under Interference
| null | null | null | null |
cs.PF
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Throughput-oriented computing via co-running multiple applications in the
same machine has been widely adopted to achieve high hardware utilization and
energy saving on modern supercomputers and data centers. However, efficiently
co-running applications raises new design challenges, mainly because
applications with diverse requirements can stress out shared hardware resources
(IO, Network and Cache) at various levels. The disparities in resource usage
can result in interference, which in turn can lead to unpredictable co-running
behaviors. To better understand application interference, prior work provided
detailed execution characterization. However, these characterization approaches
either emphasize on traditional benchmarks or fall into a single application
domain. To address this issue, we study 25 up-to-date applications and
benchmarks from various application domains and form 625 consolidation pairs to
thoroughly analyze the execution interference caused by application co-running.
Moreover, we leverage mini-benchmarks and real applications to pinpoint the
provenance of co-running interference in both hardware and software aspects.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 06:44:32 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Xu",
"Hao",
""
],
[
"Song",
"Shuang",
""
],
[
"Mao",
"Ze",
""
]
] |
TITLE: Characterizing the Performance of Emerging Deep Learning, Graph, and
High Performance Computing Workloads Under Interference
ABSTRACT: Throughput-oriented computing via co-running multiple applications in the
same machine has been widely adopted to achieve high hardware utilization and
energy saving on modern supercomputers and data centers. However, efficiently
co-running applications raises new design challenges, mainly because
applications with diverse requirements can stress out shared hardware resources
(IO, Network and Cache) at various levels. The disparities in resource usage
can result in interference, which in turn can lead to unpredictable co-running
behaviors. To better understand application interference, prior work provided
detailed execution characterization. However, these characterization approaches
either emphasize on traditional benchmarks or fall into a single application
domain. To address this issue, we study 25 up-to-date applications and
benchmarks from various application domains and form 625 consolidation pairs to
thoroughly analyze the execution interference caused by application co-running.
Moreover, we leverage mini-benchmarks and real applications to pinpoint the
provenance of co-running interference in both hardware and software aspects.
|
no_new_dataset
| 0.706368
|
2303.15768
|
Jaeseong Lee
|
Jaeseong Lee, Taewoo Kim, Sunghyun Park, Younggun Lee, Jaegul Choo
|
RobustSwap: A Simple yet Robust Face Swapping Model against Attribute
Leakage
|
21 pages
| null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Face swapping aims at injecting a source image's identity (i.e., facial
features) into a target image, while strictly preserving the target's
attributes, which are irrelevant to identity. However, we observed that
previous approaches still suffer from source attribute leakage, where the
source image's attributes interfere with the target image's. In this paper, we
analyze the latent space of StyleGAN and find the adequate combination of the
latents geared for face swapping task. Based on the findings, we develop a
simple yet robust face swapping model, RobustSwap, which is resistant to the
potential source attribute leakage. Moreover, we exploit the coordination of
3DMM's implicit and explicit information as a guidance to incorporate the
structure of the source image and the precise pose of the target image. Despite
our method solely utilizing an image dataset without identity labels for
training, our model has the capability to generate high-fidelity and temporally
consistent videos. Through extensive qualitative and quantitative evaluations,
we demonstrate that our method shows significant improvements compared with the
previous face swapping models in synthesizing both images and videos. Project
page is available at https://robustswap.github.io/
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 07:03:31 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Lee",
"Jaeseong",
""
],
[
"Kim",
"Taewoo",
""
],
[
"Park",
"Sunghyun",
""
],
[
"Lee",
"Younggun",
""
],
[
"Choo",
"Jaegul",
""
]
] |
TITLE: RobustSwap: A Simple yet Robust Face Swapping Model against Attribute
Leakage
ABSTRACT: Face swapping aims at injecting a source image's identity (i.e., facial
features) into a target image, while strictly preserving the target's
attributes, which are irrelevant to identity. However, we observed that
previous approaches still suffer from source attribute leakage, where the
source image's attributes interfere with the target image's. In this paper, we
analyze the latent space of StyleGAN and find the adequate combination of the
latents geared for face swapping task. Based on the findings, we develop a
simple yet robust face swapping model, RobustSwap, which is resistant to the
potential source attribute leakage. Moreover, we exploit the coordination of
3DMM's implicit and explicit information as a guidance to incorporate the
structure of the source image and the precise pose of the target image. Despite
our method solely utilizing an image dataset without identity labels for
training, our model has the capability to generate high-fidelity and temporally
consistent videos. Through extensive qualitative and quantitative evaluations,
we demonstrate that our method shows significant improvements compared with the
previous face swapping models in synthesizing both images and videos. Project
page is available at https://robustswap.github.io/
|
no_new_dataset
| 0.709617
|
2303.15796
|
Xiangyang Li
|
Xiangyang Li, Zihan Wang, Jiahao Yang, Yaowei Wang, Shuqiang Jiang
|
KERM: Knowledge Enhanced Reasoning for Vision-and-Language Navigation
|
Accepted by CVPR 2023. The code is available at
https://github.com/XiangyangLi20/KERM
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Vision-and-language navigation (VLN) is the task to enable an embodied agent
to navigate to a remote location following the natural language instruction in
real scenes. Most of the previous approaches utilize the entire features or
object-centric features to represent navigable candidates. However, these
representations are not efficient enough for an agent to perform actions to
arrive the target location. As knowledge provides crucial information which is
complementary to visible content, in this paper, we propose a Knowledge
Enhanced Reasoning Model (KERM) to leverage knowledge to improve agent
navigation ability. Specifically, we first retrieve facts (i.e., knowledge
described by language descriptions) for the navigation views based on local
regions from the constructed knowledge base. The retrieved facts range from
properties of a single object (e.g., color, shape) to relationships between
objects (e.g., action, spatial position), providing crucial information for
VLN. We further present the KERM which contains the purification, fact-aware
interaction, and instruction-guided aggregation modules to integrate visual,
history, instruction, and fact features. The proposed KERM can automatically
select and gather crucial and relevant cues, obtaining more accurate action
prediction. Experimental results on the REVERIE, R2R, and SOON datasets
demonstrate the effectiveness of the proposed method.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 08:00:46 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Li",
"Xiangyang",
""
],
[
"Wang",
"Zihan",
""
],
[
"Yang",
"Jiahao",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Jiang",
"Shuqiang",
""
]
] |
TITLE: KERM: Knowledge Enhanced Reasoning for Vision-and-Language Navigation
ABSTRACT: Vision-and-language navigation (VLN) is the task to enable an embodied agent
to navigate to a remote location following the natural language instruction in
real scenes. Most of the previous approaches utilize the entire features or
object-centric features to represent navigable candidates. However, these
representations are not efficient enough for an agent to perform actions to
arrive the target location. As knowledge provides crucial information which is
complementary to visible content, in this paper, we propose a Knowledge
Enhanced Reasoning Model (KERM) to leverage knowledge to improve agent
navigation ability. Specifically, we first retrieve facts (i.e., knowledge
described by language descriptions) for the navigation views based on local
regions from the constructed knowledge base. The retrieved facts range from
properties of a single object (e.g., color, shape) to relationships between
objects (e.g., action, spatial position), providing crucial information for
VLN. We further present the KERM which contains the purification, fact-aware
interaction, and instruction-guided aggregation modules to integrate visual,
history, instruction, and fact features. The proposed KERM can automatically
select and gather crucial and relevant cues, obtaining more accurate action
prediction. Experimental results on the REVERIE, R2R, and SOON datasets
demonstrate the effectiveness of the proposed method.
|
no_new_dataset
| 0.710635
|
2303.15810
|
Haoran Xu
|
Haoran Xu, Li Jiang, Jianxiong Li, Zhuoran Yang, Zhaoran Wang, Victor
Wai Kin Chan, Xianyuan Zhan
|
Offline RL with No OOD Actions: In-Sample Learning via Implicit Value
Regularization
|
ICLR 2023 notable top 5%
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Most offline reinforcement learning (RL) methods suffer from the trade-off
between improving the policy to surpass the behavior policy and constraining
the policy to limit the deviation from the behavior policy as computing
$Q$-values using out-of-distribution (OOD) actions will suffer from errors due
to distributional shift. The recently proposed \textit{In-sample Learning}
paradigm (i.e., IQL), which improves the policy by quantile regression using
only data samples, shows great promise because it learns an optimal policy
without querying the value function of any unseen actions. However, it remains
unclear how this type of method handles the distributional shift in learning
the value function. In this work, we make a key finding that the in-sample
learning paradigm arises under the \textit{Implicit Value Regularization} (IVR)
framework. This gives a deeper understanding of why the in-sample learning
paradigm works, i.e., it applies implicit value regularization to the policy.
Based on the IVR framework, we further propose two practical algorithms, Sparse
$Q$-learning (SQL) and Exponential $Q$-learning (EQL), which adopt the same
value regularization used in existing works, but in a complete in-sample
manner. Compared with IQL, we find that our algorithms introduce sparsity in
learning the value function, making them more robust in noisy data regimes. We
also verify the effectiveness of SQL and EQL on D4RL benchmark datasets and
show the benefits of in-sample learning by comparing them with CQL in small
data regimes.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 08:30:01 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Xu",
"Haoran",
""
],
[
"Jiang",
"Li",
""
],
[
"Li",
"Jianxiong",
""
],
[
"Yang",
"Zhuoran",
""
],
[
"Wang",
"Zhaoran",
""
],
[
"Chan",
"Victor Wai Kin",
""
],
[
"Zhan",
"Xianyuan",
""
]
] |
TITLE: Offline RL with No OOD Actions: In-Sample Learning via Implicit Value
Regularization
ABSTRACT: Most offline reinforcement learning (RL) methods suffer from the trade-off
between improving the policy to surpass the behavior policy and constraining
the policy to limit the deviation from the behavior policy as computing
$Q$-values using out-of-distribution (OOD) actions will suffer from errors due
to distributional shift. The recently proposed \textit{In-sample Learning}
paradigm (i.e., IQL), which improves the policy by quantile regression using
only data samples, shows great promise because it learns an optimal policy
without querying the value function of any unseen actions. However, it remains
unclear how this type of method handles the distributional shift in learning
the value function. In this work, we make a key finding that the in-sample
learning paradigm arises under the \textit{Implicit Value Regularization} (IVR)
framework. This gives a deeper understanding of why the in-sample learning
paradigm works, i.e., it applies implicit value regularization to the policy.
Based on the IVR framework, we further propose two practical algorithms, Sparse
$Q$-learning (SQL) and Exponential $Q$-learning (EQL), which adopt the same
value regularization used in existing works, but in a complete in-sample
manner. Compared with IQL, we find that our algorithms introduce sparsity in
learning the value function, making them more robust in noisy data regimes. We
also verify the effectiveness of SQL and EQL on D4RL benchmark datasets and
show the benefits of in-sample learning by comparing them with CQL in small
data regimes.
|
no_new_dataset
| 0.708843
|
2303.15822
|
Deze Wang
|
Deze Wang, Boxing Chen, Shanshan Li, Wei Luo, Shaoliang Peng, Wei
Dong, Xiangke Liao
|
One Adapter for All Programming Languages? Adapter Tuning for Code
Search and Summarization
|
Accepted to the 45th International Conference on Software Engineering
(ICSE 2023)
| null | null | null |
cs.SE cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
As pre-trained models automate many code intelligence tasks, a widely used
paradigm is to fine-tune a model on the task dataset for each programming
language. A recent study reported that multilingual fine-tuning benefits a
range of tasks and models. However, we find that multilingual fine-tuning leads
to performance degradation on recent models UniXcoder and CodeT5.
To alleviate the potentially catastrophic forgetting issue in multilingual
models, we fix all pre-trained model parameters, insert the parameter-efficient
structure adapter, and fine-tune it. Updating only 0.6\% of the overall
parameters compared to full-model fine-tuning for each programming language,
adapter tuning yields consistent improvements on code search and summarization
tasks, achieving state-of-the-art results. In addition, we experimentally show
its effectiveness in cross-lingual and low-resource scenarios. Multilingual
fine-tuning with 200 samples per programming language approaches the results
fine-tuned with the entire dataset on code summarization. Our experiments on
three probing tasks show that adapter tuning significantly outperforms
full-model fine-tuning and effectively overcomes catastrophic forgetting.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 08:49:54 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Deze",
""
],
[
"Chen",
"Boxing",
""
],
[
"Li",
"Shanshan",
""
],
[
"Luo",
"Wei",
""
],
[
"Peng",
"Shaoliang",
""
],
[
"Dong",
"Wei",
""
],
[
"Liao",
"Xiangke",
""
]
] |
TITLE: One Adapter for All Programming Languages? Adapter Tuning for Code
Search and Summarization
ABSTRACT: As pre-trained models automate many code intelligence tasks, a widely used
paradigm is to fine-tune a model on the task dataset for each programming
language. A recent study reported that multilingual fine-tuning benefits a
range of tasks and models. However, we find that multilingual fine-tuning leads
to performance degradation on recent models UniXcoder and CodeT5.
To alleviate the potentially catastrophic forgetting issue in multilingual
models, we fix all pre-trained model parameters, insert the parameter-efficient
structure adapter, and fine-tune it. Updating only 0.6\% of the overall
parameters compared to full-model fine-tuning for each programming language,
adapter tuning yields consistent improvements on code search and summarization
tasks, achieving state-of-the-art results. In addition, we experimentally show
its effectiveness in cross-lingual and low-resource scenarios. Multilingual
fine-tuning with 200 samples per programming language approaches the results
fine-tuned with the entire dataset on code summarization. Our experiments on
three probing tasks show that adapter tuning significantly outperforms
full-model fine-tuning and effectively overcomes catastrophic forgetting.
|
no_new_dataset
| 0.709221
|
2303.15842
|
Wei Tian
|
Zukai Li, Wei Tian, Jingjin Wu
|
Modeling and Joint Optimization of Security, Latency, and Computational
Cost in Blockchain-based Healthcare Systems
| null | null | null | null |
cs.NE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In the era of the Internet of Things (IoT), blockchain is a promising
technology for improving the efficiency of healthcare systems, as it enables
secure storage, management, and sharing of real-time health data collected by
the IoT devices. As the implementations of blockchain-based healthcare systems
usually involve multiple conflicting metrics, it is essential to balance them
according to the requirements of specific scenarios. In this paper, we
formulate a joint optimization model with three metrics, namely latency,
security, and computational cost, that are particularly important for
IoT-enabled healthcare. However, it is computationally intractable to identify
the exact optimal solution of this problem for practical sized systems. Thus,
we propose an algorithm called the Adaptive Discrete Particle Swarm Algorithm
(ADPSA) to obtain near-optimal solutions in a low-complexity manner. With its
roots in the classical Particle Swarm Optimization (PSO) algorithm, our
proposed ADPSA can effectively manage the numerous binary and integer variables
in the formulation. We demonstrate by extensive numerical experiments that the
ADPSA consistently outperforms existing benchmark approaches, including the
original PSO, exhaustive search and Simulated Annealing, in a wide range of
scenarios.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 09:29:29 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Li",
"Zukai",
""
],
[
"Tian",
"Wei",
""
],
[
"Wu",
"Jingjin",
""
]
] |
TITLE: Modeling and Joint Optimization of Security, Latency, and Computational
Cost in Blockchain-based Healthcare Systems
ABSTRACT: In the era of the Internet of Things (IoT), blockchain is a promising
technology for improving the efficiency of healthcare systems, as it enables
secure storage, management, and sharing of real-time health data collected by
the IoT devices. As the implementations of blockchain-based healthcare systems
usually involve multiple conflicting metrics, it is essential to balance them
according to the requirements of specific scenarios. In this paper, we
formulate a joint optimization model with three metrics, namely latency,
security, and computational cost, that are particularly important for
IoT-enabled healthcare. However, it is computationally intractable to identify
the exact optimal solution of this problem for practical sized systems. Thus,
we propose an algorithm called the Adaptive Discrete Particle Swarm Algorithm
(ADPSA) to obtain near-optimal solutions in a low-complexity manner. With its
roots in the classical Particle Swarm Optimization (PSO) algorithm, our
proposed ADPSA can effectively manage the numerous binary and integer variables
in the formulation. We demonstrate by extensive numerical experiments that the
ADPSA consistently outperforms existing benchmark approaches, including the
original PSO, exhaustive search and Simulated Annealing, in a wide range of
scenarios.
|
no_new_dataset
| 0.709221
|
2303.15848
|
Zhuoran Zheng
|
Zhuoran Zheng and Xiuyi Jia
|
4K-HAZE: A Dehazing Benchmark with 4K Resolution Hazy and Haze-Free
Images
| null | null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Currently, mobile and IoT devices are in dire need of a series of methods to
enhance 4K images with limited resource expenditure. The absence of large-scale
4K benchmark datasets hampers progress in this area, especially for dehazing.
The challenges in building ultra-high-definition (UHD) dehazing datasets are
the absence of estimation methods for UHD depth maps, high-quality 4K depth
estimation datasets, and migration strategies for UHD haze images from
synthetic to real domains. To address these problems, we develop a novel
synthetic method to simulate 4K hazy images (including nighttime and daytime
scenes) from clear images, which first estimates the scene depth, simulates the
light rays and object reflectance, then migrates the synthetic images to real
domains by using a GAN, and finally yields the hazy effects on 4K resolution
images. We wrap these synthesized images into a benchmark called the 4K-HAZE
dataset. Specifically, we design the CS-Mixer (an MLP-based model that
integrates \textbf{C}hannel domain and \textbf{S}patial domain) to estimate the
depth map of 4K clear images, the GU-Net to migrate a 4K synthetic image to the
real hazy domain. The most appealing aspect of our approach (depth estimation
and domain migration) is the capability to run a 4K image on a single GPU with
24G RAM in real-time (33fps). Additionally, this work presents an objective
assessment of several state-of-the-art single-image dehazing methods that are
evaluated using the 4K-HAZE dataset. At the end of the paper, we discuss the
limitations of the 4K-HAZE dataset and its social implications.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 09:39:29 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zheng",
"Zhuoran",
""
],
[
"Jia",
"Xiuyi",
""
]
] |
TITLE: 4K-HAZE: A Dehazing Benchmark with 4K Resolution Hazy and Haze-Free
Images
ABSTRACT: Currently, mobile and IoT devices are in dire need of a series of methods to
enhance 4K images with limited resource expenditure. The absence of large-scale
4K benchmark datasets hampers progress in this area, especially for dehazing.
The challenges in building ultra-high-definition (UHD) dehazing datasets are
the absence of estimation methods for UHD depth maps, high-quality 4K depth
estimation datasets, and migration strategies for UHD haze images from
synthetic to real domains. To address these problems, we develop a novel
synthetic method to simulate 4K hazy images (including nighttime and daytime
scenes) from clear images, which first estimates the scene depth, simulates the
light rays and object reflectance, then migrates the synthetic images to real
domains by using a GAN, and finally yields the hazy effects on 4K resolution
images. We wrap these synthesized images into a benchmark called the 4K-HAZE
dataset. Specifically, we design the CS-Mixer (an MLP-based model that
integrates \textbf{C}hannel domain and \textbf{S}patial domain) to estimate the
depth map of 4K clear images, the GU-Net to migrate a 4K synthetic image to the
real hazy domain. The most appealing aspect of our approach (depth estimation
and domain migration) is the capability to run a 4K image on a single GPU with
24G RAM in real-time (33fps). Additionally, this work presents an objective
assessment of several state-of-the-art single-image dehazing methods that are
evaluated using the 4K-HAZE dataset. At the end of the paper, we discuss the
limitations of the 4K-HAZE dataset and its social implications.
|
no_new_dataset
| 0.702964
|
2303.15850
|
Kilian Zepf
|
Kilian Zepf, Eike Petersen, Jes Frellsen, Aasa Feragen
|
That Label's Got Style: Handling Label Style Bias for Uncertain Image
Segmentation
| null | null | null | null |
cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Segmentation uncertainty models predict a distribution over plausible
segmentations for a given input, which they learn from the annotator variation
in the training set. However, in practice these annotations can differ
systematically in the way they are generated, for example through the use of
different labeling tools. This results in datasets that contain both data
variability and differing label styles. In this paper, we demonstrate that
applying state-of-the-art segmentation uncertainty models on such datasets can
lead to model bias caused by the different label styles. We present an updated
modelling objective conditioning on labeling style for aleatoric uncertainty
estimation, and modify two state-of-the-art-architectures for segmentation
uncertainty accordingly. We show with extensive experiments that this method
reduces label style bias, while improving segmentation performance, increasing
the applicability of segmentation uncertainty models in the wild. We curate two
datasets, with annotations in different label styles, which we will make
publicly available along with our code upon publication.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 09:43:16 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Zepf",
"Kilian",
""
],
[
"Petersen",
"Eike",
""
],
[
"Frellsen",
"Jes",
""
],
[
"Feragen",
"Aasa",
""
]
] |
TITLE: That Label's Got Style: Handling Label Style Bias for Uncertain Image
Segmentation
ABSTRACT: Segmentation uncertainty models predict a distribution over plausible
segmentations for a given input, which they learn from the annotator variation
in the training set. However, in practice these annotations can differ
systematically in the way they are generated, for example through the use of
different labeling tools. This results in datasets that contain both data
variability and differing label styles. In this paper, we demonstrate that
applying state-of-the-art segmentation uncertainty models on such datasets can
lead to model bias caused by the different label styles. We present an updated
modelling objective conditioning on labeling style for aleatoric uncertainty
estimation, and modify two state-of-the-art-architectures for segmentation
uncertainty accordingly. We show with extensive experiments that this method
reduces label style bias, while improving segmentation performance, increasing
the applicability of segmentation uncertainty models in the wild. We curate two
datasets, with annotations in different label styles, which we will make
publicly available along with our code upon publication.
|
new_dataset
| 0.7152
|
2303.15859
|
Cheng Wang
|
Cheng Wang, Guoli Wang, Qian Zhang, Peng Guo, Wenyu Liu, Xinggang Wang
|
OpenInst: A Simple Query-Based Method for Open-World Instance
Segmentation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Open-world instance segmentation has recently gained significant
popularitydue to its importance in many real-world applications, such as
autonomous driving, robot perception, and remote sensing. However, previous
methods have either produced unsatisfactory results or relied on complex
systems and paradigms. We wonder if there is a simple way to obtain
state-of-the-art results. Fortunately, we have identified two observations that
help us achieve the best of both worlds: 1) query-based methods demonstrate
superiority over dense proposal-based methods in open-world instance
segmentation, and 2) learning localization cues is sufficient for open world
instance segmentation. Based on these observations, we propose a simple
query-based method named OpenInst for open world instance segmentation.
OpenInst leverages advanced query-based methods like QueryInst and focuses on
learning localization cues. Notably, OpenInst is an extremely simple and
straightforward framework without any auxiliary modules or post-processing, yet
achieves state-of-the-art results on multiple benchmarks. Specifically, in the
COCO$\to$UVO scenario, OpenInst achieves a mask AR of 53.3, outperforming the
previous best methods by 2.0 AR with a simpler structure. We hope that OpenInst
can serve as a solid baselines for future research in this area.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 10:03:32 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Cheng",
""
],
[
"Wang",
"Guoli",
""
],
[
"Zhang",
"Qian",
""
],
[
"Guo",
"Peng",
""
],
[
"Liu",
"Wenyu",
""
],
[
"Wang",
"Xinggang",
""
]
] |
TITLE: OpenInst: A Simple Query-Based Method for Open-World Instance
Segmentation
ABSTRACT: Open-world instance segmentation has recently gained significant
popularitydue to its importance in many real-world applications, such as
autonomous driving, robot perception, and remote sensing. However, previous
methods have either produced unsatisfactory results or relied on complex
systems and paradigms. We wonder if there is a simple way to obtain
state-of-the-art results. Fortunately, we have identified two observations that
help us achieve the best of both worlds: 1) query-based methods demonstrate
superiority over dense proposal-based methods in open-world instance
segmentation, and 2) learning localization cues is sufficient for open world
instance segmentation. Based on these observations, we propose a simple
query-based method named OpenInst for open world instance segmentation.
OpenInst leverages advanced query-based methods like QueryInst and focuses on
learning localization cues. Notably, OpenInst is an extremely simple and
straightforward framework without any auxiliary modules or post-processing, yet
achieves state-of-the-art results on multiple benchmarks. Specifically, in the
COCO$\to$UVO scenario, OpenInst achieves a mask AR of 53.3, outperforming the
previous best methods by 2.0 AR with a simpler structure. We hope that OpenInst
can serve as a solid baselines for future research in this area.
|
no_new_dataset
| 0.709025
|
2303.15879
|
Tao Wu
|
Tao Wu and Mengqi Cao and Ziteng Gao and Gangshan Wu and Limin Wang
|
STMixer: A One-Stage Sparse Action Detector
|
Accepted by CVPR 2023
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Traditional video action detectors typically adopt the two-stage pipeline,
where a person detector is first employed to generate actor boxes and then 3D
RoIAlign is used to extract actor-specific features for classification. This
detection paradigm requires multi-stage training and inference, and cannot
capture context information outside the bounding box. Recently, a few
query-based action detectors are proposed to predict action instances in an
end-to-end manner. However, they still lack adaptability in feature sampling
and decoding, thus suffering from the issues of inferior performance or slower
convergence. In this paper, we propose a new one-stage sparse action detector,
termed STMixer. STMixer is based on two core designs. First, we present a
query-based adaptive feature sampling module, which endows our STMixer with the
flexibility of mining a set of discriminative features from the entire
spatiotemporal domain. Second, we devise a dual-branch feature mixing module,
which allows our STMixer to dynamically attend to and mix video features along
the spatial and the temporal dimension respectively for better feature
decoding. Coupling these two designs with a video backbone yields an efficient
end-to-end action detector. Without bells and whistles, our STMixer obtains the
state-of-the-art results on the datasets of AVA, UCF101-24, and JHMDB.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 10:47:06 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wu",
"Tao",
""
],
[
"Cao",
"Mengqi",
""
],
[
"Gao",
"Ziteng",
""
],
[
"Wu",
"Gangshan",
""
],
[
"Wang",
"Limin",
""
]
] |
TITLE: STMixer: A One-Stage Sparse Action Detector
ABSTRACT: Traditional video action detectors typically adopt the two-stage pipeline,
where a person detector is first employed to generate actor boxes and then 3D
RoIAlign is used to extract actor-specific features for classification. This
detection paradigm requires multi-stage training and inference, and cannot
capture context information outside the bounding box. Recently, a few
query-based action detectors are proposed to predict action instances in an
end-to-end manner. However, they still lack adaptability in feature sampling
and decoding, thus suffering from the issues of inferior performance or slower
convergence. In this paper, we propose a new one-stage sparse action detector,
termed STMixer. STMixer is based on two core designs. First, we present a
query-based adaptive feature sampling module, which endows our STMixer with the
flexibility of mining a set of discriminative features from the entire
spatiotemporal domain. Second, we devise a dual-branch feature mixing module,
which allows our STMixer to dynamically attend to and mix video features along
the spatial and the temporal dimension respectively for better feature
decoding. Coupling these two designs with a video backbone yields an efficient
end-to-end action detector. Without bells and whistles, our STMixer obtains the
state-of-the-art results on the datasets of AVA, UCF101-24, and JHMDB.
|
no_new_dataset
| 0.70638
|
2303.15880
|
Guillaume Rochette
|
Guillaume Rochette, Chris Russell, Richard Bowden
|
Novel View Synthesis of Humans using Differentiable Rendering
|
Accepted at IEEE transactions on Biometrics, Behavior, and Identity
Science, 10 pages, 11 figures. arXiv admin note: substantial text overlap
with arXiv:2111.12731
| null |
10.1109/TBIOM.2022.3218903
| null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present a new approach for synthesizing novel views of people in new
poses. Our novel differentiable renderer enables the synthesis of highly
realistic images from any viewpoint. Rather than operating over mesh-based
structures, our renderer makes use of diffuse Gaussian primitives that directly
represent the underlying skeletal structure of a human. Rendering these
primitives gives results in a high-dimensional latent image, which is then
transformed into an RGB image by a decoder network. The formulation gives rise
to a fully differentiable framework that can be trained end-to-end. We
demonstrate the effectiveness of our approach to image reconstruction on both
the Human3.6M and Panoptic Studio datasets. We show how our approach can be
used for motion transfer between individuals; novel view synthesis of
individuals captured from just a single camera; to synthesize individuals from
any virtual viewpoint; and to re-render people in novel poses. Code and video
results are available at
https://github.com/GuillaumeRochette/HumanViewSynthesis.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 10:48:33 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Rochette",
"Guillaume",
""
],
[
"Russell",
"Chris",
""
],
[
"Bowden",
"Richard",
""
]
] |
TITLE: Novel View Synthesis of Humans using Differentiable Rendering
ABSTRACT: We present a new approach for synthesizing novel views of people in new
poses. Our novel differentiable renderer enables the synthesis of highly
realistic images from any viewpoint. Rather than operating over mesh-based
structures, our renderer makes use of diffuse Gaussian primitives that directly
represent the underlying skeletal structure of a human. Rendering these
primitives gives results in a high-dimensional latent image, which is then
transformed into an RGB image by a decoder network. The formulation gives rise
to a fully differentiable framework that can be trained end-to-end. We
demonstrate the effectiveness of our approach to image reconstruction on both
the Human3.6M and Panoptic Studio datasets. We show how our approach can be
used for motion transfer between individuals; novel view synthesis of
individuals captured from just a single camera; to synthesize individuals from
any virtual viewpoint; and to re-render people in novel poses. Code and video
results are available at
https://github.com/GuillaumeRochette/HumanViewSynthesis.
|
no_new_dataset
| 0.712188
|
2303.15900
|
Jeonghwan Kim
|
Jeonghwan Kim, Tianyu Li, Sehoon Ha
|
ARMP: Autoregressive Motion Planning for Quadruped Locomotion and
Navigation in Complex Indoor Environments
|
Submitted to IROS
| null | null | null |
cs.RO
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Generating natural and physically feasible motions for legged robots has been
a challenging problem due to its complex dynamics. In this work, we introduce a
novel learning-based framework of autoregressive motion planner (ARMP) for
quadruped locomotion and navigation. Our method can generate motion plans with
an arbitrary length in an autoregressive fashion, unlike most offline
trajectory optimization algorithms for a fixed trajectory length. To this end,
we first construct the motion library by solving a dense set of trajectory
optimization problems for diverse scenarios and parameter settings. Then we
learn the motion manifold from the dataset in a supervised learning fashion. We
show that the proposed ARMP can generate physically plausible motions for
various tasks and situations. We also showcase that our method can be
successfully integrated with the recent robot navigation frameworks as a
low-level controller and unleash the full capability of legged robots for
complex indoor navigation.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 11:26:13 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Kim",
"Jeonghwan",
""
],
[
"Li",
"Tianyu",
""
],
[
"Ha",
"Sehoon",
""
]
] |
TITLE: ARMP: Autoregressive Motion Planning for Quadruped Locomotion and
Navigation in Complex Indoor Environments
ABSTRACT: Generating natural and physically feasible motions for legged robots has been
a challenging problem due to its complex dynamics. In this work, we introduce a
novel learning-based framework of autoregressive motion planner (ARMP) for
quadruped locomotion and navigation. Our method can generate motion plans with
an arbitrary length in an autoregressive fashion, unlike most offline
trajectory optimization algorithms for a fixed trajectory length. To this end,
we first construct the motion library by solving a dense set of trajectory
optimization problems for diverse scenarios and parameter settings. Then we
learn the motion manifold from the dataset in a supervised learning fashion. We
show that the proposed ARMP can generate physically plausible motions for
various tasks and situations. We also showcase that our method can be
successfully integrated with the recent robot navigation frameworks as a
low-level controller and unleash the full capability of legged robots for
complex indoor navigation.
|
no_new_dataset
| 0.709403
|
2303.15904
|
Lei Ke
|
Lei Ke, Martin Danelljan, Henghui Ding, Yu-Wing Tai, Chi-Keung Tang,
Fisher Yu
|
Mask-Free Video Instance Segmentation
|
Accepted in CVPR 2023; Code: https://github.com/SysCV/MaskFreeVis;
Project page: http://vis.xyz/pub/maskfreevis
| null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The recent advancement in Video Instance Segmentation (VIS) has largely been
driven by the use of deeper and increasingly data-hungry transformer-based
models. However, video masks are tedious and expensive to annotate, limiting
the scale and diversity of existing VIS datasets. In this work, we aim to
remove the mask-annotation requirement. We propose MaskFreeVIS, achieving
highly competitive VIS performance, while only using bounding box annotations
for the object state. We leverage the rich temporal mask consistency
constraints in videos by introducing the Temporal KNN-patch Loss (TK-Loss),
providing strong mask supervision without any labels. Our TK-Loss finds
one-to-many matches across frames, through an efficient patch-matching step
followed by a K-nearest neighbor selection. A consistency loss is then enforced
on the found matches. Our mask-free objective is simple to implement, has no
trainable parameters, is computationally efficient, yet outperforms baselines
employing, e.g., state-of-the-art optical flow to enforce temporal mask
consistency. We validate MaskFreeVIS on the YouTube-VIS 2019/2021, OVIS and
BDD100K MOTS benchmarks. The results clearly demonstrate the efficacy of our
method by drastically narrowing the gap between fully and weakly-supervised VIS
performance. Our code and trained models are available at
https://github.com/SysCV/MaskFreeVis.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 11:48:07 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Ke",
"Lei",
""
],
[
"Danelljan",
"Martin",
""
],
[
"Ding",
"Henghui",
""
],
[
"Tai",
"Yu-Wing",
""
],
[
"Tang",
"Chi-Keung",
""
],
[
"Yu",
"Fisher",
""
]
] |
TITLE: Mask-Free Video Instance Segmentation
ABSTRACT: The recent advancement in Video Instance Segmentation (VIS) has largely been
driven by the use of deeper and increasingly data-hungry transformer-based
models. However, video masks are tedious and expensive to annotate, limiting
the scale and diversity of existing VIS datasets. In this work, we aim to
remove the mask-annotation requirement. We propose MaskFreeVIS, achieving
highly competitive VIS performance, while only using bounding box annotations
for the object state. We leverage the rich temporal mask consistency
constraints in videos by introducing the Temporal KNN-patch Loss (TK-Loss),
providing strong mask supervision without any labels. Our TK-Loss finds
one-to-many matches across frames, through an efficient patch-matching step
followed by a K-nearest neighbor selection. A consistency loss is then enforced
on the found matches. Our mask-free objective is simple to implement, has no
trainable parameters, is computationally efficient, yet outperforms baselines
employing, e.g., state-of-the-art optical flow to enforce temporal mask
consistency. We validate MaskFreeVIS on the YouTube-VIS 2019/2021, OVIS and
BDD100K MOTS benchmarks. The results clearly demonstrate the efficacy of our
method by drastically narrowing the gap between fully and weakly-supervised VIS
performance. Our code and trained models are available at
https://github.com/SysCV/MaskFreeVis.
|
no_new_dataset
| 0.710434
|
2303.15937
|
HsiaoYuan Hsu
|
HsiaoYuan Hsu, Xiangteng He, Yuxin Peng, Hao Kong and Qing Zhang
|
PosterLayout: A New Benchmark and Approach for Content-aware
Visual-Textual Presentation Layout
|
Accepted to CVPR 2023. Dataset and code are available at
https://github.com/PKU-ICST-MIPL/PosterLayout-CVPR2023
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Content-aware visual-textual presentation layout aims at arranging spatial
space on the given canvas for pre-defined elements, including text, logo, and
underlay, which is a key to automatic template-free creative graphic design. In
practical applications, e.g., poster designs, the canvas is originally
non-empty, and both inter-element relationships as well as inter-layer
relationships should be concerned when generating a proper layout. A few recent
works deal with them simultaneously, but they still suffer from poor graphic
performance, such as a lack of layout variety or spatial non-alignment. Since
content-aware visual-textual presentation layout is a novel task, we first
construct a new dataset named PosterLayout, which consists of 9,974
poster-layout pairs and 905 images, i.e., non-empty canvases. It is more
challenging and useful for greater layout variety, domain diversity, and
content diversity. Then, we propose design sequence formation (DSF) that
reorganizes elements in layouts to imitate the design processes of human
designers, and a novel CNN-LSTM-based conditional generative adversarial
network (GAN) is presented to generate proper layouts. Specifically, the
discriminator is design-sequence-aware and will supervise the "design" process
of the generator. Experimental results verify the usefulness of the new
benchmark and the effectiveness of the proposed approach, which achieves the
best performance by generating suitable layouts for diverse canvases.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 12:48:36 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Hsu",
"HsiaoYuan",
""
],
[
"He",
"Xiangteng",
""
],
[
"Peng",
"Yuxin",
""
],
[
"Kong",
"Hao",
""
],
[
"Zhang",
"Qing",
""
]
] |
TITLE: PosterLayout: A New Benchmark and Approach for Content-aware
Visual-Textual Presentation Layout
ABSTRACT: Content-aware visual-textual presentation layout aims at arranging spatial
space on the given canvas for pre-defined elements, including text, logo, and
underlay, which is a key to automatic template-free creative graphic design. In
practical applications, e.g., poster designs, the canvas is originally
non-empty, and both inter-element relationships as well as inter-layer
relationships should be concerned when generating a proper layout. A few recent
works deal with them simultaneously, but they still suffer from poor graphic
performance, such as a lack of layout variety or spatial non-alignment. Since
content-aware visual-textual presentation layout is a novel task, we first
construct a new dataset named PosterLayout, which consists of 9,974
poster-layout pairs and 905 images, i.e., non-empty canvases. It is more
challenging and useful for greater layout variety, domain diversity, and
content diversity. Then, we propose design sequence formation (DSF) that
reorganizes elements in layouts to imitate the design processes of human
designers, and a novel CNN-LSTM-based conditional generative adversarial
network (GAN) is presented to generate proper layouts. Specifically, the
discriminator is design-sequence-aware and will supervise the "design" process
of the generator. Experimental results verify the usefulness of the new
benchmark and the effectiveness of the proposed approach, which achieves the
best performance by generating suitable layouts for diverse canvases.
|
new_dataset
| 0.712664
|
2303.15946
|
Edoardo D'Amico
|
Edoardo D'Amico, Khalil Muhammad, Elias Tragos, Barry Smyth, Neil
Hurley, Aonghus Lawlor
|
Item Graph Convolution Collaborative Filtering for Inductive
Recommendations
| null |
LNCS,volume 13980, pp 249-263, 2023
|
10.1007/978-3-031-28244-7_16
| null |
cs.IR cs.LG
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Graph Convolutional Networks (GCN) have been recently employed as core
component in the construction of recommender system algorithms, interpreting
user-item interactions as the edges of a bipartite graph. However, in the
absence of side information, the majority of existing models adopt an approach
of randomly initialising the user embeddings and optimising them throughout the
training process. This strategy makes these algorithms inherently transductive,
curtailing their ability to generate predictions for users that were unseen at
training time. To address this issue, we propose a convolution-based algorithm,
which is inductive from the user perspective, while at the same time, depending
only on implicit user-item interaction data. We propose the construction of an
item-item graph through a weighted projection of the bipartite interaction
network and to employ convolution to inject higher order associations into item
embeddings, while constructing user representations as weighted sums of the
items with which they have interacted. Despite not training individual
embeddings for each user our approach achieves state of-the-art recommendation
performance with respect to transductive baselines on four real-world datasets,
showing at the same time robust inductive performance.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 12:58:41 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"D'Amico",
"Edoardo",
""
],
[
"Muhammad",
"Khalil",
""
],
[
"Tragos",
"Elias",
""
],
[
"Smyth",
"Barry",
""
],
[
"Hurley",
"Neil",
""
],
[
"Lawlor",
"Aonghus",
""
]
] |
TITLE: Item Graph Convolution Collaborative Filtering for Inductive
Recommendations
ABSTRACT: Graph Convolutional Networks (GCN) have been recently employed as core
component in the construction of recommender system algorithms, interpreting
user-item interactions as the edges of a bipartite graph. However, in the
absence of side information, the majority of existing models adopt an approach
of randomly initialising the user embeddings and optimising them throughout the
training process. This strategy makes these algorithms inherently transductive,
curtailing their ability to generate predictions for users that were unseen at
training time. To address this issue, we propose a convolution-based algorithm,
which is inductive from the user perspective, while at the same time, depending
only on implicit user-item interaction data. We propose the construction of an
item-item graph through a weighted projection of the bipartite interaction
network and to employ convolution to inject higher order associations into item
embeddings, while constructing user representations as weighted sums of the
items with which they have interacted. Despite not training individual
embeddings for each user our approach achieves state of-the-art recommendation
performance with respect to transductive baselines on four real-world datasets,
showing at the same time robust inductive performance.
|
no_new_dataset
| 0.709648
|
2303.15947
|
Ryo Hachiuma
|
Ryo Hachiuma, Tomohiro Shimizu, Hideo Saito, Hiroki Kajita, Yoshifumi
Takatsume
|
Deep Selection: A Fully Supervised Camera Selection Network for Surgery
Recordings
|
MICCAI 2020
| null | null | null |
cs.CV cs.AI cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Recording surgery in operating rooms is an essential task for education and
evaluation of medical treatment. However, recording the desired targets, such
as the surgery field, surgical tools, or doctor's hands, is difficult because
the targets are heavily occluded during surgery. We use a recording system in
which multiple cameras are embedded in the surgical lamp, and we assume that at
least one camera is recording the target without occlusion at any given time.
As the embedded cameras obtain multiple video sequences, we address the task of
selecting the camera with the best view of the surgery. Unlike the conventional
method, which selects the camera based on the area size of the surgery field,
we propose a deep neural network that predicts the camera selection probability
from multiple video sequences by learning the supervision of the expert
annotation. We created a dataset in which six different types of plastic
surgery are recorded, and we provided the annotation of camera switching. Our
experiments show that our approach successfully switched between cameras and
outperformed three baseline methods.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 13:00:08 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Hachiuma",
"Ryo",
""
],
[
"Shimizu",
"Tomohiro",
""
],
[
"Saito",
"Hideo",
""
],
[
"Kajita",
"Hiroki",
""
],
[
"Takatsume",
"Yoshifumi",
""
]
] |
TITLE: Deep Selection: A Fully Supervised Camera Selection Network for Surgery
Recordings
ABSTRACT: Recording surgery in operating rooms is an essential task for education and
evaluation of medical treatment. However, recording the desired targets, such
as the surgery field, surgical tools, or doctor's hands, is difficult because
the targets are heavily occluded during surgery. We use a recording system in
which multiple cameras are embedded in the surgical lamp, and we assume that at
least one camera is recording the target without occlusion at any given time.
As the embedded cameras obtain multiple video sequences, we address the task of
selecting the camera with the best view of the surgery. Unlike the conventional
method, which selects the camera based on the area size of the surgery field,
we propose a deep neural network that predicts the camera selection probability
from multiple video sequences by learning the supervision of the expert
annotation. We created a dataset in which six different types of plastic
surgery are recorded, and we provided the annotation of camera switching. Our
experiments show that our approach successfully switched between cameras and
outperformed three baseline methods.
|
new_dataset
| 0.713856
|
2303.15948
|
Stefanos Eleftheriadis PhD
|
Stefanos Eleftheriadis, Dominic Richards, James Hensman
|
Sparse Gaussian Processes with Spherical Harmonic Features Revisited
| null | null | null | null |
stat.ML cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We revisit the Gaussian process model with spherical harmonic features and
study connections between the associated RKHS, its eigenstructure and deep
models. Based on this, we introduce a new class of kernels which correspond to
deep models of continuous depth. In our formulation, depth can be estimated as
a kernel hyper-parameter by optimizing the evidence lower bound. Further, we
introduce sparseness in the eigenbasis by variational learning of the spherical
harmonic phases. This enables scaling to larger input dimensions than
previously, while also allowing for learning of high frequency variations. We
validate our approach on machine learning benchmark datasets.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 13:02:48 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Eleftheriadis",
"Stefanos",
""
],
[
"Richards",
"Dominic",
""
],
[
"Hensman",
"James",
""
]
] |
TITLE: Sparse Gaussian Processes with Spherical Harmonic Features Revisited
ABSTRACT: We revisit the Gaussian process model with spherical harmonic features and
study connections between the associated RKHS, its eigenstructure and deep
models. Based on this, we introduce a new class of kernels which correspond to
deep models of continuous depth. In our formulation, depth can be estimated as
a kernel hyper-parameter by optimizing the evidence lower bound. Further, we
introduce sparseness in the eigenbasis by variational learning of the spherical
harmonic phases. This enables scaling to larger input dimensions than
previously, while also allowing for learning of high frequency variations. We
validate our approach on machine learning benchmark datasets.
|
no_new_dataset
| 0.711387
|
2303.15951
|
Peng Wang
|
Peng Wang, Yuan Liu, Zhaoxi Chen, Lingjie Liu, Ziwei Liu, Taku Komura,
Christian Theobalt, Wenping Wang
|
F$^{2}$-NeRF: Fast Neural Radiance Field Training with Free Camera
Trajectories
|
CVPR 2023. Project page: https://totoro97.github.io/projects/f2-nerf
| null | null | null |
cs.CV cs.GR
|
http://creativecommons.org/licenses/by/4.0/
|
This paper presents a novel grid-based NeRF called F2-NeRF (Fast-Free-NeRF)
for novel view synthesis, which enables arbitrary input camera trajectories and
only costs a few minutes for training. Existing fast grid-based NeRF training
frameworks, like Instant-NGP, Plenoxels, DVGO, or TensoRF, are mainly designed
for bounded scenes and rely on space warping to handle unbounded scenes.
Existing two widely-used space-warping methods are only designed for the
forward-facing trajectory or the 360-degree object-centric trajectory but
cannot process arbitrary trajectories. In this paper, we delve deep into the
mechanism of space warping to handle unbounded scenes. Based on our analysis,
we further propose a novel space-warping method called perspective warping,
which allows us to handle arbitrary trajectories in the grid-based NeRF
framework. Extensive experiments demonstrate that F2-NeRF is able to use the
same perspective warping to render high-quality images on two standard datasets
and a new free trajectory dataset collected by us. Project page:
https://totoro97.github.io/projects/f2-nerf.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 13:09:44 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Wang",
"Peng",
""
],
[
"Liu",
"Yuan",
""
],
[
"Chen",
"Zhaoxi",
""
],
[
"Liu",
"Lingjie",
""
],
[
"Liu",
"Ziwei",
""
],
[
"Komura",
"Taku",
""
],
[
"Theobalt",
"Christian",
""
],
[
"Wang",
"Wenping",
""
]
] |
TITLE: F$^{2}$-NeRF: Fast Neural Radiance Field Training with Free Camera
Trajectories
ABSTRACT: This paper presents a novel grid-based NeRF called F2-NeRF (Fast-Free-NeRF)
for novel view synthesis, which enables arbitrary input camera trajectories and
only costs a few minutes for training. Existing fast grid-based NeRF training
frameworks, like Instant-NGP, Plenoxels, DVGO, or TensoRF, are mainly designed
for bounded scenes and rely on space warping to handle unbounded scenes.
Existing two widely-used space-warping methods are only designed for the
forward-facing trajectory or the 360-degree object-centric trajectory but
cannot process arbitrary trajectories. In this paper, we delve deep into the
mechanism of space warping to handle unbounded scenes. Based on our analysis,
we further propose a novel space-warping method called perspective warping,
which allows us to handle arbitrary trajectories in the grid-based NeRF
framework. Extensive experiments demonstrate that F2-NeRF is able to use the
same perspective warping to render high-quality images on two standard datasets
and a new free trajectory dataset collected by us. Project page:
https://totoro97.github.io/projects/f2-nerf.
|
new_dataset
| 0.713444
|
2303.15960
|
Raghavendra Badiger
|
Raghavendra Badiger, M. Prabhakar
|
ASCNet-ECG: Deep Autoencoder based Attention aware Skip Connection
network for ECG filtering
| null |
International Journal of Engineering Trends and Technology Volume
71 Issue 2, 382-398, February 2023
|
10.14445/22315381/IJETT-V71I2P240
| null |
eess.SP cs.CY
|
http://creativecommons.org/licenses/by/4.0/
|
Currently, the telehealth monitoring field has gained huge attention due to
its noteworthy use in day-to-day life. This advancement has led to an increase
in the data collection of electrophysiological signals. Due to this
advancement, electrocardiogram (ECG) signal monitoring has become a leading
task in the medical field. ECG plays an important role in the medical field by
analysing cardiac physiology and abnormalities. However, these signals are
affected due to numerous varieties of noises, such as electrode motion,
baseline wander and white noise etc., which affects the diagnosis accuracy.
Therefore, filtering ECG signals became an important task. Currently, deep
learning schemes are widely employed in signal-filtering tasks due to their
efficient architecture of feature learning. This work presents a deep
learning-based scheme for ECG signal filtering, which is based on the deep
autoencoder module. According to this scheme, the data is processed through the
encoder and decoder layer to reconstruct by eliminating noises. The proposed
deep learning architecture uses a modified ReLU function to improve the
learning of attributes because standard ReLU cannot adapt to huge variations.
Further, a skip connection is also incorporated in the proposed architecture,
which retains the key feature of the encoder layer while mapping these features
to the decoder layer. Similarly, an attention model is also included, which
performs channel and spatial attention, which generates the robust map by using
channel and average pooling operations, resulting in improving the learning
performance. The proposed approach is tested on a publicly available MIT-BIH
dataset where different types of noise, such as electrode motion, baseline
water and motion artifacts, are added to the original signal at varied SNR
levels.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 13:23:03 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Badiger",
"Raghavendra",
""
],
[
"Prabhakar",
"M.",
""
]
] |
TITLE: ASCNet-ECG: Deep Autoencoder based Attention aware Skip Connection
network for ECG filtering
ABSTRACT: Currently, the telehealth monitoring field has gained huge attention due to
its noteworthy use in day-to-day life. This advancement has led to an increase
in the data collection of electrophysiological signals. Due to this
advancement, electrocardiogram (ECG) signal monitoring has become a leading
task in the medical field. ECG plays an important role in the medical field by
analysing cardiac physiology and abnormalities. However, these signals are
affected due to numerous varieties of noises, such as electrode motion,
baseline wander and white noise etc., which affects the diagnosis accuracy.
Therefore, filtering ECG signals became an important task. Currently, deep
learning schemes are widely employed in signal-filtering tasks due to their
efficient architecture of feature learning. This work presents a deep
learning-based scheme for ECG signal filtering, which is based on the deep
autoencoder module. According to this scheme, the data is processed through the
encoder and decoder layer to reconstruct by eliminating noises. The proposed
deep learning architecture uses a modified ReLU function to improve the
learning of attributes because standard ReLU cannot adapt to huge variations.
Further, a skip connection is also incorporated in the proposed architecture,
which retains the key feature of the encoder layer while mapping these features
to the decoder layer. Similarly, an attention model is also included, which
performs channel and spatial attention, which generates the robust map by using
channel and average pooling operations, resulting in improving the learning
performance. The proposed approach is tested on a publicly available MIT-BIH
dataset where different types of noise, such as electrode motion, baseline
water and motion artifacts, are added to the original signal at varied SNR
levels.
|
no_new_dataset
| 0.710434
|
2303.15973
|
Suman Adhya
|
Suman Adhya, Avishek Lahiri, Debarshi Kumar Sanyal
|
Do Neural Topic Models Really Need Dropout? Analysis of the Effect of
Dropout in Topic Modeling
|
Accepted at EACL 2023
| null | null | null |
cs.CL cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Dropout is a widely used regularization trick to resolve the overfitting
issue in large feedforward neural networks trained on a small dataset, which
performs poorly on the held-out test subset. Although the effectiveness of this
regularization trick has been extensively studied for convolutional neural
networks, there is a lack of analysis of it for unsupervised models and in
particular, VAE-based neural topic models. In this paper, we have analyzed the
consequences of dropout in the encoder as well as in the decoder of the VAE
architecture in three widely used neural topic models, namely, contextualized
topic model (CTM), ProdLDA, and embedded topic model (ETM) using four publicly
available datasets. We characterize the dropout effect on these models in terms
of the quality and predictive performance of the generated topics.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 13:45:39 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Adhya",
"Suman",
""
],
[
"Lahiri",
"Avishek",
""
],
[
"Sanyal",
"Debarshi Kumar",
""
]
] |
TITLE: Do Neural Topic Models Really Need Dropout? Analysis of the Effect of
Dropout in Topic Modeling
ABSTRACT: Dropout is a widely used regularization trick to resolve the overfitting
issue in large feedforward neural networks trained on a small dataset, which
performs poorly on the held-out test subset. Although the effectiveness of this
regularization trick has been extensively studied for convolutional neural
networks, there is a lack of analysis of it for unsupervised models and in
particular, VAE-based neural topic models. In this paper, we have analyzed the
consequences of dropout in the encoder as well as in the decoder of the VAE
architecture in three widely used neural topic models, namely, contextualized
topic model (CTM), ProdLDA, and embedded topic model (ETM) using four publicly
available datasets. We characterize the dropout effect on these models in terms
of the quality and predictive performance of the generated topics.
|
no_new_dataset
| 0.710603
|
2303.15990
|
Giovanni Rosa
|
Giovanni Rosa, Antonio Mastropaolo, Simone Scalabrino, Gabriele
Bavota, Rocco Oliveto
|
Automatically Generating Dockerfiles via Deep Learning: Challenges and
Promises
| null | null | null | null |
cs.SE
|
http://creativecommons.org/licenses/by/4.0/
|
Containerization allows developers to define the execution environment in
which their software needs to be installed. Docker is the leading platform in
this field, and developers that use it are required to write a Dockerfile for
their software. Writing Dockerfiles is far from trivial, especially when the
system has unusual requirements for its execution environment. Despite several
tools exist to support developers in writing Dockerfiles, none of them is able
to generate entire Dockerfiles from scratch given a high-level specification of
the requirements of the execution environment. In this paper, we present a
study in which we aim at understanding to what extent Deep Learning (DL), which
has been proven successful for other coding tasks, can be used for this
specific coding task. We preliminarily defined a structured natural language
specification for Dockerfile requirements and a methodology that we use to
automatically infer the requirements from the largest dataset of Dockerfiles
currently available. We used the obtained dataset, with 670,982 instances, to
train and test a Text-to-Text Transfer Transformer (T5) model, following the
current state-of-the-art procedure for coding tasks, to automatically generate
Dockerfiles from the structured specifications. The results of our evaluation
show that T5 performs similarly to the more trivial IR-based baselines we
considered. We also report the open challenges associated with the application
of deep learning in the context of Dockerfile generation.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 14:06:28 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Rosa",
"Giovanni",
""
],
[
"Mastropaolo",
"Antonio",
""
],
[
"Scalabrino",
"Simone",
""
],
[
"Bavota",
"Gabriele",
""
],
[
"Oliveto",
"Rocco",
""
]
] |
TITLE: Automatically Generating Dockerfiles via Deep Learning: Challenges and
Promises
ABSTRACT: Containerization allows developers to define the execution environment in
which their software needs to be installed. Docker is the leading platform in
this field, and developers that use it are required to write a Dockerfile for
their software. Writing Dockerfiles is far from trivial, especially when the
system has unusual requirements for its execution environment. Despite several
tools exist to support developers in writing Dockerfiles, none of them is able
to generate entire Dockerfiles from scratch given a high-level specification of
the requirements of the execution environment. In this paper, we present a
study in which we aim at understanding to what extent Deep Learning (DL), which
has been proven successful for other coding tasks, can be used for this
specific coding task. We preliminarily defined a structured natural language
specification for Dockerfile requirements and a methodology that we use to
automatically infer the requirements from the largest dataset of Dockerfiles
currently available. We used the obtained dataset, with 670,982 instances, to
train and test a Text-to-Text Transfer Transformer (T5) model, following the
current state-of-the-art procedure for coding tasks, to automatically generate
Dockerfiles from the structured specifications. The results of our evaluation
show that T5 performs similarly to the more trivial IR-based baselines we
considered. We also report the open challenges associated with the application
of deep learning in the context of Dockerfile generation.
|
no_new_dataset
| 0.702951
|
2303.15993
|
Hojjat Mokhtarabadi
|
Hojjat Mokhtarabadi, Kave Bahraman, Mehrdad HosseinZadeh, Mahdi
Eftekhari
|
SELF-VS: Self-supervised Encoding Learning For Video Summarization
|
9 pages, 5 figures
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Despite its wide range of applications, video summarization is still held
back by the scarcity of extensive datasets, largely due to the labor-intensive
and costly nature of frame-level annotations. As a result, existing video
summarization methods are prone to overfitting. To mitigate this challenge, we
propose a novel self-supervised video representation learning method using
knowledge distillation to pre-train a transformer encoder. Our method matches
its semantic video representation, which is constructed with respect to frame
importance scores, to a representation derived from a CNN trained on video
classification. Empirical evaluations on correlation-based metrics, such as
Kendall's $\tau$ and Spearman's $\rho$ demonstrate the superiority of our
approach compared to existing state-of-the-art methods in assigning relative
scores to the input frames.
|
[
{
"version": "v1",
"created": "Tue, 28 Mar 2023 14:08:05 GMT"
}
] | 2023-03-29T00:00:00
|
[
[
"Mokhtarabadi",
"Hojjat",
""
],
[
"Bahraman",
"Kave",
""
],
[
"HosseinZadeh",
"Mehrdad",
""
],
[
"Eftekhari",
"Mahdi",
""
]
] |
TITLE: SELF-VS: Self-supervised Encoding Learning For Video Summarization
ABSTRACT: Despite its wide range of applications, video summarization is still held
back by the scarcity of extensive datasets, largely due to the labor-intensive
and costly nature of frame-level annotations. As a result, existing video
summarization methods are prone to overfitting. To mitigate this challenge, we
propose a novel self-supervised video representation learning method using
knowledge distillation to pre-train a transformer encoder. Our method matches
its semantic video representation, which is constructed with respect to frame
importance scores, to a representation derived from a CNN trained on video
classification. Empirical evaluations on correlation-based metrics, such as
Kendall's $\tau$ and Spearman's $\rho$ demonstrate the superiority of our
approach compared to existing state-of-the-art methods in assigning relative
scores to the input frames.
|
no_new_dataset
| 0.709636
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.