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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2306.09441 | Amin Yousefpour | Amin Yousefpour, Mehdi Shishehbor, Zahra Zanjani Foumani, Ramin
Bostanabad | Unsupervised Anomaly Detection via Nonlinear Manifold Learning | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Anomalies are samples that significantly deviate from the rest of the data
and their detection plays a major role in building machine learning models that
can be reliably used in applications such as data-driven design and novelty
detection. The majority of existing anomaly detection methods either are
exclusively developed for (semi) supervised settings, or provide poor
performance in unsupervised applications where there is no training data with
labeled anomalous samples. To bridge this research gap, we introduce a robust,
efficient, and interpretable methodology based on nonlinear manifold learning
to detect anomalies in unsupervised settings. The essence of our approach is to
learn a low-dimensional and interpretable latent representation (aka manifold)
for all the data points such that normal samples are automatically clustered
together and hence can be easily and robustly identified. We learn this
low-dimensional manifold by designing a learning algorithm that leverages
either a latent map Gaussian process (LMGP) or a deep autoencoder (AE). Our
LMGP-based approach, in particular, provides a probabilistic perspective on the
learning task and is ideal for high-dimensional applications with scarce data.
We demonstrate the superior performance of our approach over existing
technologies via multiple analytic examples and real-world datasets.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 18:48:10 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Yousefpour",
"Amin",
""
],
[
"Shishehbor",
"Mehdi",
""
],
[
"Foumani",
"Zahra Zanjani",
""
],
[
"Bostanabad",
"Ramin",
""
]
] | TITLE: Unsupervised Anomaly Detection via Nonlinear Manifold Learning
ABSTRACT: Anomalies are samples that significantly deviate from the rest of the data
and their detection plays a major role in building machine learning models that
can be reliably used in applications such as data-driven design and novelty
detection. The majority of existing anomaly detection methods either are
exclusively developed for (semi) supervised settings, or provide poor
performance in unsupervised applications where there is no training data with
labeled anomalous samples. To bridge this research gap, we introduce a robust,
efficient, and interpretable methodology based on nonlinear manifold learning
to detect anomalies in unsupervised settings. The essence of our approach is to
learn a low-dimensional and interpretable latent representation (aka manifold)
for all the data points such that normal samples are automatically clustered
together and hence can be easily and robustly identified. We learn this
low-dimensional manifold by designing a learning algorithm that leverages
either a latent map Gaussian process (LMGP) or a deep autoencoder (AE). Our
LMGP-based approach, in particular, provides a probabilistic perspective on the
learning task and is ideal for high-dimensional applications with scarce data.
We demonstrate the superior performance of our approach over existing
technologies via multiple analytic examples and real-world datasets.
|
2306.09451 | Burak Kantarci | Zhiyan Chen, Murat Simsek, Burak Kantarci, Mehran Bagheri, Petar
Djukic | Host-Based Network Intrusion Detection via Feature Flattening and
Two-stage Collaborative Classifier | 20 pages, 11 figures, 4 tables | null | null | null | cs.CR cs.AI cs.NI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Network Intrusion Detection Systems (NIDS) have been extensively investigated
by monitoring real network traffic and analyzing suspicious activities.
However, there are limitations in detecting specific types of attacks with
NIDS, such as Advanced Persistent Threats (APT). Additionally, NIDS is
restricted in observing complete traffic information due to encrypted traffic
or a lack of authority. To address these limitations, a Host-based Intrusion
Detection system (HIDS) evaluates resources in the host, including logs, files,
and folders, to identify APT attacks that routinely inject malicious files into
victimized nodes. In this study, a hybrid network intrusion detection system
that combines NIDS and HIDS is proposed to improve intrusion detection
performance. The feature flattening technique is applied to flatten
two-dimensional host-based features into one-dimensional vectors, which can be
directly used by traditional Machine Learning (ML) models. A two-stage
collaborative classifier is introduced that deploys two levels of ML algorithms
to identify network intrusions. In the first stage, a binary classifier is used
to detect benign samples. All detected attack types undergo a multi-class
classifier to reduce the complexity of the original problem and improve the
overall detection performance. The proposed method is shown to generalize
across two well-known datasets, CICIDS 2018 and NDSec-1. Performance of
XGBoost, which represents conventional ML, is evaluated. Combining host and
network features enhances attack detection performance (macro average F1 score)
by 8.1% under the CICIDS 2018 dataset and 3.7% under the NDSec-1 dataset.
Meanwhile, the two-stage collaborative classifier improves detection
performance for most single classes, especially for DoS-LOIC-UDP and
DoS-SlowHTTPTest, with improvements of 30.7% and 84.3%, respectively, when
compared with the traditional ML XGBoost.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 19:09:00 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Chen",
"Zhiyan",
""
],
[
"Simsek",
"Murat",
""
],
[
"Kantarci",
"Burak",
""
],
[
"Bagheri",
"Mehran",
""
],
[
"Djukic",
"Petar",
""
]
] | TITLE: Host-Based Network Intrusion Detection via Feature Flattening and
Two-stage Collaborative Classifier
ABSTRACT: Network Intrusion Detection Systems (NIDS) have been extensively investigated
by monitoring real network traffic and analyzing suspicious activities.
However, there are limitations in detecting specific types of attacks with
NIDS, such as Advanced Persistent Threats (APT). Additionally, NIDS is
restricted in observing complete traffic information due to encrypted traffic
or a lack of authority. To address these limitations, a Host-based Intrusion
Detection system (HIDS) evaluates resources in the host, including logs, files,
and folders, to identify APT attacks that routinely inject malicious files into
victimized nodes. In this study, a hybrid network intrusion detection system
that combines NIDS and HIDS is proposed to improve intrusion detection
performance. The feature flattening technique is applied to flatten
two-dimensional host-based features into one-dimensional vectors, which can be
directly used by traditional Machine Learning (ML) models. A two-stage
collaborative classifier is introduced that deploys two levels of ML algorithms
to identify network intrusions. In the first stage, a binary classifier is used
to detect benign samples. All detected attack types undergo a multi-class
classifier to reduce the complexity of the original problem and improve the
overall detection performance. The proposed method is shown to generalize
across two well-known datasets, CICIDS 2018 and NDSec-1. Performance of
XGBoost, which represents conventional ML, is evaluated. Combining host and
network features enhances attack detection performance (macro average F1 score)
by 8.1% under the CICIDS 2018 dataset and 3.7% under the NDSec-1 dataset.
Meanwhile, the two-stage collaborative classifier improves detection
performance for most single classes, especially for DoS-LOIC-UDP and
DoS-SlowHTTPTest, with improvements of 30.7% and 84.3%, respectively, when
compared with the traditional ML XGBoost.
|
2306.09457 | Fnu Shilpika | Shilpika, Bethany Lusch, Murali Emani, Filippo Simini, Venkatram
Vishwanath, Michael E. Papka, and Kwan-Liu Ma | A Multi-Level, Multi-Scale Visual Analytics Approach to Assessment of
Multifidelity HPC Systems | null | null | null | null | cs.HC cs.CV | http://creativecommons.org/licenses/by/4.0/ | The ability to monitor and interpret of hardware system events and behaviors
are crucial to improving the robustness and reliability of these systems,
especially in a supercomputing facility. The growing complexity and scale of
these systems demand an increase in monitoring data collected at multiple
fidelity levels and varying temporal resolutions. In this work, we aim to build
a holistic analytical system that helps make sense of such massive data, mainly
the hardware logs, job logs, and environment logs collected from disparate
subsystems and components of a supercomputer system. This end-to-end log
analysis system, coupled with visual analytics support, allows users to glean
and promptly extract supercomputer usage and error patterns at varying temporal
and spatial resolutions. We use multiresolution dynamic mode decomposition
(mrDMD), a technique that depicts high-dimensional data as correlated
spatial-temporal variations patterns or modes, to extract variation patterns
isolated at specified frequencies. Our improvements to the mrDMD algorithm help
promptly reveal useful information in the massive environment log dataset,
which is then associated with the processed hardware and job log datasets using
our visual analytics system. Furthermore, our system can identify the usage and
error patterns filtered at user, project, and subcomponent levels. We exemplify
the effectiveness of our approach with two use scenarios with the Cray XC40
supercomputer.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 19:23:50 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Shilpika",
"",
""
],
[
"Lusch",
"Bethany",
""
],
[
"Emani",
"Murali",
""
],
[
"Simini",
"Filippo",
""
],
[
"Vishwanath",
"Venkatram",
""
],
[
"Papka",
"Michael E.",
""
],
[
"Ma",
"Kwan-Liu",
""
]
] | TITLE: A Multi-Level, Multi-Scale Visual Analytics Approach to Assessment of
Multifidelity HPC Systems
ABSTRACT: The ability to monitor and interpret of hardware system events and behaviors
are crucial to improving the robustness and reliability of these systems,
especially in a supercomputing facility. The growing complexity and scale of
these systems demand an increase in monitoring data collected at multiple
fidelity levels and varying temporal resolutions. In this work, we aim to build
a holistic analytical system that helps make sense of such massive data, mainly
the hardware logs, job logs, and environment logs collected from disparate
subsystems and components of a supercomputer system. This end-to-end log
analysis system, coupled with visual analytics support, allows users to glean
and promptly extract supercomputer usage and error patterns at varying temporal
and spatial resolutions. We use multiresolution dynamic mode decomposition
(mrDMD), a technique that depicts high-dimensional data as correlated
spatial-temporal variations patterns or modes, to extract variation patterns
isolated at specified frequencies. Our improvements to the mrDMD algorithm help
promptly reveal useful information in the massive environment log dataset,
which is then associated with the processed hardware and job log datasets using
our visual analytics system. Furthermore, our system can identify the usage and
error patterns filtered at user, project, and subcomponent levels. We exemplify
the effectiveness of our approach with two use scenarios with the Cray XC40
supercomputer.
|
2306.09470 | Wael Suleiman | Jorge Ocampo Jimenez and Wael Suleiman | Improving Path Planning Performance through Multimodal Generative Models
with Local Critics | Under review | null | null | null | cs.RO cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel method for accelerating path planning tasks in
unknown scenes with obstacles by utilizing Wasserstein Generative Adversarial
Networks (WGANs) with Gradient Penalty (GP) to approximate the distribution of
the free conditioned configuration space. Our proposed approach involves
conditioning the WGAN-GP with a Variational Auto-Encoder in a continuous latent
space to handle multimodal datasets. However, training a Variational
Auto-Encoder with WGAN-GP can be challenging for image-to-configuration-space
problems, as the Kullback-Leibler loss function often converges to a random
distribution. To overcome this issue, we simplify the configuration space as a
set of Gaussian distributions and divide the dataset into several local models.
This enables us to not only learn the model but also speed up its convergence.
We evaluate the reconstructed configuration space using the homology rank of
manifolds for datasets with the geometry score. Furthermore, we propose a novel
transformation of the robot's configuration space that enables us to measure
how well collision-free regions are reconstructed, which could be used with
other rank of homology metrics. Our experiments show promising results for
accelerating path planning tasks in unknown scenes while generating
quasi-optimal paths with our WGAN-GP. The source code is openly available.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 19:51:35 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Jimenez",
"Jorge Ocampo",
""
],
[
"Suleiman",
"Wael",
""
]
] | TITLE: Improving Path Planning Performance through Multimodal Generative Models
with Local Critics
ABSTRACT: This paper presents a novel method for accelerating path planning tasks in
unknown scenes with obstacles by utilizing Wasserstein Generative Adversarial
Networks (WGANs) with Gradient Penalty (GP) to approximate the distribution of
the free conditioned configuration space. Our proposed approach involves
conditioning the WGAN-GP with a Variational Auto-Encoder in a continuous latent
space to handle multimodal datasets. However, training a Variational
Auto-Encoder with WGAN-GP can be challenging for image-to-configuration-space
problems, as the Kullback-Leibler loss function often converges to a random
distribution. To overcome this issue, we simplify the configuration space as a
set of Gaussian distributions and divide the dataset into several local models.
This enables us to not only learn the model but also speed up its convergence.
We evaluate the reconstructed configuration space using the homology rank of
manifolds for datasets with the geometry score. Furthermore, we propose a novel
transformation of the robot's configuration space that enables us to measure
how well collision-free regions are reconstructed, which could be used with
other rank of homology metrics. Our experiments show promising results for
accelerating path planning tasks in unknown scenes while generating
quasi-optimal paths with our WGAN-GP. The source code is openly available.
|
2306.09482 | Sarthak Bhagat | Sarthak Bhagat, Simon Stepputtis, Joseph Campbell, Katia Sycara | Sample-Efficient Learning of Novel Visual Concepts | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the advances made in visual object recognition, state-of-the-art deep
learning models struggle to effectively recognize novel objects in a few-shot
setting where only a limited number of examples are provided. Unlike humans who
excel at such tasks, these models often fail to leverage known relationships
between entities in order to draw conclusions about such objects. In this work,
we show that incorporating a symbolic knowledge graph into a state-of-the-art
recognition model enables a new approach for effective few-shot classification.
In our proposed neuro-symbolic architecture and training methodology, the
knowledge graph is augmented with additional relationships extracted from a
small set of examples, improving its ability to recognize novel objects by
considering the presence of interconnected entities. Unlike existing few-shot
classifiers, we show that this enables our model to incorporate not only
objects but also abstract concepts and affordances. The existence of the
knowledge graph also makes this approach amenable to interpretability through
analysis of the relationships contained within it. We empirically show that our
approach outperforms current state-of-the-art few-shot multi-label
classification methods on the COCO dataset and evaluate the addition of
abstract concepts and affordances on the Visual Genome dataset.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 20:24:30 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Bhagat",
"Sarthak",
""
],
[
"Stepputtis",
"Simon",
""
],
[
"Campbell",
"Joseph",
""
],
[
"Sycara",
"Katia",
""
]
] | TITLE: Sample-Efficient Learning of Novel Visual Concepts
ABSTRACT: Despite the advances made in visual object recognition, state-of-the-art deep
learning models struggle to effectively recognize novel objects in a few-shot
setting where only a limited number of examples are provided. Unlike humans who
excel at such tasks, these models often fail to leverage known relationships
between entities in order to draw conclusions about such objects. In this work,
we show that incorporating a symbolic knowledge graph into a state-of-the-art
recognition model enables a new approach for effective few-shot classification.
In our proposed neuro-symbolic architecture and training methodology, the
knowledge graph is augmented with additional relationships extracted from a
small set of examples, improving its ability to recognize novel objects by
considering the presence of interconnected entities. Unlike existing few-shot
classifiers, we show that this enables our model to incorporate not only
objects but also abstract concepts and affordances. The existence of the
knowledge graph also makes this approach amenable to interpretability through
analysis of the relationships contained within it. We empirically show that our
approach outperforms current state-of-the-art few-shot multi-label
classification methods on the COCO dataset and evaluate the addition of
abstract concepts and affordances on the Visual Genome dataset.
|
2306.09483 | Takeru Oba | Takeru Oba and Norimichi Ukita | R2-Diff: Denoising by diffusion as a refinement of retrieved motion for
image-based motion prediction | 22 pages, preprint submitted to Neurocomputing | null | null | null | cs.CV cs.LG cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-based motion prediction is one of the essential techniques for robot
manipulation. Among the various prediction models, we focus on diffusion models
because they have achieved state-of-the-art performance in various
applications. In image-based motion prediction, diffusion models stochastically
predict contextually appropriate motion by gradually denoising random Gaussian
noise based on the image context. While diffusion models are able to predict
various motions by changing the random noise, they sometimes fail to predict a
contextually appropriate motion based on the image because the random noise is
sampled independently of the image context. To solve this problem, we propose
R2-Diff. In R2-Diff, a motion retrieved from a dataset based on image
similarity is fed into a diffusion model instead of random noise. Then, the
retrieved motion is refined through the denoising process of the diffusion
model. Since the retrieved motion is almost appropriate to the context, it
becomes easier to predict contextually appropriate motion. However, traditional
diffusion models are not optimized to refine the retrieved motion. Therefore,
we propose the method of tuning the hyperparameters based on the distance of
the nearest neighbor motion among the dataset to optimize the diffusion model
for refinement. Furthermore, we propose an image-based retrieval method to
retrieve the nearest neighbor motion in inference. Our proposed retrieval
efficiently computes the similarity based on the image features along the
motion trajectory. We demonstrate that R2-Diff accurately predicts appropriate
motions and achieves high task success rates compared to recent
state-of-the-art models in robot manipulation.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 20:27:06 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Oba",
"Takeru",
""
],
[
"Ukita",
"Norimichi",
""
]
] | TITLE: R2-Diff: Denoising by diffusion as a refinement of retrieved motion for
image-based motion prediction
ABSTRACT: Image-based motion prediction is one of the essential techniques for robot
manipulation. Among the various prediction models, we focus on diffusion models
because they have achieved state-of-the-art performance in various
applications. In image-based motion prediction, diffusion models stochastically
predict contextually appropriate motion by gradually denoising random Gaussian
noise based on the image context. While diffusion models are able to predict
various motions by changing the random noise, they sometimes fail to predict a
contextually appropriate motion based on the image because the random noise is
sampled independently of the image context. To solve this problem, we propose
R2-Diff. In R2-Diff, a motion retrieved from a dataset based on image
similarity is fed into a diffusion model instead of random noise. Then, the
retrieved motion is refined through the denoising process of the diffusion
model. Since the retrieved motion is almost appropriate to the context, it
becomes easier to predict contextually appropriate motion. However, traditional
diffusion models are not optimized to refine the retrieved motion. Therefore,
we propose the method of tuning the hyperparameters based on the distance of
the nearest neighbor motion among the dataset to optimize the diffusion model
for refinement. Furthermore, we propose an image-based retrieval method to
retrieve the nearest neighbor motion in inference. Our proposed retrieval
efficiently computes the similarity based on the image features along the
motion trajectory. We demonstrate that R2-Diff accurately predicts appropriate
motions and achieves high task success rates compared to recent
state-of-the-art models in robot manipulation.
|
2306.09485 | Elohim Fonseca Dos Reis | Elohim Fonseca dos Reis, Alexander Teytelboym, Abeer ElBahraw, Ignacio
De Loizaga, Andrea Baronchelli | Identifying key players in dark web marketplaces | null | null | null | null | physics.soc-ph cs.CY econ.GN q-fin.EC | http://creativecommons.org/licenses/by/4.0/ | Dark web marketplaces have been a significant outlet for illicit trade,
serving millions of users worldwide for over a decade. However, not all users
are the same. This paper aims to identify the key players in Bitcoin
transaction networks linked to dark markets and assess their role by analysing
a dataset of 40 million Bitcoin transactions involving 31 markets in the period
2011-2021. First, we propose an algorithm that categorizes users either as
buyers or sellers and shows that a large fraction of the traded volume is
concentrated in a small group of elite market participants. Then, we
investigate both market star-graphs and user-to-user networks and highlight the
importance of a new class of users, namely `multihomers' who operate on
multiple marketplaces concurrently. Specifically, we show how the networks of
multihomers and seller-to-seller interactions can shed light on the resilience
of the dark market ecosystem against external shocks. Our findings suggest that
understanding the behavior of key players in dark web marketplaces is critical
to effectively disrupting illegal activities.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 20:30:43 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Reis",
"Elohim Fonseca dos",
""
],
[
"Teytelboym",
"Alexander",
""
],
[
"ElBahraw",
"Abeer",
""
],
[
"De Loizaga",
"Ignacio",
""
],
[
"Baronchelli",
"Andrea",
""
]
] | TITLE: Identifying key players in dark web marketplaces
ABSTRACT: Dark web marketplaces have been a significant outlet for illicit trade,
serving millions of users worldwide for over a decade. However, not all users
are the same. This paper aims to identify the key players in Bitcoin
transaction networks linked to dark markets and assess their role by analysing
a dataset of 40 million Bitcoin transactions involving 31 markets in the period
2011-2021. First, we propose an algorithm that categorizes users either as
buyers or sellers and shows that a large fraction of the traded volume is
concentrated in a small group of elite market participants. Then, we
investigate both market star-graphs and user-to-user networks and highlight the
importance of a new class of users, namely `multihomers' who operate on
multiple marketplaces concurrently. Specifically, we show how the networks of
multihomers and seller-to-seller interactions can shed light on the resilience
of the dark market ecosystem against external shocks. Our findings suggest that
understanding the behavior of key players in dark web marketplaces is critical
to effectively disrupting illegal activities.
|
2306.09489 | Matthijs Douze | Ed Pizzi and Giorgos Kordopatis-Zilos and Hiral Patel and Gheorghe
Postelnicu and Sugosh Nagavara Ravindra and Akshay Gupta and Symeon
Papadopoulos and Giorgos Tolias and Matthijs Douze | The 2023 Video Similarity Dataset and Challenge | null | null | null | null | cs.CV cs.AI cs.MM | http://creativecommons.org/licenses/by/4.0/ | This work introduces a dataset, benchmark, and challenge for the problem of
video copy detection and localization. The problem comprises two distinct but
related tasks: determining whether a query video shares content with a
reference video ("detection"), and additionally temporally localizing the
shared content within each video ("localization"). The benchmark is designed to
evaluate methods on these two tasks, and simulates a realistic
needle-in-haystack setting, where the majority of both query and reference
videos are "distractors" containing no copied content. We propose a metric that
reflects both detection and localization accuracy. The associated challenge
consists of two corresponding tracks, each with restrictions that reflect
real-world settings. We provide implementation code for evaluation and
baselines. We also analyze the results and methods of the top submissions to
the challenge. The dataset, baseline methods and evaluation code is publicly
available and will be discussed at a dedicated CVPR'23 workshop.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 20:34:43 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Pizzi",
"Ed",
""
],
[
"Kordopatis-Zilos",
"Giorgos",
""
],
[
"Patel",
"Hiral",
""
],
[
"Postelnicu",
"Gheorghe",
""
],
[
"Ravindra",
"Sugosh Nagavara",
""
],
[
"Gupta",
"Akshay",
""
],
[
"Papadopoulos",
"Symeon",
""
],
[
"Tolias",
"Giorgos",
""
],
[
"Douze",
"Matthijs",
""
]
] | TITLE: The 2023 Video Similarity Dataset and Challenge
ABSTRACT: This work introduces a dataset, benchmark, and challenge for the problem of
video copy detection and localization. The problem comprises two distinct but
related tasks: determining whether a query video shares content with a
reference video ("detection"), and additionally temporally localizing the
shared content within each video ("localization"). The benchmark is designed to
evaluate methods on these two tasks, and simulates a realistic
needle-in-haystack setting, where the majority of both query and reference
videos are "distractors" containing no copied content. We propose a metric that
reflects both detection and localization accuracy. The associated challenge
consists of two corresponding tracks, each with restrictions that reflect
real-world settings. We provide implementation code for evaluation and
baselines. We also analyze the results and methods of the top submissions to
the challenge. The dataset, baseline methods and evaluation code is publicly
available and will be discussed at a dedicated CVPR'23 workshop.
|
2306.09519 | Qiao Qiao | Qiao Qiao, Yuepei Li, Kang Zhou, Qi Li | Relation-Aware Network with Attention-Based Loss for Few-Shot Knowledge
Graph Completion | conference PAKDD 2023 | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Few-shot knowledge graph completion (FKGC) task aims to predict unseen facts
of a relation with few-shot reference entity pairs. Current approaches randomly
select one negative sample for each reference entity pair to minimize a
margin-based ranking loss, which easily leads to a zero-loss problem if the
negative sample is far away from the positive sample and then out of the
margin. Moreover, the entity should have a different representation under a
different context. To tackle these issues, we propose a novel Relation-Aware
Network with Attention-Based Loss (RANA) framework. Specifically, to better
utilize the plentiful negative samples and alleviate the zero-loss issue, we
strategically select relevant negative samples and design an attention-based
loss function to further differentiate the importance of each negative sample.
The intuition is that negative samples more similar to positive samples will
contribute more to the model. Further, we design a dynamic relation-aware
entity encoder for learning a context-dependent entity representation.
Experiments demonstrate that RANA outperforms the state-of-the-art models on
two benchmark datasets.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 21:41:43 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Qiao",
"Qiao",
""
],
[
"Li",
"Yuepei",
""
],
[
"Zhou",
"Kang",
""
],
[
"Li",
"Qi",
""
]
] | TITLE: Relation-Aware Network with Attention-Based Loss for Few-Shot Knowledge
Graph Completion
ABSTRACT: Few-shot knowledge graph completion (FKGC) task aims to predict unseen facts
of a relation with few-shot reference entity pairs. Current approaches randomly
select one negative sample for each reference entity pair to minimize a
margin-based ranking loss, which easily leads to a zero-loss problem if the
negative sample is far away from the positive sample and then out of the
margin. Moreover, the entity should have a different representation under a
different context. To tackle these issues, we propose a novel Relation-Aware
Network with Attention-Based Loss (RANA) framework. Specifically, to better
utilize the plentiful negative samples and alleviate the zero-loss issue, we
strategically select relevant negative samples and design an attention-based
loss function to further differentiate the importance of each negative sample.
The intuition is that negative samples more similar to positive samples will
contribute more to the model. Further, we design a dynamic relation-aware
entity encoder for learning a context-dependent entity representation.
Experiments demonstrate that RANA outperforms the state-of-the-art models on
two benchmark datasets.
|
2306.09527 | Alberto Rodriguez | Katherine R. Dearstyne, Alberto D. Rodriguez | Leveraging Human Salience to Improve Calorie Estimation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The following paper investigates the effectiveness of incorporating human
salience into the task of calorie prediction from images of food. We observe a
32.2% relative improvement when incorporating saliency maps on the images of
food highlighting the most calorie regions. We also attempt to further improve
the accuracy by starting the best models using pre-trained weights on similar
tasks of mass estimation and food classification. However, we observe no
improvement. Surprisingly, we also find that our best model was not able to
surpass the original performance published alongside the test dataset,
Nutrition5k. We use ResNet50 and Xception as the base models for our
experiment.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 22:13:26 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Dearstyne",
"Katherine R.",
""
],
[
"Rodriguez",
"Alberto D.",
""
]
] | TITLE: Leveraging Human Salience to Improve Calorie Estimation
ABSTRACT: The following paper investigates the effectiveness of incorporating human
salience into the task of calorie prediction from images of food. We observe a
32.2% relative improvement when incorporating saliency maps on the images of
food highlighting the most calorie regions. We also attempt to further improve
the accuracy by starting the best models using pre-trained weights on similar
tasks of mass estimation and food classification. However, we observe no
improvement. Surprisingly, we also find that our best model was not able to
surpass the original performance published alongside the test dataset,
Nutrition5k. We use ResNet50 and Xception as the base models for our
experiment.
|
2306.09546 | Ali Abedi | Ali Abedi, Mobin Malmirian, and Shehroz S. Khan | Cross-Modal Video to Body-joints Augmentation for Rehabilitation
Exercise Quality Assessment | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Exercise-based rehabilitation programs have been shown to enhance quality of
life and reduce mortality and rehospitalizations. AI-driven virtual
rehabilitation programs enable patients to complete exercises independently at
home while AI algorithms can analyze exercise data to provide feedback to
patients and report their progress to clinicians. This paper introduces a novel
approach to assessing the quality of rehabilitation exercises using RGB video.
Sequences of skeletal body joints are extracted from consecutive RGB video
frames and analyzed by many-to-one sequential neural networks to evaluate
exercise quality. Existing datasets for exercise rehabilitation lack adequate
samples for training deep sequential neural networks to generalize effectively.
A cross-modal data augmentation approach is proposed to resolve this problem.
Visual augmentation techniques are applied to video data, and body joints
extracted from the resulting augmented videos are used for training sequential
neural networks. Extensive experiments conducted on the KInematic assessment of
MOvement and clinical scores for remote monitoring of physical REhabilitation
(KIMORE) dataset, demonstrate the superiority of the proposed method over
previous baseline approaches. The ablation study highlights a significant
enhancement in exercise quality assessment following cross-modal augmentation.
| [
{
"version": "v1",
"created": "Thu, 15 Jun 2023 23:23:35 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Abedi",
"Ali",
""
],
[
"Malmirian",
"Mobin",
""
],
[
"Khan",
"Shehroz S.",
""
]
] | TITLE: Cross-Modal Video to Body-joints Augmentation for Rehabilitation
Exercise Quality Assessment
ABSTRACT: Exercise-based rehabilitation programs have been shown to enhance quality of
life and reduce mortality and rehospitalizations. AI-driven virtual
rehabilitation programs enable patients to complete exercises independently at
home while AI algorithms can analyze exercise data to provide feedback to
patients and report their progress to clinicians. This paper introduces a novel
approach to assessing the quality of rehabilitation exercises using RGB video.
Sequences of skeletal body joints are extracted from consecutive RGB video
frames and analyzed by many-to-one sequential neural networks to evaluate
exercise quality. Existing datasets for exercise rehabilitation lack adequate
samples for training deep sequential neural networks to generalize effectively.
A cross-modal data augmentation approach is proposed to resolve this problem.
Visual augmentation techniques are applied to video data, and body joints
extracted from the resulting augmented videos are used for training sequential
neural networks. Extensive experiments conducted on the KInematic assessment of
MOvement and clinical scores for remote monitoring of physical REhabilitation
(KIMORE) dataset, demonstrate the superiority of the proposed method over
previous baseline approaches. The ablation study highlights a significant
enhancement in exercise quality assessment following cross-modal augmentation.
|
2306.09562 | Ian Magnusson | Ian Magnusson, Noah A. Smith, Jesse Dodge | Reproducibility in NLP: What Have We Learned from the Checklist? | To be published in ACL 2023 Findings | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Scientific progress in NLP rests on the reproducibility of researchers'
claims. The *CL conferences created the NLP Reproducibility Checklist in 2020
to be completed by authors at submission to remind them of key information to
include. We provide the first analysis of the Checklist by examining 10,405
anonymous responses to it. First, we find evidence of an increase in reporting
of information on efficiency, validation performance, summary statistics, and
hyperparameters after the Checklist's introduction. Further, we show acceptance
rate grows for submissions with more Yes responses. We find that the 44% of
submissions that gather new data are 5% less likely to be accepted than those
that did not; the average reviewer-rated reproducibility of these submissions
is also 2% lower relative to the rest. We find that only 46% of submissions
claim to open-source their code, though submissions that do have 8% higher
reproducibility score relative to those that do not, the most for any item. We
discuss what can be inferred about the state of reproducibility in NLP, and
provide a set of recommendations for future conferences, including: a) allowing
submitting code and appendices one week after the deadline, and b) measuring
dataset reproducibility by a checklist of data collection practices.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 00:39:25 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Magnusson",
"Ian",
""
],
[
"Smith",
"Noah A.",
""
],
[
"Dodge",
"Jesse",
""
]
] | TITLE: Reproducibility in NLP: What Have We Learned from the Checklist?
ABSTRACT: Scientific progress in NLP rests on the reproducibility of researchers'
claims. The *CL conferences created the NLP Reproducibility Checklist in 2020
to be completed by authors at submission to remind them of key information to
include. We provide the first analysis of the Checklist by examining 10,405
anonymous responses to it. First, we find evidence of an increase in reporting
of information on efficiency, validation performance, summary statistics, and
hyperparameters after the Checklist's introduction. Further, we show acceptance
rate grows for submissions with more Yes responses. We find that the 44% of
submissions that gather new data are 5% less likely to be accepted than those
that did not; the average reviewer-rated reproducibility of these submissions
is also 2% lower relative to the rest. We find that only 46% of submissions
claim to open-source their code, though submissions that do have 8% higher
reproducibility score relative to those that do not, the most for any item. We
discuss what can be inferred about the state of reproducibility in NLP, and
provide a set of recommendations for future conferences, including: a) allowing
submitting code and appendices one week after the deadline, and b) measuring
dataset reproducibility by a checklist of data collection practices.
|
2306.09579 | Xiaosong Wang | Dequan Wang, Xiaosong Wang, Lilong Wang, Mengzhang Li, Qian Da,
Xiaoqiang Liu, Xiangyu Gao, Jun Shen, Junjun He, Tian Shen, Qi Duan, Jie
Zhao, Kang Li, Yu Qiao, Shaoting Zhang | MedFMC: A Real-world Dataset and Benchmark For Foundation Model
Adaptation in Medical Image Classification | Preprint. Under review | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Foundation models, often pre-trained with large-scale data, have achieved
paramount success in jump-starting various vision and language applications.
Recent advances further enable adapting foundation models in downstream tasks
efficiently using only a few training samples, e.g., in-context learning. Yet,
the application of such learning paradigms in medical image analysis remains
scarce due to the shortage of publicly accessible data and benchmarks. In this
paper, we aim at approaches adapting the foundation models for medical image
classification and present a novel dataset and benchmark for the evaluation,
i.e., examining the overall performance of accommodating the large-scale
foundation models downstream on a set of diverse real-world clinical tasks. We
collect five sets of medical imaging data from multiple institutes targeting a
variety of real-world clinical tasks (22,349 images in total), i.e., thoracic
diseases screening in X-rays, pathological lesion tissue screening, lesion
detection in endoscopy images, neonatal jaundice evaluation, and diabetic
retinopathy grading. Results of multiple baseline methods are demonstrated
using the proposed dataset from both accuracy and cost-effective perspectives.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 01:46:07 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Wang",
"Dequan",
""
],
[
"Wang",
"Xiaosong",
""
],
[
"Wang",
"Lilong",
""
],
[
"Li",
"Mengzhang",
""
],
[
"Da",
"Qian",
""
],
[
"Liu",
"Xiaoqiang",
""
],
[
"Gao",
"Xiangyu",
""
],
[
"Shen",
"Jun",
""
],
[
"He",
"Junjun",
""
],
[
"Shen",
"Tian",
""
],
[
"Duan",
"Qi",
""
],
[
"Zhao",
"Jie",
""
],
[
"Li",
"Kang",
""
],
[
"Qiao",
"Yu",
""
],
[
"Zhang",
"Shaoting",
""
]
] | TITLE: MedFMC: A Real-world Dataset and Benchmark For Foundation Model
Adaptation in Medical Image Classification
ABSTRACT: Foundation models, often pre-trained with large-scale data, have achieved
paramount success in jump-starting various vision and language applications.
Recent advances further enable adapting foundation models in downstream tasks
efficiently using only a few training samples, e.g., in-context learning. Yet,
the application of such learning paradigms in medical image analysis remains
scarce due to the shortage of publicly accessible data and benchmarks. In this
paper, we aim at approaches adapting the foundation models for medical image
classification and present a novel dataset and benchmark for the evaluation,
i.e., examining the overall performance of accommodating the large-scale
foundation models downstream on a set of diverse real-world clinical tasks. We
collect five sets of medical imaging data from multiple institutes targeting a
variety of real-world clinical tasks (22,349 images in total), i.e., thoracic
diseases screening in X-rays, pathological lesion tissue screening, lesion
detection in endoscopy images, neonatal jaundice evaluation, and diabetic
retinopathy grading. Results of multiple baseline methods are demonstrated
using the proposed dataset from both accuracy and cost-effective perspectives.
|
2306.09592 | Yang Li | Rui Zhang, Ziqi Wang, Yang Li, Jiabao Wang, Zhiteng Wang | FewSAR: A Few-shot SAR Image Classification Benchmark | 7 pages, 4 figures | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Few-shot learning (FSL) is one of the significant and hard problems in the
field of image classification. However, in contrast to the rapid development of
the visible light dataset, the progress in SAR target image classification is
much slower. The lack of unified benchmark is a key reason for this phenomenon,
which may be severely overlooked by the current literature. The researchers of
SAR target image classification always report their new results on their own
datasets and experimental setup. It leads to inefficiency in result comparison
and impedes the further progress of this area. Motivated by this observation,
we propose a novel few-shot SAR image classification benchmark (FewSAR) to
address this issue. FewSAR consists of an open-source Python code library of 15
classic methods in three categories for few-shot SAR image classification. It
provides an accessible and customizable testbed for different few-shot SAR
image classification task. To further understanding the performance of
different few-shot methods, we establish evaluation protocols and conduct
extensive experiments within the benchmark. By analyzing the quantitative
results and runtime under the same setting, we observe that the accuracy of
metric learning methods can achieve the best results. Meta-learning methods and
fine-tuning methods perform poorly on few-shot SAR images, which is primarily
due to the bias of existing datasets. We believe that FewSAR will open up a new
avenue for future research and development, on real-world challenges at the
intersection of SAR image classification and few-shot deep learning. We will
provide our code for the proposed FewSAR at https://github.com/solarlee/FewSAR.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 02:35:00 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Zhang",
"Rui",
""
],
[
"Wang",
"Ziqi",
""
],
[
"Li",
"Yang",
""
],
[
"Wang",
"Jiabao",
""
],
[
"Wang",
"Zhiteng",
""
]
] | TITLE: FewSAR: A Few-shot SAR Image Classification Benchmark
ABSTRACT: Few-shot learning (FSL) is one of the significant and hard problems in the
field of image classification. However, in contrast to the rapid development of
the visible light dataset, the progress in SAR target image classification is
much slower. The lack of unified benchmark is a key reason for this phenomenon,
which may be severely overlooked by the current literature. The researchers of
SAR target image classification always report their new results on their own
datasets and experimental setup. It leads to inefficiency in result comparison
and impedes the further progress of this area. Motivated by this observation,
we propose a novel few-shot SAR image classification benchmark (FewSAR) to
address this issue. FewSAR consists of an open-source Python code library of 15
classic methods in three categories for few-shot SAR image classification. It
provides an accessible and customizable testbed for different few-shot SAR
image classification task. To further understanding the performance of
different few-shot methods, we establish evaluation protocols and conduct
extensive experiments within the benchmark. By analyzing the quantitative
results and runtime under the same setting, we observe that the accuracy of
metric learning methods can achieve the best results. Meta-learning methods and
fine-tuning methods perform poorly on few-shot SAR images, which is primarily
due to the bias of existing datasets. We believe that FewSAR will open up a new
avenue for future research and development, on real-world challenges at the
intersection of SAR image classification and few-shot deep learning. We will
provide our code for the proposed FewSAR at https://github.com/solarlee/FewSAR.
|
2306.09593 | Guangtao Lyu | Guangtao Lyu, Kun Liu, Anna Zhu, Seiichi Uchida, Brian Kenji Iwana | FETNet: Feature Erasing and Transferring Network for Scene Text Removal | Accepted by Pattern Recognition 2023 | Pattern Recognition 2023 | 10.1016/j.patcog.2023.109531 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The scene text removal (STR) task aims to remove text regions and recover the
background smoothly in images for private information protection. Most existing
STR methods adopt encoder-decoder-based CNNs, with direct copies of the
features in the skip connections. However, the encoded features contain both
text texture and structure information. The insufficient utilization of text
features hampers the performance of background reconstruction in text removal
regions. To tackle these problems, we propose a novel Feature Erasing and
Transferring (FET) mechanism to reconfigure the encoded features for STR in
this paper. In FET, a Feature Erasing Module (FEM) is designed to erase text
features. An attention module is responsible for generating the feature
similarity guidance. The Feature Transferring Module (FTM) is introduced to
transfer the corresponding features in different layers based on the attention
guidance. With this mechanism, a one-stage, end-to-end trainable network called
FETNet is constructed for scene text removal. In addition, to facilitate
research on both scene text removal and segmentation tasks, we introduce a
novel dataset, Flickr-ST, with multi-category annotations. A sufficient number
of experiments and ablation studies are conducted on the public datasets and
Flickr-ST. Our proposed method achieves state-of-the-art performance using most
metrics, with remarkably higher quality scene text removal results. The source
code of our work is available at:
\href{https://github.com/GuangtaoLyu/FETNet}{https://github.com/GuangtaoLyu/FETNet.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 02:38:30 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Lyu",
"Guangtao",
""
],
[
"Liu",
"Kun",
""
],
[
"Zhu",
"Anna",
""
],
[
"Uchida",
"Seiichi",
""
],
[
"Iwana",
"Brian Kenji",
""
]
] | TITLE: FETNet: Feature Erasing and Transferring Network for Scene Text Removal
ABSTRACT: The scene text removal (STR) task aims to remove text regions and recover the
background smoothly in images for private information protection. Most existing
STR methods adopt encoder-decoder-based CNNs, with direct copies of the
features in the skip connections. However, the encoded features contain both
text texture and structure information. The insufficient utilization of text
features hampers the performance of background reconstruction in text removal
regions. To tackle these problems, we propose a novel Feature Erasing and
Transferring (FET) mechanism to reconfigure the encoded features for STR in
this paper. In FET, a Feature Erasing Module (FEM) is designed to erase text
features. An attention module is responsible for generating the feature
similarity guidance. The Feature Transferring Module (FTM) is introduced to
transfer the corresponding features in different layers based on the attention
guidance. With this mechanism, a one-stage, end-to-end trainable network called
FETNet is constructed for scene text removal. In addition, to facilitate
research on both scene text removal and segmentation tasks, we introduce a
novel dataset, Flickr-ST, with multi-category annotations. A sufficient number
of experiments and ablation studies are conducted on the public datasets and
Flickr-ST. Our proposed method achieves state-of-the-art performance using most
metrics, with remarkably higher quality scene text removal results. The source
code of our work is available at:
\href{https://github.com/GuangtaoLyu/FETNet}{https://github.com/GuangtaoLyu/FETNet.
|
2306.09612 | Wen-Zhi Li | Wen-Zhi Li, Chang-Dong Wang, Hui Xiong, Jian-Huang Lai | GraphSHA: Synthesizing Harder Samples for Class-Imbalanced Node
Classification | Accepted to KDD 2023 Research Track | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Class imbalance is the phenomenon that some classes have much fewer instances
than others, which is ubiquitous in real-world graph-structured scenarios.
Recent studies find that off-the-shelf Graph Neural Networks (GNNs) would
under-represent minor class samples. We investigate this phenomenon and
discover that the subspaces of minor classes being squeezed by those of the
major ones in the latent space is the main cause of this failure. We are
naturally inspired to enlarge the decision boundaries of minor classes and
propose a general framework GraphSHA by Synthesizing HArder minor samples.
Furthermore, to avoid the enlarged minor boundary violating the subspaces of
neighbor classes, we also propose a module called SemiMixup to transmit
enlarged boundary information to the interior of the minor classes while
blocking information propagation from minor classes to neighbor classes.
Empirically, GraphSHA shows its effectiveness in enlarging the decision
boundaries of minor classes, as it outperforms various baseline methods in
class-imbalanced node classification with different GNN backbone encoders over
seven public benchmark datasets. Code is avilable at
https://github.com/wenzhilics/GraphSHA.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 04:05:58 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Li",
"Wen-Zhi",
""
],
[
"Wang",
"Chang-Dong",
""
],
[
"Xiong",
"Hui",
""
],
[
"Lai",
"Jian-Huang",
""
]
] | TITLE: GraphSHA: Synthesizing Harder Samples for Class-Imbalanced Node
Classification
ABSTRACT: Class imbalance is the phenomenon that some classes have much fewer instances
than others, which is ubiquitous in real-world graph-structured scenarios.
Recent studies find that off-the-shelf Graph Neural Networks (GNNs) would
under-represent minor class samples. We investigate this phenomenon and
discover that the subspaces of minor classes being squeezed by those of the
major ones in the latent space is the main cause of this failure. We are
naturally inspired to enlarge the decision boundaries of minor classes and
propose a general framework GraphSHA by Synthesizing HArder minor samples.
Furthermore, to avoid the enlarged minor boundary violating the subspaces of
neighbor classes, we also propose a module called SemiMixup to transmit
enlarged boundary information to the interior of the minor classes while
blocking information propagation from minor classes to neighbor classes.
Empirically, GraphSHA shows its effectiveness in enlarging the decision
boundaries of minor classes, as it outperforms various baseline methods in
class-imbalanced node classification with different GNN backbone encoders over
seven public benchmark datasets. Code is avilable at
https://github.com/wenzhilics/GraphSHA.
|
2306.09613 | Pha Nguyen | Pha Nguyen, Kha Gia Quach, John Gauch, Samee U. Khan, Bhiksha Raj,
Khoa Luu | UTOPIA: Unconstrained Tracking Objects without Preliminary Examination
via Cross-Domain Adaptation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multiple Object Tracking (MOT) aims to find bounding boxes and identities of
targeted objects in consecutive video frames. While fully-supervised MOT
methods have achieved high accuracy on existing datasets, they cannot
generalize well on a newly obtained dataset or a new unseen domain. In this
work, we first address the MOT problem from the cross-domain point of view,
imitating the process of new data acquisition in practice. Then, a new
cross-domain MOT adaptation from existing datasets is proposed without any
pre-defined human knowledge in understanding and modeling objects. It can also
learn and update itself from the target data feedback. The intensive
experiments are designed on four challenging settings, including MOTSynth to
MOT17, MOT17 to MOT20, MOT17 to VisDrone, and MOT17 to DanceTrack. We then
prove the adaptability of the proposed self-supervised learning strategy. The
experiments also show superior performance on tracking metrics MOTA and IDF1,
compared to fully supervised, unsupervised, and self-supervised
state-of-the-art methods.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 04:06:15 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Nguyen",
"Pha",
""
],
[
"Quach",
"Kha Gia",
""
],
[
"Gauch",
"John",
""
],
[
"Khan",
"Samee U.",
""
],
[
"Raj",
"Bhiksha",
""
],
[
"Luu",
"Khoa",
""
]
] | TITLE: UTOPIA: Unconstrained Tracking Objects without Preliminary Examination
via Cross-Domain Adaptation
ABSTRACT: Multiple Object Tracking (MOT) aims to find bounding boxes and identities of
targeted objects in consecutive video frames. While fully-supervised MOT
methods have achieved high accuracy on existing datasets, they cannot
generalize well on a newly obtained dataset or a new unseen domain. In this
work, we first address the MOT problem from the cross-domain point of view,
imitating the process of new data acquisition in practice. Then, a new
cross-domain MOT adaptation from existing datasets is proposed without any
pre-defined human knowledge in understanding and modeling objects. It can also
learn and update itself from the target data feedback. The intensive
experiments are designed on four challenging settings, including MOTSynth to
MOT17, MOT17 to MOT20, MOT17 to VisDrone, and MOT17 to DanceTrack. We then
prove the adaptability of the proposed self-supervised learning strategy. The
experiments also show superior performance on tracking metrics MOTA and IDF1,
compared to fully supervised, unsupervised, and self-supervised
state-of-the-art methods.
|
2306.09614 | Wen-Zhi Li | Wen-Zhi Li, Chang-Dong Wang, Hui Xiong, Jian-Huang Lai | HomoGCL: Rethinking Homophily in Graph Contrastive Learning | Accepted to KDD 2023 Research Track | null | null | null | cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Contrastive learning (CL) has become the de-facto learning paradigm in
self-supervised learning on graphs, which generally follows the
"augmenting-contrasting" learning scheme. However, we observe that unlike CL in
computer vision domain, CL in graph domain performs decently even without
augmentation. We conduct a systematic analysis of this phenomenon and argue
that homophily, i.e., the principle that "like attracts like", plays a key role
in the success of graph CL. Inspired to leverage this property explicitly, we
propose HomoGCL, a model-agnostic framework to expand the positive set using
neighbor nodes with neighbor-specific significances. Theoretically, HomoGCL
introduces a stricter lower bound of the mutual information between raw node
features and node embeddings in augmented views. Furthermore, HomoGCL can be
combined with existing graph CL models in a plug-and-play way with light extra
computational overhead. Extensive experiments demonstrate that HomoGCL yields
multiple state-of-the-art results across six public datasets and consistently
brings notable performance improvements when applied to various graph CL
methods. Code is avilable at https://github.com/wenzhilics/HomoGCL.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 04:06:52 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Li",
"Wen-Zhi",
""
],
[
"Wang",
"Chang-Dong",
""
],
[
"Xiong",
"Hui",
""
],
[
"Lai",
"Jian-Huang",
""
]
] | TITLE: HomoGCL: Rethinking Homophily in Graph Contrastive Learning
ABSTRACT: Contrastive learning (CL) has become the de-facto learning paradigm in
self-supervised learning on graphs, which generally follows the
"augmenting-contrasting" learning scheme. However, we observe that unlike CL in
computer vision domain, CL in graph domain performs decently even without
augmentation. We conduct a systematic analysis of this phenomenon and argue
that homophily, i.e., the principle that "like attracts like", plays a key role
in the success of graph CL. Inspired to leverage this property explicitly, we
propose HomoGCL, a model-agnostic framework to expand the positive set using
neighbor nodes with neighbor-specific significances. Theoretically, HomoGCL
introduces a stricter lower bound of the mutual information between raw node
features and node embeddings in augmented views. Furthermore, HomoGCL can be
combined with existing graph CL models in a plug-and-play way with light extra
computational overhead. Extensive experiments demonstrate that HomoGCL yields
multiple state-of-the-art results across six public datasets and consistently
brings notable performance improvements when applied to various graph CL
methods. Code is avilable at https://github.com/wenzhilics/HomoGCL.
|
2306.09615 | Yaqi Zhang | Yaqi Zhang, Yan Lu, Bin Liu, Zhiwei Zhao, Qi Chu, Nenghai Yu | EVOPOSE: A Recursive Transformer For 3D Human Pose Estimation With
Kinematic Structure Priors | 5 pages, 2 figures, 4 tables, published in the proceedings of IEEE
ICASSP 2023 | null | 10.1109/ICASSP49357.2023.10095302 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transformer is popular in recent 3D human pose estimation, which utilizes
long-term modeling to lift 2D keypoints into the 3D space. However, current
transformer-based methods do not fully exploit the prior knowledge of the human
skeleton provided by the kinematic structure. In this paper, we propose a novel
transformer-based model EvoPose to introduce the human body prior knowledge for
3D human pose estimation effectively. Specifically, a Structural Priors
Representation (SPR) module represents human priors as structural features
carrying rich body patterns, e.g. joint relationships. The structural features
are interacted with 2D pose sequences and help the model to achieve more
informative spatiotemporal features. Moreover, a Recursive Refinement (RR)
module is applied to refine the 3D pose outputs by utilizing estimated results
and further injects human priors simultaneously. Extensive experiments
demonstrate the effectiveness of EvoPose which achieves a new state of the art
on two most popular benchmarks, Human3.6M and MPI-INF-3DHP.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 04:09:16 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Zhang",
"Yaqi",
""
],
[
"Lu",
"Yan",
""
],
[
"Liu",
"Bin",
""
],
[
"Zhao",
"Zhiwei",
""
],
[
"Chu",
"Qi",
""
],
[
"Yu",
"Nenghai",
""
]
] | TITLE: EVOPOSE: A Recursive Transformer For 3D Human Pose Estimation With
Kinematic Structure Priors
ABSTRACT: Transformer is popular in recent 3D human pose estimation, which utilizes
long-term modeling to lift 2D keypoints into the 3D space. However, current
transformer-based methods do not fully exploit the prior knowledge of the human
skeleton provided by the kinematic structure. In this paper, we propose a novel
transformer-based model EvoPose to introduce the human body prior knowledge for
3D human pose estimation effectively. Specifically, a Structural Priors
Representation (SPR) module represents human priors as structural features
carrying rich body patterns, e.g. joint relationships. The structural features
are interacted with 2D pose sequences and help the model to achieve more
informative spatiotemporal features. Moreover, a Recursive Refinement (RR)
module is applied to refine the 3D pose outputs by utilizing estimated results
and further injects human priors simultaneously. Extensive experiments
demonstrate the effectiveness of EvoPose which achieves a new state of the art
on two most popular benchmarks, Human3.6M and MPI-INF-3DHP.
|
2306.09631 | Junwei Bao Doctor | Yu Lu, Junwei Bao, Zichen Ma, Xiaoguang Han, Youzheng Wu, Shuguang
Cui, Xiaodong He | AUGUST: an Automatic Generation Understudy for Synthesizing
Conversational Recommendation Datasets | 10 pages | null | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | High-quality data is essential for conversational recommendation systems and
serves as the cornerstone of the network architecture development and training
strategy design. Existing works contribute heavy human efforts to manually
labeling or designing and extending recommender dialogue templates. However,
they suffer from (i) the limited number of human annotators results in that
datasets can hardly capture rich and large-scale cases in the real world, (ii)
the limited experience and knowledge of annotators account for the
uninformative corpus and inappropriate recommendations. In this paper, we
propose a novel automatic dataset synthesis approach that can generate both
large-scale and high-quality recommendation dialogues through a data2text
generation process, where unstructured recommendation conversations are
generated from structured graphs based on user-item information from the real
world. In doing so, we comprehensively exploit: (i) rich personalized user
profiles from traditional recommendation datasets, (ii) rich external knowledge
from knowledge graphs, and (iii) the conversation ability contained in
human-to-human conversational recommendation datasets. Extensive experiments
validate the benefit brought by the automatically synthesized data under
low-resource scenarios and demonstrate the promising potential to facilitate
the development of a more effective conversational recommendation system.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 05:27:14 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Lu",
"Yu",
""
],
[
"Bao",
"Junwei",
""
],
[
"Ma",
"Zichen",
""
],
[
"Han",
"Xiaoguang",
""
],
[
"Wu",
"Youzheng",
""
],
[
"Cui",
"Shuguang",
""
],
[
"He",
"Xiaodong",
""
]
] | TITLE: AUGUST: an Automatic Generation Understudy for Synthesizing
Conversational Recommendation Datasets
ABSTRACT: High-quality data is essential for conversational recommendation systems and
serves as the cornerstone of the network architecture development and training
strategy design. Existing works contribute heavy human efforts to manually
labeling or designing and extending recommender dialogue templates. However,
they suffer from (i) the limited number of human annotators results in that
datasets can hardly capture rich and large-scale cases in the real world, (ii)
the limited experience and knowledge of annotators account for the
uninformative corpus and inappropriate recommendations. In this paper, we
propose a novel automatic dataset synthesis approach that can generate both
large-scale and high-quality recommendation dialogues through a data2text
generation process, where unstructured recommendation conversations are
generated from structured graphs based on user-item information from the real
world. In doing so, we comprehensively exploit: (i) rich personalized user
profiles from traditional recommendation datasets, (ii) rich external knowledge
from knowledge graphs, and (iii) the conversation ability contained in
human-to-human conversational recommendation datasets. Extensive experiments
validate the benefit brought by the automatically synthesized data under
low-resource scenarios and demonstrate the promising potential to facilitate
the development of a more effective conversational recommendation system.
|
2306.09643 | Phillip Lippe | Phillip Lippe, Sara Magliacane, Sindy L\"owe, Yuki M. Asano, Taco
Cohen, Efstratios Gavves | BISCUIT: Causal Representation Learning from Binary Interactions | Published in: Uncertainty in Artificial Intelligence (UAI 2023).
Project page: https://phlippe.github.io/BISCUIT/ | null | null | null | cs.LG cs.AI stat.ME | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Identifying the causal variables of an environment and how to intervene on
them is of core value in applications such as robotics and embodied AI. While
an agent can commonly interact with the environment and may implicitly perturb
the behavior of some of these causal variables, often the targets it affects
remain unknown. In this paper, we show that causal variables can still be
identified for many common setups, e.g., additive Gaussian noise models, if the
agent's interactions with a causal variable can be described by an unknown
binary variable. This happens when each causal variable has two different
mechanisms, e.g., an observational and an interventional one. Using this
identifiability result, we propose BISCUIT, a method for simultaneously
learning causal variables and their corresponding binary interaction variables.
On three robotic-inspired datasets, BISCUIT accurately identifies causal
variables and can even be scaled to complex, realistic environments for
embodied AI.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 06:10:55 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Lippe",
"Phillip",
""
],
[
"Magliacane",
"Sara",
""
],
[
"Löwe",
"Sindy",
""
],
[
"Asano",
"Yuki M.",
""
],
[
"Cohen",
"Taco",
""
],
[
"Gavves",
"Efstratios",
""
]
] | TITLE: BISCUIT: Causal Representation Learning from Binary Interactions
ABSTRACT: Identifying the causal variables of an environment and how to intervene on
them is of core value in applications such as robotics and embodied AI. While
an agent can commonly interact with the environment and may implicitly perturb
the behavior of some of these causal variables, often the targets it affects
remain unknown. In this paper, we show that causal variables can still be
identified for many common setups, e.g., additive Gaussian noise models, if the
agent's interactions with a causal variable can be described by an unknown
binary variable. This happens when each causal variable has two different
mechanisms, e.g., an observational and an interventional one. Using this
identifiability result, we propose BISCUIT, a method for simultaneously
learning causal variables and their corresponding binary interaction variables.
On three robotic-inspired datasets, BISCUIT accurately identifies causal
variables and can even be scaled to complex, realistic environments for
embodied AI.
|
2306.09646 | Misha Glazunov | Misha Glazunov, Apostolis Zarras | Vacant Holes for Unsupervised Detection of the Outliers in Compact
Latent Representation | Accepted for the 39th Conference on Uncertainty in Artificial
Intelligence (UAI 2023) | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Detection of the outliers is pivotal for any machine learning model deployed
and operated in real-world. It is essential for the Deep Neural Networks that
were shown to be overconfident with such inputs. Moreover, even deep generative
models that allow estimation of the probability density of the input fail in
achieving this task. In this work, we concentrate on the specific type of these
models: Variational Autoencoders (VAEs). First, we unveil a significant
theoretical flaw in the assumption of the classical VAE model. Second, we
enforce an accommodating topological property to the image of the deep neural
mapping to the latent space: compactness to alleviate the flaw and obtain the
means to provably bound the image within the determined limits by squeezing
both inliers and outliers together. We enforce compactness using two
approaches: (i) Alexandroff extension and (ii) fixed Lipschitz continuity
constant on the mapping of the encoder of the VAEs. Finally and most
importantly, we discover that the anomalous inputs predominantly tend to land
on the vacant latent holes within the compact space, enabling their successful
identification. For that reason, we introduce a specifically devised score for
hole detection and evaluate the solution against several baseline benchmarks
achieving promising results.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 06:21:48 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Glazunov",
"Misha",
""
],
[
"Zarras",
"Apostolis",
""
]
] | TITLE: Vacant Holes for Unsupervised Detection of the Outliers in Compact
Latent Representation
ABSTRACT: Detection of the outliers is pivotal for any machine learning model deployed
and operated in real-world. It is essential for the Deep Neural Networks that
were shown to be overconfident with such inputs. Moreover, even deep generative
models that allow estimation of the probability density of the input fail in
achieving this task. In this work, we concentrate on the specific type of these
models: Variational Autoencoders (VAEs). First, we unveil a significant
theoretical flaw in the assumption of the classical VAE model. Second, we
enforce an accommodating topological property to the image of the deep neural
mapping to the latent space: compactness to alleviate the flaw and obtain the
means to provably bound the image within the determined limits by squeezing
both inliers and outliers together. We enforce compactness using two
approaches: (i) Alexandroff extension and (ii) fixed Lipschitz continuity
constant on the mapping of the encoder of the VAEs. Finally and most
importantly, we discover that the anomalous inputs predominantly tend to land
on the vacant latent holes within the compact space, enabling their successful
identification. For that reason, we introduce a specifically devised score for
hole detection and evaluate the solution against several baseline benchmarks
achieving promising results.
|
2306.09665 | Ali Khatami | Ali Khatami, Andy Zaidman | State-Of-The-Practice in Quality Assurance in Java-Based Open Source
Software Development | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | To ensure the quality of software systems, software engineers can make use of
a variety of quality assurance approaches, such as software testing, modern
code review, automated static analysis, and build automation. Each of these
quality assurance practices has been studied in depth in isolation, but there
is a clear knowledge gap when it comes to our understanding of how these
approaches are being used in conjunction or not. In our study, we broadly
investigate whether and how these quality assurance approaches are being used
in conjunction in the development of 1,454 popular open source software
projects on GitHub. Our study indicates that typically projects do not follow
all quality assurance practices together with high intensity. In fact, we only
observe weak correlation among some quality assurance practices. In general,
our study provides a deeper understanding of how existing quality assurance
approaches are currently being used in Java-based open source software
development. Besides, we specifically zoomed in on the more mature projects in
our dataset, and generally, we observe that more mature projects are more
intense in their application of the quality assurance practices, with more
focus on their ASAT usage and code reviewing, but no strong change in their CI
usage.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 07:43:11 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Khatami",
"Ali",
""
],
[
"Zaidman",
"Andy",
""
]
] | TITLE: State-Of-The-Practice in Quality Assurance in Java-Based Open Source
Software Development
ABSTRACT: To ensure the quality of software systems, software engineers can make use of
a variety of quality assurance approaches, such as software testing, modern
code review, automated static analysis, and build automation. Each of these
quality assurance practices has been studied in depth in isolation, but there
is a clear knowledge gap when it comes to our understanding of how these
approaches are being used in conjunction or not. In our study, we broadly
investigate whether and how these quality assurance approaches are being used
in conjunction in the development of 1,454 popular open source software
projects on GitHub. Our study indicates that typically projects do not follow
all quality assurance practices together with high intensity. In fact, we only
observe weak correlation among some quality assurance practices. In general,
our study provides a deeper understanding of how existing quality assurance
approaches are currently being used in Java-based open source software
development. Besides, we specifically zoomed in on the more mature projects in
our dataset, and generally, we observe that more mature projects are more
intense in their application of the quality assurance practices, with more
focus on their ASAT usage and code reviewing, but no strong change in their CI
usage.
|
2306.09687 | Nguyen Phi | Phi Nguyen Van, Hieu Pham Huy, Long Tran Quoc | Echocardiography Segmentation Using Neural ODE-based Diffeomorphic
Registration Field | Submitted to IEEE TMI in June 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Convolutional neural networks (CNNs) have recently proven their excellent
ability to segment 2D cardiac ultrasound images. However, the majority of
attempts to perform full-sequence segmentation of cardiac ultrasound videos
either rely on models trained only on keyframe images or fail to maintain the
topology over time. To address these issues, in this work, we consider
segmentation of ultrasound video as a registration estimation problem and
present a novel method for diffeomorphic image registration using neural
ordinary differential equations (Neural ODE). In particular, we consider the
registration field vector field between frames as a continuous trajectory ODE.
The estimated registration field is then applied to the segmentation mask of
the first frame to obtain a segment for the whole cardiac cycle. The proposed
method, Echo-ODE, introduces several key improvements compared to the previous
state-of-the-art. Firstly, by solving a continuous ODE, the proposed method
achieves smoother segmentation, preserving the topology of segmentation maps
over the whole sequence (Hausdorff distance: 3.7-4.4). Secondly, it maintains
temporal consistency between frames without explicitly optimizing for temporal
consistency attributes, achieving temporal consistency in 91% of the videos in
the dataset. Lastly, the proposed method is able to maintain the clinical
accuracy of the segmentation maps (MAE of the LVEF: 2.7-3.1). The results show
that our method surpasses the previous state-of-the-art in multiple aspects,
demonstrating the importance of spatial-temporal data processing for the
implementation of Neural ODEs in medical imaging applications. These findings
open up new research directions for solving echocardiography segmentation
tasks.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 08:37:27 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Van",
"Phi Nguyen",
""
],
[
"Huy",
"Hieu Pham",
""
],
[
"Quoc",
"Long Tran",
""
]
] | TITLE: Echocardiography Segmentation Using Neural ODE-based Diffeomorphic
Registration Field
ABSTRACT: Convolutional neural networks (CNNs) have recently proven their excellent
ability to segment 2D cardiac ultrasound images. However, the majority of
attempts to perform full-sequence segmentation of cardiac ultrasound videos
either rely on models trained only on keyframe images or fail to maintain the
topology over time. To address these issues, in this work, we consider
segmentation of ultrasound video as a registration estimation problem and
present a novel method for diffeomorphic image registration using neural
ordinary differential equations (Neural ODE). In particular, we consider the
registration field vector field between frames as a continuous trajectory ODE.
The estimated registration field is then applied to the segmentation mask of
the first frame to obtain a segment for the whole cardiac cycle. The proposed
method, Echo-ODE, introduces several key improvements compared to the previous
state-of-the-art. Firstly, by solving a continuous ODE, the proposed method
achieves smoother segmentation, preserving the topology of segmentation maps
over the whole sequence (Hausdorff distance: 3.7-4.4). Secondly, it maintains
temporal consistency between frames without explicitly optimizing for temporal
consistency attributes, achieving temporal consistency in 91% of the videos in
the dataset. Lastly, the proposed method is able to maintain the clinical
accuracy of the segmentation maps (MAE of the LVEF: 2.7-3.1). The results show
that our method surpasses the previous state-of-the-art in multiple aspects,
demonstrating the importance of spatial-temporal data processing for the
implementation of Neural ODEs in medical imaging applications. These findings
open up new research directions for solving echocardiography segmentation
tasks.
|
2306.09697 | Qingyu Tan | Qingyu Tan, Lu Xu, Lidong Bing, Hwee Tou Ng | Class-Adaptive Self-Training for Relation Extraction with Incompletely
Annotated Training Data | ACL 2023 Findings | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Relation extraction (RE) aims to extract relations from sentences and
documents. Existing relation extraction models typically rely on supervised
machine learning. However, recent studies showed that many RE datasets are
incompletely annotated. This is known as the false negative problem in which
valid relations are falsely annotated as 'no_relation'. Models trained with
such data inevitably make similar mistakes during the inference stage.
Self-training has been proven effective in alleviating the false negative
problem. However, traditional self-training is vulnerable to confirmation bias
and exhibits poor performance in minority classes. To overcome this limitation,
we proposed a novel class-adaptive re-sampling self-training framework.
Specifically, we re-sampled the pseudo-labels for each class by precision and
recall scores. Our re-sampling strategy favored the pseudo-labels of classes
with high precision and low recall, which improved the overall recall without
significantly compromising precision. We conducted experiments on
document-level and biomedical relation extraction datasets, and the results
showed that our proposed self-training framework consistently outperforms
existing competitive methods on the Re-DocRED and ChemDisgene datasets when the
training data are incompletely annotated. Our code is released at
https://github.com/DAMO-NLP-SG/CAST.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 09:01:45 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Tan",
"Qingyu",
""
],
[
"Xu",
"Lu",
""
],
[
"Bing",
"Lidong",
""
],
[
"Ng",
"Hwee Tou",
""
]
] | TITLE: Class-Adaptive Self-Training for Relation Extraction with Incompletely
Annotated Training Data
ABSTRACT: Relation extraction (RE) aims to extract relations from sentences and
documents. Existing relation extraction models typically rely on supervised
machine learning. However, recent studies showed that many RE datasets are
incompletely annotated. This is known as the false negative problem in which
valid relations are falsely annotated as 'no_relation'. Models trained with
such data inevitably make similar mistakes during the inference stage.
Self-training has been proven effective in alleviating the false negative
problem. However, traditional self-training is vulnerable to confirmation bias
and exhibits poor performance in minority classes. To overcome this limitation,
we proposed a novel class-adaptive re-sampling self-training framework.
Specifically, we re-sampled the pseudo-labels for each class by precision and
recall scores. Our re-sampling strategy favored the pseudo-labels of classes
with high precision and low recall, which improved the overall recall without
significantly compromising precision. We conducted experiments on
document-level and biomedical relation extraction datasets, and the results
showed that our proposed self-training framework consistently outperforms
existing competitive methods on the Re-DocRED and ChemDisgene datasets when the
training data are incompletely annotated. Our code is released at
https://github.com/DAMO-NLP-SG/CAST.
|
2306.09718 | Mengdi Gao | Hongyang Jiang, Mengdi Gao, Yan Hu, Qiushi Ren, Zhaoheng Xie, Jiang
Liu | Label-noise-tolerant medical image classification via self-attention and
self-supervised learning | 11pages, 8 figures | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Deep neural networks (DNNs) have been widely applied in medical image
classification and achieve remarkable classification performance. These
achievements heavily depend on large-scale accurately annotated training data.
However, label noise is inevitably introduced in the medical image annotation,
as the labeling process heavily relies on the expertise and experience of
annotators. Meanwhile, DNNs suffer from overfitting noisy labels, degrading the
performance of models. Therefore, in this work, we innovatively devise
noise-robust training approach to mitigate the adverse effects of noisy labels
in medical image classification. Specifically, we incorporate contrastive
learning and intra-group attention mixup strategies into the vanilla supervised
learning. The contrastive learning for feature extractor helps to enhance
visual representation of DNNs. The intra-group attention mixup module
constructs groups and assigns self-attention weights for group-wise samples,
and subsequently interpolates massive noisy-suppressed samples through weighted
mixup operation. We conduct comparative experiments on both synthetic and
real-world noisy medical datasets under various noise levels. Rigorous
experiments validate that our noise-robust method with contrastive learning and
attention mixup can effectively handle with label noise, and is superior to
state-of-the-art methods. An ablation study also shows that both components
contribute to boost model performance. The proposed method demonstrates its
capability of curb label noise and has certain potential toward real-world
clinic applications.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 09:37:16 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Jiang",
"Hongyang",
""
],
[
"Gao",
"Mengdi",
""
],
[
"Hu",
"Yan",
""
],
[
"Ren",
"Qiushi",
""
],
[
"Xie",
"Zhaoheng",
""
],
[
"Liu",
"Jiang",
""
]
] | TITLE: Label-noise-tolerant medical image classification via self-attention and
self-supervised learning
ABSTRACT: Deep neural networks (DNNs) have been widely applied in medical image
classification and achieve remarkable classification performance. These
achievements heavily depend on large-scale accurately annotated training data.
However, label noise is inevitably introduced in the medical image annotation,
as the labeling process heavily relies on the expertise and experience of
annotators. Meanwhile, DNNs suffer from overfitting noisy labels, degrading the
performance of models. Therefore, in this work, we innovatively devise
noise-robust training approach to mitigate the adverse effects of noisy labels
in medical image classification. Specifically, we incorporate contrastive
learning and intra-group attention mixup strategies into the vanilla supervised
learning. The contrastive learning for feature extractor helps to enhance
visual representation of DNNs. The intra-group attention mixup module
constructs groups and assigns self-attention weights for group-wise samples,
and subsequently interpolates massive noisy-suppressed samples through weighted
mixup operation. We conduct comparative experiments on both synthetic and
real-world noisy medical datasets under various noise levels. Rigorous
experiments validate that our noise-robust method with contrastive learning and
attention mixup can effectively handle with label noise, and is superior to
state-of-the-art methods. An ablation study also shows that both components
contribute to boost model performance. The proposed method demonstrates its
capability of curb label noise and has certain potential toward real-world
clinic applications.
|
2306.09752 | Victor Steinborn | Victor Steinborn and Antonis Maronikolakis and Hinrich Sch\"utze | Politeness Stereotypes and Attack Vectors: Gender Stereotypes in
Japanese and Korean Language Models | null | null | null | null | cs.CL cs.AI cs.CY cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In efforts to keep up with the rapid progress and use of large language
models, gender bias research is becoming more prevalent in NLP. Non-English
bias research, however, is still in its infancy with most work focusing on
English. In our work, we study how grammatical gender bias relating to
politeness levels manifests in Japanese and Korean language models. Linguistic
studies in these languages have identified a connection between gender bias and
politeness levels, however it is not yet known if language models reproduce
these biases. We analyze relative prediction probabilities of the male and
female grammatical genders using templates and find that informal polite speech
is most indicative of the female grammatical gender, while rude and formal
speech is most indicative of the male grammatical gender. Further, we find
politeness levels to be an attack vector for allocational gender bias in
cyberbullying detection models. Cyberbullies can evade detection through simple
techniques abusing politeness levels. We introduce an attack dataset to (i)
identify representational gender bias across politeness levels, (ii)
demonstrate how gender biases can be abused to bypass cyberbullying detection
models and (iii) show that allocational biases can be mitigated via training on
our proposed dataset. Through our findings we highlight the importance of bias
research moving beyond its current English-centrism.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 10:36:18 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Steinborn",
"Victor",
""
],
[
"Maronikolakis",
"Antonis",
""
],
[
"Schütze",
"Hinrich",
""
]
] | TITLE: Politeness Stereotypes and Attack Vectors: Gender Stereotypes in
Japanese and Korean Language Models
ABSTRACT: In efforts to keep up with the rapid progress and use of large language
models, gender bias research is becoming more prevalent in NLP. Non-English
bias research, however, is still in its infancy with most work focusing on
English. In our work, we study how grammatical gender bias relating to
politeness levels manifests in Japanese and Korean language models. Linguistic
studies in these languages have identified a connection between gender bias and
politeness levels, however it is not yet known if language models reproduce
these biases. We analyze relative prediction probabilities of the male and
female grammatical genders using templates and find that informal polite speech
is most indicative of the female grammatical gender, while rude and formal
speech is most indicative of the male grammatical gender. Further, we find
politeness levels to be an attack vector for allocational gender bias in
cyberbullying detection models. Cyberbullies can evade detection through simple
techniques abusing politeness levels. We introduce an attack dataset to (i)
identify representational gender bias across politeness levels, (ii)
demonstrate how gender biases can be abused to bypass cyberbullying detection
models and (iii) show that allocational biases can be mitigated via training on
our proposed dataset. Through our findings we highlight the importance of bias
research moving beyond its current English-centrism.
|
2306.09753 | Matteo Busso | Matteo Busso and Xiaoyue Li | A context model for collecting diversity-aware data | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Diversity-aware data are essential for a robust modeling of human behavior in
context. In addition, being the human behavior of interest for numerous
applications, data must also be reusable across domain, to ensure diversity of
interpretations. Current data collection techniques allow only a partial
representation of the diversity of people and often generate data that is
difficult to reuse. To fill this gap, we propose a data collection methodology,
within a hybrid machine-artificial intelligence approach, and its related
dataset, based on a comprehensive ontological notion of context which enables
data reusability. The dataset has a sample of 158 participants and is collected
via the iLog smartphone application. It contains more than 170 GB of subjective
and objective data, which comes from 27 smartphone sensors that are associated
with 168,095 self-reported annotations on the participants context. The dataset
is highly reusable, as demonstrated by its diverse applications.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 10:36:38 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Busso",
"Matteo",
""
],
[
"Li",
"Xiaoyue",
""
]
] | TITLE: A context model for collecting diversity-aware data
ABSTRACT: Diversity-aware data are essential for a robust modeling of human behavior in
context. In addition, being the human behavior of interest for numerous
applications, data must also be reusable across domain, to ensure diversity of
interpretations. Current data collection techniques allow only a partial
representation of the diversity of people and often generate data that is
difficult to reuse. To fill this gap, we propose a data collection methodology,
within a hybrid machine-artificial intelligence approach, and its related
dataset, based on a comprehensive ontological notion of context which enables
data reusability. The dataset has a sample of 158 participants and is collected
via the iLog smartphone application. It contains more than 170 GB of subjective
and objective data, which comes from 27 smartphone sensors that are associated
with 168,095 self-reported annotations on the participants context. The dataset
is highly reusable, as demonstrated by its diverse applications.
|
2306.09762 | Roy Voetman | Roy Voetman and Maya Aghaei and Klaas Dijkstra | The Big Data Myth: Using Diffusion Models for Dataset Generation to
Train Deep Detection Models | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the notable accomplishments of deep object detection models, a major
challenge that persists is the requirement for extensive amounts of training
data. The process of procuring such real-world data is a laborious undertaking,
which has prompted researchers to explore new avenues of research, such as
synthetic data generation techniques. This study presents a framework for the
generation of synthetic datasets by fine-tuning pretrained stable diffusion
models. The synthetic datasets are then manually annotated and employed for
training various object detection models. These detectors are evaluated on a
real-world test set of 331 images and compared against a baseline model that
was trained on real-world images. The results of this study reveal that the
object detection models trained on synthetic data perform similarly to the
baseline model. In the context of apple detection in orchards, the average
precision deviation with the baseline ranges from 0.09 to 0.12. This study
illustrates the potential of synthetic data generation techniques as a viable
alternative to the collection of extensive training data for the training of
deep models.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 10:48:52 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Voetman",
"Roy",
""
],
[
"Aghaei",
"Maya",
""
],
[
"Dijkstra",
"Klaas",
""
]
] | TITLE: The Big Data Myth: Using Diffusion Models for Dataset Generation to
Train Deep Detection Models
ABSTRACT: Despite the notable accomplishments of deep object detection models, a major
challenge that persists is the requirement for extensive amounts of training
data. The process of procuring such real-world data is a laborious undertaking,
which has prompted researchers to explore new avenues of research, such as
synthetic data generation techniques. This study presents a framework for the
generation of synthetic datasets by fine-tuning pretrained stable diffusion
models. The synthetic datasets are then manually annotated and employed for
training various object detection models. These detectors are evaluated on a
real-world test set of 331 images and compared against a baseline model that
was trained on real-world images. The results of this study reveal that the
object detection models trained on synthetic data perform similarly to the
baseline model. In the context of apple detection in orchards, the average
precision deviation with the baseline ranges from 0.09 to 0.12. This study
illustrates the potential of synthetic data generation techniques as a viable
alternative to the collection of extensive training data for the training of
deep models.
|
2306.09789 | Francesco Daghero | Francesco Daghero, Alessio Burrello, Enrico Macii, Paolo Montuschi,
Massimo Poncino and Daniele Jahier Pagliari | Dynamic Decision Tree Ensembles for Energy-Efficient Inference on IoT
Edge Nodes | This article has been accepted for publication in IEEE Internet of
Things Journal | null | 10.1109/JIOT.2023.3286276 | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | With the increasing popularity of Internet of Things (IoT) devices, there is
a growing need for energy-efficient Machine Learning (ML) models that can run
on constrained edge nodes. Decision tree ensembles, such as Random Forests
(RFs) and Gradient Boosting (GBTs), are particularly suited for this task,
given their relatively low complexity compared to other alternatives. However,
their inference time and energy costs are still significant for edge hardware.
Given that said costs grow linearly with the ensemble size, this paper proposes
the use of dynamic ensembles, that adjust the number of executed trees based
both on a latency/energy target and on the complexity of the processed input,
to trade-off computational cost and accuracy. We focus on deploying these
algorithms on multi-core low-power IoT devices, designing a tool that
automatically converts a Python ensemble into optimized C code, and exploring
several optimizations that account for the available parallelism and memory
hierarchy. We extensively benchmark both static and dynamic RFs and GBTs on
three state-of-the-art IoT-relevant datasets, using an 8-core ultra-lowpower
System-on-Chip (SoC), GAP8, as the target platform. Thanks to the proposed
early-stopping mechanisms, we achieve an energy reduction of up to 37.9% with
respect to static GBTs (8.82 uJ vs 14.20 uJ per inference) and 41.7% with
respect to static RFs (2.86 uJ vs 4.90 uJ per inference), without losing
accuracy compared to the static model.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 11:59:18 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Daghero",
"Francesco",
""
],
[
"Burrello",
"Alessio",
""
],
[
"Macii",
"Enrico",
""
],
[
"Montuschi",
"Paolo",
""
],
[
"Poncino",
"Massimo",
""
],
[
"Pagliari",
"Daniele Jahier",
""
]
] | TITLE: Dynamic Decision Tree Ensembles for Energy-Efficient Inference on IoT
Edge Nodes
ABSTRACT: With the increasing popularity of Internet of Things (IoT) devices, there is
a growing need for energy-efficient Machine Learning (ML) models that can run
on constrained edge nodes. Decision tree ensembles, such as Random Forests
(RFs) and Gradient Boosting (GBTs), are particularly suited for this task,
given their relatively low complexity compared to other alternatives. However,
their inference time and energy costs are still significant for edge hardware.
Given that said costs grow linearly with the ensemble size, this paper proposes
the use of dynamic ensembles, that adjust the number of executed trees based
both on a latency/energy target and on the complexity of the processed input,
to trade-off computational cost and accuracy. We focus on deploying these
algorithms on multi-core low-power IoT devices, designing a tool that
automatically converts a Python ensemble into optimized C code, and exploring
several optimizations that account for the available parallelism and memory
hierarchy. We extensively benchmark both static and dynamic RFs and GBTs on
three state-of-the-art IoT-relevant datasets, using an 8-core ultra-lowpower
System-on-Chip (SoC), GAP8, as the target platform. Thanks to the proposed
early-stopping mechanisms, we achieve an energy reduction of up to 37.9% with
respect to static GBTs (8.82 uJ vs 14.20 uJ per inference) and 41.7% with
respect to static RFs (2.86 uJ vs 4.90 uJ per inference), without losing
accuracy compared to the static model.
|
2306.09819 | Matthias Bitzer | Matthias Bitzer, Mona Meister, Christoph Zimmer | Amortized Inference for Gaussian Process Hyperparameters of Structured
Kernels | Accepted at UAI 2023 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning the kernel parameters for Gaussian processes is often the
computational bottleneck in applications such as online learning, Bayesian
optimization, or active learning. Amortizing parameter inference over different
datasets is a promising approach to dramatically speed up training time.
However, existing methods restrict the amortized inference procedure to a fixed
kernel structure. The amortization network must be redesigned manually and
trained again in case a different kernel is employed, which leads to a large
overhead in design time and training time. We propose amortizing kernel
parameter inference over a complete kernel-structure-family rather than a fixed
kernel structure. We do that via defining an amortization network over pairs of
datasets and kernel structures. This enables fast kernel inference for each
element in the kernel family without retraining the amortization network. As a
by-product, our amortization network is able to do fast ensembling over kernel
structures. In our experiments, we show drastically reduced inference time
combined with competitive test performance for a large set of kernels and
datasets.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:02:57 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Bitzer",
"Matthias",
""
],
[
"Meister",
"Mona",
""
],
[
"Zimmer",
"Christoph",
""
]
] | TITLE: Amortized Inference for Gaussian Process Hyperparameters of Structured
Kernels
ABSTRACT: Learning the kernel parameters for Gaussian processes is often the
computational bottleneck in applications such as online learning, Bayesian
optimization, or active learning. Amortizing parameter inference over different
datasets is a promising approach to dramatically speed up training time.
However, existing methods restrict the amortized inference procedure to a fixed
kernel structure. The amortization network must be redesigned manually and
trained again in case a different kernel is employed, which leads to a large
overhead in design time and training time. We propose amortizing kernel
parameter inference over a complete kernel-structure-family rather than a fixed
kernel structure. We do that via defining an amortization network over pairs of
datasets and kernel structures. This enables fast kernel inference for each
element in the kernel family without retraining the amortization network. As a
by-product, our amortization network is able to do fast ensembling over kernel
structures. In our experiments, we show drastically reduced inference time
combined with competitive test performance for a large set of kernels and
datasets.
|
2306.09824 | Kaushik Roy | Kaushik Roy, Yuxin Zi, Manas Gaur, Jinendra Malekar, Qi Zhang, Vignesh
Narayanan, Amit Sheth | Process Knowledge-infused Learning for Clinician-friendly Explanations | Accepted for Publication at AAAI Second Symposium on Human
Partnership with Medical Artificial Intelligence (HUMAN.AI Summer 2023):
Design, Operationalization, and Ethics. July 17-19, 2023 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Language models have the potential to assess mental health using social media
data. By analyzing online posts and conversations, these models can detect
patterns indicating mental health conditions like depression, anxiety, or
suicidal thoughts. They examine keywords, language markers, and sentiment to
gain insights into an individual's mental well-being. This information is
crucial for early detection, intervention, and support, improving mental health
care and prevention strategies. However, using language models for mental
health assessments from social media has two limitations: (1) They do not
compare posts against clinicians' diagnostic processes, and (2) It's
challenging to explain language model outputs using concepts that the clinician
can understand, i.e., clinician-friendly explanations. In this study, we
introduce Process Knowledge-infused Learning (PK-iL), a new learning paradigm
that layers clinical process knowledge structures on language model outputs,
enabling clinician-friendly explanations of the underlying language model
predictions. We rigorously test our methods on existing benchmark datasets,
augmented with such clinical process knowledge, and release a new dataset for
assessing suicidality. PK-iL performs competitively, achieving a 70% agreement
with users, while other XAI methods only achieve 47% agreement (average
inter-rater agreement of 0.72). Our evaluations demonstrate that PK-iL
effectively explains model predictions to clinicians.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:08:17 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Roy",
"Kaushik",
""
],
[
"Zi",
"Yuxin",
""
],
[
"Gaur",
"Manas",
""
],
[
"Malekar",
"Jinendra",
""
],
[
"Zhang",
"Qi",
""
],
[
"Narayanan",
"Vignesh",
""
],
[
"Sheth",
"Amit",
""
]
] | TITLE: Process Knowledge-infused Learning for Clinician-friendly Explanations
ABSTRACT: Language models have the potential to assess mental health using social media
data. By analyzing online posts and conversations, these models can detect
patterns indicating mental health conditions like depression, anxiety, or
suicidal thoughts. They examine keywords, language markers, and sentiment to
gain insights into an individual's mental well-being. This information is
crucial for early detection, intervention, and support, improving mental health
care and prevention strategies. However, using language models for mental
health assessments from social media has two limitations: (1) They do not
compare posts against clinicians' diagnostic processes, and (2) It's
challenging to explain language model outputs using concepts that the clinician
can understand, i.e., clinician-friendly explanations. In this study, we
introduce Process Knowledge-infused Learning (PK-iL), a new learning paradigm
that layers clinical process knowledge structures on language model outputs,
enabling clinician-friendly explanations of the underlying language model
predictions. We rigorously test our methods on existing benchmark datasets,
augmented with such clinical process knowledge, and release a new dataset for
assessing suicidality. PK-iL performs competitively, achieving a 70% agreement
with users, while other XAI methods only achieve 47% agreement (average
inter-rater agreement of 0.72). Our evaluations demonstrate that PK-iL
effectively explains model predictions to clinicians.
|
2306.09826 | Meisam Kabiri | Meisam Kabiri, Claudio Cimarelli, Hriday Bavle, Jose Luis
Sanchez-Lopez, and Holger Voos | Pose Graph Optimization for a MAV Indoor Localization Fusing 5GNR TOA
with an IMU | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by-sa/4.0/ | This paper explores the potential of 5G new radio (NR) Time-of-Arrival (TOA)
data for indoor drone localization under different scenarios and conditions
when fused with inertial measurement unit (IMU) data. Our approach involves
performing graph-based optimization to estimate the drone's position and
orientation from the multiple sensor measurements. Due to the lack of
real-world data, we use Matlab 5G toolbox and QuaDRiGa (quasi-deterministic
radio channel generator) channel simulator to generate TOA measurements for the
EuRoC MAV indoor dataset that provides IMU readings and ground truths 6DoF
poses of a flying drone. Hence, we create twelve sequences combining three
predefined indoor scenarios setups of QuaDRiGa with 2 to 5 base station
antennas. Therefore, experimental results demonstrate that, for a sufficient
number of base stations and a high bandwidth 5G configuration, the pose graph
optimization approach achieves accurate drone localization, with an average
error of less than 15 cm on the overall trajectory. Furthermore, the adopted
graph-based optimization algorithm is fast and can be easily implemented for
onboard real-time pose tracking on a micro aerial vehicle (MAV).
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:10:23 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Kabiri",
"Meisam",
""
],
[
"Cimarelli",
"Claudio",
""
],
[
"Bavle",
"Hriday",
""
],
[
"Sanchez-Lopez",
"Jose Luis",
""
],
[
"Voos",
"Holger",
""
]
] | TITLE: Pose Graph Optimization for a MAV Indoor Localization Fusing 5GNR TOA
with an IMU
ABSTRACT: This paper explores the potential of 5G new radio (NR) Time-of-Arrival (TOA)
data for indoor drone localization under different scenarios and conditions
when fused with inertial measurement unit (IMU) data. Our approach involves
performing graph-based optimization to estimate the drone's position and
orientation from the multiple sensor measurements. Due to the lack of
real-world data, we use Matlab 5G toolbox and QuaDRiGa (quasi-deterministic
radio channel generator) channel simulator to generate TOA measurements for the
EuRoC MAV indoor dataset that provides IMU readings and ground truths 6DoF
poses of a flying drone. Hence, we create twelve sequences combining three
predefined indoor scenarios setups of QuaDRiGa with 2 to 5 base station
antennas. Therefore, experimental results demonstrate that, for a sufficient
number of base stations and a high bandwidth 5G configuration, the pose graph
optimization approach achieves accurate drone localization, with an average
error of less than 15 cm on the overall trajectory. Furthermore, the adopted
graph-based optimization algorithm is fast and can be easily implemented for
onboard real-time pose tracking on a micro aerial vehicle (MAV).
|
2306.09827 | Florian Eilers | Florian Eilers and Xiaoyi Jiang | Building Blocks for a Complex-Valued Transformer Architecture | ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP). IEEE, 2023 | null | 10.1109/ICASSP49357.2023.10095349 | null | cs.LG cs.CV cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most deep learning pipelines are built on real-valued operations to deal with
real-valued inputs such as images, speech or music signals. However, a lot of
applications naturally make use of complex-valued signals or images, such as
MRI or remote sensing. Additionally the Fourier transform of signals is
complex-valued and has numerous applications. We aim to make deep learning
directly applicable to these complex-valued signals without using projections
into $\mathbb{R}^2$. Thus we add to the recent developments of complex-valued
neural networks by presenting building blocks to transfer the transformer
architecture to the complex domain. We present multiple versions of a
complex-valued Scaled Dot-Product Attention mechanism as well as a
complex-valued layer normalization. We test on a classification and a sequence
generation task on the MusicNet dataset and show improved robustness to
overfitting while maintaining on-par performance when compared to the
real-valued transformer architecture.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:11:15 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Eilers",
"Florian",
""
],
[
"Jiang",
"Xiaoyi",
""
]
] | TITLE: Building Blocks for a Complex-Valued Transformer Architecture
ABSTRACT: Most deep learning pipelines are built on real-valued operations to deal with
real-valued inputs such as images, speech or music signals. However, a lot of
applications naturally make use of complex-valued signals or images, such as
MRI or remote sensing. Additionally the Fourier transform of signals is
complex-valued and has numerous applications. We aim to make deep learning
directly applicable to these complex-valued signals without using projections
into $\mathbb{R}^2$. Thus we add to the recent developments of complex-valued
neural networks by presenting building blocks to transfer the transformer
architecture to the complex domain. We present multiple versions of a
complex-valued Scaled Dot-Product Attention mechanism as well as a
complex-valued layer normalization. We test on a classification and a sequence
generation task on the MusicNet dataset and show improved robustness to
overfitting while maintaining on-par performance when compared to the
real-valued transformer architecture.
|
2306.09844 | Yifan Jiang | Xingjian Bai, Guangyi He, Yifan Jiang, Jan Obloj | Wasserstein distributional robustness of neural networks | 23 pages, 6 figures, 8 tables | null | null | null | cs.LG cs.CV math.OC math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks are known to be vulnerable to adversarial attacks (AA).
For an image recognition task, this means that a small perturbation of the
original can result in the image being misclassified. Design of such attacks as
well as methods of adversarial training against them are subject of intense
research. We re-cast the problem using techniques of Wasserstein
distributionally robust optimization (DRO) and obtain novel contributions
leveraging recent insights from DRO sensitivity analysis. We consider a set of
distributional threat models. Unlike the traditional pointwise attacks, which
assume a uniform bound on perturbation of each input data point, distributional
threat models allow attackers to perturb inputs in a non-uniform way. We link
these more general attacks with questions of out-of-sample performance and
Knightian uncertainty. To evaluate the distributional robustness of neural
networks, we propose a first-order AA algorithm and its multi-step version. Our
attack algorithms include Fast Gradient Sign Method (FGSM) and Projected
Gradient Descent (PGD) as special cases. Furthermore, we provide a new
asymptotic estimate of the adversarial accuracy against distributional threat
models. The bound is fast to compute and first-order accurate, offering new
insights even for the pointwise AA. It also naturally yields out-of-sample
performance guarantees. We conduct numerical experiments on the CIFAR-10
dataset using DNNs on RobustBench to illustrate our theoretical results. Our
code is available at https://github.com/JanObloj/W-DRO-Adversarial-Methods.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:41:24 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Bai",
"Xingjian",
""
],
[
"He",
"Guangyi",
""
],
[
"Jiang",
"Yifan",
""
],
[
"Obloj",
"Jan",
""
]
] | TITLE: Wasserstein distributional robustness of neural networks
ABSTRACT: Deep neural networks are known to be vulnerable to adversarial attacks (AA).
For an image recognition task, this means that a small perturbation of the
original can result in the image being misclassified. Design of such attacks as
well as methods of adversarial training against them are subject of intense
research. We re-cast the problem using techniques of Wasserstein
distributionally robust optimization (DRO) and obtain novel contributions
leveraging recent insights from DRO sensitivity analysis. We consider a set of
distributional threat models. Unlike the traditional pointwise attacks, which
assume a uniform bound on perturbation of each input data point, distributional
threat models allow attackers to perturb inputs in a non-uniform way. We link
these more general attacks with questions of out-of-sample performance and
Knightian uncertainty. To evaluate the distributional robustness of neural
networks, we propose a first-order AA algorithm and its multi-step version. Our
attack algorithms include Fast Gradient Sign Method (FGSM) and Projected
Gradient Descent (PGD) as special cases. Furthermore, we provide a new
asymptotic estimate of the adversarial accuracy against distributional threat
models. The bound is fast to compute and first-order accurate, offering new
insights even for the pointwise AA. It also naturally yields out-of-sample
performance guarantees. We conduct numerical experiments on the CIFAR-10
dataset using DNNs on RobustBench to illustrate our theoretical results. Our
code is available at https://github.com/JanObloj/W-DRO-Adversarial-Methods.
|
2306.09848 | Ege Gursoy | Ege Gursoy, Sonny Tarbouriech, Andrea Cherubini | Can robots mold soft plastic materials by shaping depth images? | Accepted to IEEE Transactions on Robotics (T-RO) | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | Can robots mold soft plastic materials by shaping depth images? The short
answer is no: current day robots can't. In this article, we address the problem
of shaping plastic material with an anthropomorphic arm/hand robot, which
observes the material with a fixed depth camera. Robots capable of molding
could assist humans in many tasks, such as cooking, scooping or gardening. Yet,
the problem is complex, due to its high-dimensionality at both perception and
control levels. To address it, we design three alternative data-based methods
for predicting the effect of robot actions on the material. Then, the robot can
plan the sequence of actions and their positions, to mold the material into a
desired shape. To make the prediction problem tractable, we rely on two
original ideas. First, we prove that under reasonable assumptions, the shaping
problem can be mapped from point cloud to depth image space, with many benefits
(simpler processing, no need for registration, lower computation time and
memory requirements). Second, we design a novel, simple metric for quickly
measuring the distance between two depth images. The metric is based on the
inherent point cloud representation of depth images, which enables direct and
consistent comparison of image pairs through a non-uniform scaling approach,
and therefore opens promising perspectives for designing \textit{depth image --
based} robot controllers. We assess our approach in a series of unprecedented
experiments, where a robotic arm/hand molds flour from initial to final shapes,
either with its own dataset, or by transfer learning from a human dataset. We
conclude the article by discussing the limitations of our framework and those
of current day hardware, which make human-like robot molding a challenging open
research problem.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 13:46:15 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Gursoy",
"Ege",
""
],
[
"Tarbouriech",
"Sonny",
""
],
[
"Cherubini",
"Andrea",
""
]
] | TITLE: Can robots mold soft plastic materials by shaping depth images?
ABSTRACT: Can robots mold soft plastic materials by shaping depth images? The short
answer is no: current day robots can't. In this article, we address the problem
of shaping plastic material with an anthropomorphic arm/hand robot, which
observes the material with a fixed depth camera. Robots capable of molding
could assist humans in many tasks, such as cooking, scooping or gardening. Yet,
the problem is complex, due to its high-dimensionality at both perception and
control levels. To address it, we design three alternative data-based methods
for predicting the effect of robot actions on the material. Then, the robot can
plan the sequence of actions and their positions, to mold the material into a
desired shape. To make the prediction problem tractable, we rely on two
original ideas. First, we prove that under reasonable assumptions, the shaping
problem can be mapped from point cloud to depth image space, with many benefits
(simpler processing, no need for registration, lower computation time and
memory requirements). Second, we design a novel, simple metric for quickly
measuring the distance between two depth images. The metric is based on the
inherent point cloud representation of depth images, which enables direct and
consistent comparison of image pairs through a non-uniform scaling approach,
and therefore opens promising perspectives for designing \textit{depth image --
based} robot controllers. We assess our approach in a series of unprecedented
experiments, where a robotic arm/hand molds flour from initial to final shapes,
either with its own dataset, or by transfer learning from a human dataset. We
conclude the article by discussing the limitations of our framework and those
of current day hardware, which make human-like robot molding a challenging open
research problem.
|
2306.09886 | Jakub Nalepa | Bartosz Grabowski, Maciej Ziaja, Michal Kawulok, Piotr Bosowski,
Nicolas Long\'ep\'e, Bertrand Le Saux, Jakub Nalepa | Squeezing nnU-Nets with Knowledge Distillation for On-Board Cloud
Detection | arXiv admin note: substantial text overlap with arXiv:2210.13659 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cloud detection is a pivotal satellite image pre-processing step that can be
performed both on the ground and on board a satellite to tag useful images. In
the latter case, it can reduce the amount of data to downlink by pruning the
cloudy areas, or to make a satellite more autonomous through data-driven
acquisition re-scheduling. We approach this task with nnU-Nets, a
self-reconfigurable framework able to perform meta-learning of a segmentation
network over various datasets. Unfortunately, such models are commonly
memory-inefficient due to their (very) large architectures. To benefit from
them in on-board processing, we compress nnU-Nets with knowledge distillation
into much smaller and compact U-Nets. Our experiments, performed over
Sentinel-2 and Landsat-8 images revealed that nnU-Nets deliver state-of-the-art
performance without any manual design. Our approach was ranked within the top
7% best solutions (across 847 teams) in the On Cloud N: Cloud Cover Detection
Challenge, where we reached the Jaccard index of 0.882 over more than 10k
unseen Sentinel-2 images (the winners obtained 0.897, the baseline U-Net with
the ResNet-34 backbone: 0.817, and the classic Sentinel-2 image thresholding:
0.652). Finally, we showed that knowledge distillation enables to elaborate
dramatically smaller (almost 280x) U-Nets when compared to nnU-Nets while still
maintaining their segmentation capabilities.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 14:53:36 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Grabowski",
"Bartosz",
""
],
[
"Ziaja",
"Maciej",
""
],
[
"Kawulok",
"Michal",
""
],
[
"Bosowski",
"Piotr",
""
],
[
"Longépé",
"Nicolas",
""
],
[
"Saux",
"Bertrand Le",
""
],
[
"Nalepa",
"Jakub",
""
]
] | TITLE: Squeezing nnU-Nets with Knowledge Distillation for On-Board Cloud
Detection
ABSTRACT: Cloud detection is a pivotal satellite image pre-processing step that can be
performed both on the ground and on board a satellite to tag useful images. In
the latter case, it can reduce the amount of data to downlink by pruning the
cloudy areas, or to make a satellite more autonomous through data-driven
acquisition re-scheduling. We approach this task with nnU-Nets, a
self-reconfigurable framework able to perform meta-learning of a segmentation
network over various datasets. Unfortunately, such models are commonly
memory-inefficient due to their (very) large architectures. To benefit from
them in on-board processing, we compress nnU-Nets with knowledge distillation
into much smaller and compact U-Nets. Our experiments, performed over
Sentinel-2 and Landsat-8 images revealed that nnU-Nets deliver state-of-the-art
performance without any manual design. Our approach was ranked within the top
7% best solutions (across 847 teams) in the On Cloud N: Cloud Cover Detection
Challenge, where we reached the Jaccard index of 0.882 over more than 10k
unseen Sentinel-2 images (the winners obtained 0.897, the baseline U-Net with
the ResNet-34 backbone: 0.817, and the classic Sentinel-2 image thresholding:
0.652). Finally, we showed that knowledge distillation enables to elaborate
dramatically smaller (almost 280x) U-Nets when compared to nnU-Nets while still
maintaining their segmentation capabilities.
|
2306.09918 | Animesh Nighojkar | Animesh Nighojkar and Antonio Laverghetta Jr. and John Licato | No Strong Feelings One Way or Another: Re-operationalizing Neutrality in
Natural Language Inference | Appearing at the 17th Linguistic Annotation Workshop at ACL 2023 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Natural Language Inference (NLI) has been a cornerstone task in evaluating
language models' inferential reasoning capabilities. However, the standard
three-way classification scheme used in NLI has well-known shortcomings in
evaluating models' ability to capture the nuances of natural human reasoning.
In this paper, we argue that the operationalization of the neutral label in
current NLI datasets has low validity, is interpreted inconsistently, and that
at least one important sense of neutrality is often ignored. We uncover the
detrimental impact of these shortcomings, which in some cases leads to
annotation datasets that actually decrease performance on downstream tasks. We
compare approaches of handling annotator disagreement and identify flaws in a
recent NLI dataset that designs an annotator study based on a problematic
operationalization. Our findings highlight the need for a more refined
evaluation framework for NLI, and we hope to spark further discussion and
action in the NLP community.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 15:45:08 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Nighojkar",
"Animesh",
""
],
[
"Laverghetta",
"Antonio",
"Jr."
],
[
"Licato",
"John",
""
]
] | TITLE: No Strong Feelings One Way or Another: Re-operationalizing Neutrality in
Natural Language Inference
ABSTRACT: Natural Language Inference (NLI) has been a cornerstone task in evaluating
language models' inferential reasoning capabilities. However, the standard
three-way classification scheme used in NLI has well-known shortcomings in
evaluating models' ability to capture the nuances of natural human reasoning.
In this paper, we argue that the operationalization of the neutral label in
current NLI datasets has low validity, is interpreted inconsistently, and that
at least one important sense of neutrality is often ignored. We uncover the
detrimental impact of these shortcomings, which in some cases leads to
annotation datasets that actually decrease performance on downstream tasks. We
compare approaches of handling annotator disagreement and identify flaws in a
recent NLI dataset that designs an annotator study based on a problematic
operationalization. Our findings highlight the need for a more refined
evaluation framework for NLI, and we hope to spark further discussion and
action in the NLP community.
|
2306.09938 | Iain Mackie | Iain Mackie, Ivan Sekulic, Shubham Chatterjee, Jeffrey Dalton, Fabio
Crestani | GRM: Generative Relevance Modeling Using Relevance-Aware Sample
Estimation for Document Retrieval | null | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent studies show that Generative Relevance Feedback (GRF), using text
generated by Large Language Models (LLMs), can enhance the effectiveness of
query expansion. However, LLMs can generate irrelevant information that harms
retrieval effectiveness. To address this, we propose Generative Relevance
Modeling (GRM) that uses Relevance-Aware Sample Estimation (RASE) for more
accurate weighting of expansion terms. Specifically, we identify similar real
documents for each generated document and use a neural re-ranker to estimate
their relevance. Experiments on three standard document ranking benchmarks show
that GRM improves MAP by 6-9% and R@1k by 2-4%, surpassing previous methods.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 16:18:11 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Mackie",
"Iain",
""
],
[
"Sekulic",
"Ivan",
""
],
[
"Chatterjee",
"Shubham",
""
],
[
"Dalton",
"Jeffrey",
""
],
[
"Crestani",
"Fabio",
""
]
] | TITLE: GRM: Generative Relevance Modeling Using Relevance-Aware Sample
Estimation for Document Retrieval
ABSTRACT: Recent studies show that Generative Relevance Feedback (GRF), using text
generated by Large Language Models (LLMs), can enhance the effectiveness of
query expansion. However, LLMs can generate irrelevant information that harms
retrieval effectiveness. To address this, we propose Generative Relevance
Modeling (GRM) that uses Relevance-Aware Sample Estimation (RASE) for more
accurate weighting of expansion terms. Specifically, we identify similar real
documents for each generated document and use a neural re-ranker to estimate
their relevance. Experiments on three standard document ranking benchmarks show
that GRM improves MAP by 6-9% and R@1k by 2-4%, surpassing previous methods.
|
2306.09944 | Jiajun Wu | Samuel Clarke, Ruohan Gao, Mason Wang, Mark Rau, Julia Xu, Jui-Hsien
Wang, Doug L. James, Jiajun Wu | RealImpact: A Dataset of Impact Sound Fields for Real Objects | CVPR 2023 (Highlight). Project page:
https://samuelpclarke.com/realimpact/ | null | null | null | cs.SD cs.CV cs.GR eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Objects make unique sounds under different perturbations, environment
conditions, and poses relative to the listener. While prior works have modeled
impact sounds and sound propagation in simulation, we lack a standard dataset
of impact sound fields of real objects for audio-visual learning and
calibration of the sim-to-real gap. We present RealImpact, a large-scale
dataset of real object impact sounds recorded under controlled conditions.
RealImpact contains 150,000 recordings of impact sounds of 50 everyday objects
with detailed annotations, including their impact locations, microphone
locations, contact force profiles, material labels, and RGBD images. We make
preliminary attempts to use our dataset as a reference to current simulation
methods for estimating object impact sounds that match the real world.
Moreover, we demonstrate the usefulness of our dataset as a testbed for
acoustic and audio-visual learning via the evaluation of two benchmark tasks,
including listener location classification and visual acoustic matching.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 16:25:41 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Clarke",
"Samuel",
""
],
[
"Gao",
"Ruohan",
""
],
[
"Wang",
"Mason",
""
],
[
"Rau",
"Mark",
""
],
[
"Xu",
"Julia",
""
],
[
"Wang",
"Jui-Hsien",
""
],
[
"James",
"Doug L.",
""
],
[
"Wu",
"Jiajun",
""
]
] | TITLE: RealImpact: A Dataset of Impact Sound Fields for Real Objects
ABSTRACT: Objects make unique sounds under different perturbations, environment
conditions, and poses relative to the listener. While prior works have modeled
impact sounds and sound propagation in simulation, we lack a standard dataset
of impact sound fields of real objects for audio-visual learning and
calibration of the sim-to-real gap. We present RealImpact, a large-scale
dataset of real object impact sounds recorded under controlled conditions.
RealImpact contains 150,000 recordings of impact sounds of 50 everyday objects
with detailed annotations, including their impact locations, microphone
locations, contact force profiles, material labels, and RGBD images. We make
preliminary attempts to use our dataset as a reference to current simulation
methods for estimating object impact sounds that match the real world.
Moreover, we demonstrate the usefulness of our dataset as a testbed for
acoustic and audio-visual learning via the evaluation of two benchmark tasks,
including listener location classification and visual acoustic matching.
|
2306.09949 | Othmane Laousy | Othmane Laousy, Alexandre Araujo, Guillaume Chassagnon, Marie-Pierre
Revel, Siddharth Garg, Farshad Khorrami, Maria Vakalopoulou | Towards Better Certified Segmentation via Diffusion Models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The robustness of image segmentation has been an important research topic in
the past few years as segmentation models have reached production-level
accuracy. However, like classification models, segmentation models can be
vulnerable to adversarial perturbations, which hinders their use in
critical-decision systems like healthcare or autonomous driving. Recently,
randomized smoothing has been proposed to certify segmentation predictions by
adding Gaussian noise to the input to obtain theoretical guarantees. However,
this method exhibits a trade-off between the amount of added noise and the
level of certification achieved. In this paper, we address the problem of
certifying segmentation prediction using a combination of randomized smoothing
and diffusion models. Our experiments show that combining randomized smoothing
and diffusion models significantly improves certified robustness, with results
indicating a mean improvement of 21 points in accuracy compared to previous
state-of-the-art methods on Pascal-Context and Cityscapes public datasets. Our
method is independent of the selected segmentation model and does not need any
additional specialized training procedure.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 16:30:39 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Laousy",
"Othmane",
""
],
[
"Araujo",
"Alexandre",
""
],
[
"Chassagnon",
"Guillaume",
""
],
[
"Revel",
"Marie-Pierre",
""
],
[
"Garg",
"Siddharth",
""
],
[
"Khorrami",
"Farshad",
""
],
[
"Vakalopoulou",
"Maria",
""
]
] | TITLE: Towards Better Certified Segmentation via Diffusion Models
ABSTRACT: The robustness of image segmentation has been an important research topic in
the past few years as segmentation models have reached production-level
accuracy. However, like classification models, segmentation models can be
vulnerable to adversarial perturbations, which hinders their use in
critical-decision systems like healthcare or autonomous driving. Recently,
randomized smoothing has been proposed to certify segmentation predictions by
adding Gaussian noise to the input to obtain theoretical guarantees. However,
this method exhibits a trade-off between the amount of added noise and the
level of certification achieved. In this paper, we address the problem of
certifying segmentation prediction using a combination of randomized smoothing
and diffusion models. Our experiments show that combining randomized smoothing
and diffusion models significantly improves certified robustness, with results
indicating a mean improvement of 21 points in accuracy compared to previous
state-of-the-art methods on Pascal-Context and Cityscapes public datasets. Our
method is independent of the selected segmentation model and does not need any
additional specialized training procedure.
|
2306.09950 | Steinar Laenen | Steinar Laenen, Bogdan-Adrian Manghiuc, He Sun | Nearly-Optimal Hierarchical Clustering for Well-Clustered Graphs | This work is accepted at the 40th International Conference on Machine
Learning (ICML'23) | null | null | null | cs.DS cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents two efficient hierarchical clustering (HC) algorithms
with respect to Dasgupta's cost function. For any input graph $G$ with a clear
cluster-structure, our designed algorithms run in nearly-linear time in the
input size of $G$, and return an $O(1)$-approximate HC tree with respect to
Dasgupta's cost function. We compare the performance of our algorithm against
the previous state-of-the-art on synthetic and real-world datasets and show
that our designed algorithm produces comparable or better HC trees with much
lower running time.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 16:31:46 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Laenen",
"Steinar",
""
],
[
"Manghiuc",
"Bogdan-Adrian",
""
],
[
"Sun",
"He",
""
]
] | TITLE: Nearly-Optimal Hierarchical Clustering for Well-Clustered Graphs
ABSTRACT: This paper presents two efficient hierarchical clustering (HC) algorithms
with respect to Dasgupta's cost function. For any input graph $G$ with a clear
cluster-structure, our designed algorithms run in nearly-linear time in the
input size of $G$, and return an $O(1)$-approximate HC tree with respect to
Dasgupta's cost function. We compare the performance of our algorithm against
the previous state-of-the-art on synthetic and real-world datasets and show
that our designed algorithm produces comparable or better HC trees with much
lower running time.
|
2306.09973 | Mohammad Hasan Ahmadilivani | Mohammad Hasan Ahmadilivani, Mahdi Taheri, Jaan Raik, Masoud
Daneshtalab, and Maksim Jenihhin | Enhancing Fault Resilience of QNNs by Selective Neuron Splitting | 5 pages, 4 figures, 1 table. The paper is accepted at the AICAS'23
conference | null | null | null | cs.LG cs.AR cs.NE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The superior performance of Deep Neural Networks (DNNs) has led to their
application in various aspects of human life. Safety-critical applications are
no exception and impose rigorous reliability requirements on DNNs. Quantized
Neural Networks (QNNs) have emerged to tackle the complexity of DNN
accelerators, however, they are more prone to reliability issues.
In this paper, a recent analytical resilience assessment method is adapted
for QNNs to identify critical neurons based on a Neuron Vulnerability Factor
(NVF). Thereafter, a novel method for splitting the critical neurons is
proposed that enables the design of a Lightweight Correction Unit (LCU) in the
accelerator without redesigning its computational part.
The method is validated by experiments on different QNNs and datasets. The
results demonstrate that the proposed method for correcting the faults has a
twice smaller overhead than a selective Triple Modular Redundancy (TMR) while
achieving a similar level of fault resiliency.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 17:11:55 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Ahmadilivani",
"Mohammad Hasan",
""
],
[
"Taheri",
"Mahdi",
""
],
[
"Raik",
"Jaan",
""
],
[
"Daneshtalab",
"Masoud",
""
],
[
"Jenihhin",
"Maksim",
""
]
] | TITLE: Enhancing Fault Resilience of QNNs by Selective Neuron Splitting
ABSTRACT: The superior performance of Deep Neural Networks (DNNs) has led to their
application in various aspects of human life. Safety-critical applications are
no exception and impose rigorous reliability requirements on DNNs. Quantized
Neural Networks (QNNs) have emerged to tackle the complexity of DNN
accelerators, however, they are more prone to reliability issues.
In this paper, a recent analytical resilience assessment method is adapted
for QNNs to identify critical neurons based on a Neuron Vulnerability Factor
(NVF). Thereafter, a novel method for splitting the critical neurons is
proposed that enables the design of a Lightweight Correction Unit (LCU) in the
accelerator without redesigning its computational part.
The method is validated by experiments on different QNNs and datasets. The
results demonstrate that the proposed method for correcting the faults has a
twice smaller overhead than a selective Triple Modular Redundancy (TMR) while
achieving a similar level of fault resiliency.
|
2306.09979 | Frederico Santos De Oliveira | Frederico S. Oliveira, Edresson Casanova, Arnaldo C\^andido J\'unior,
Lucas R. S. Gris, Anderson S. Soares, and Arlindo R. Galv\~ao Filho | Evaluation of Speech Representations for MOS prediction | 12 pages, 4 figures, Accepted to the 26th International Conference of
Text, Speech and Dialogue (TSD2023) | null | null | null | cs.SD cs.AI eess.AS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we evaluate feature extraction models for predicting speech
quality. We also propose a model architecture to compare embeddings of
supervised learning and self-supervised learning models with embeddings of
speaker verification models to predict the metric MOS. Our experiments were
performed on the VCC2018 dataset and a Brazilian-Portuguese dataset called
BRSpeechMOS, which was created for this work. The results show that the Whisper
model is appropriate in all scenarios: with both the VCC2018 and BRSpeech- MOS
datasets. Among the supervised and self-supervised learning models using
BRSpeechMOS, Whisper-Small achieved the best linear correlation of 0.6980, and
the speaker verification model, SpeakerNet, had linear correlation of 0.6963.
Using VCC2018, the best supervised and self-supervised learning model,
Whisper-Large, achieved linear correlation of 0.7274, and the best model
speaker verification, TitaNet, achieved a linear correlation of 0.6933.
Although the results of the speaker verification models are slightly lower, the
SpeakerNet model has only 5M parameters, making it suitable for real-time
applications, and the TitaNet model produces an embedding of size 192, the
smallest among all the evaluated models. The experiment results are
reproducible with publicly available source-code1 .
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 17:21:42 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Oliveira",
"Frederico S.",
""
],
[
"Casanova",
"Edresson",
""
],
[
"Júnior",
"Arnaldo Cândido",
""
],
[
"Gris",
"Lucas R. S.",
""
],
[
"Soares",
"Anderson S.",
""
],
[
"Filho",
"Arlindo R. Galvão",
""
]
] | TITLE: Evaluation of Speech Representations for MOS prediction
ABSTRACT: In this paper, we evaluate feature extraction models for predicting speech
quality. We also propose a model architecture to compare embeddings of
supervised learning and self-supervised learning models with embeddings of
speaker verification models to predict the metric MOS. Our experiments were
performed on the VCC2018 dataset and a Brazilian-Portuguese dataset called
BRSpeechMOS, which was created for this work. The results show that the Whisper
model is appropriate in all scenarios: with both the VCC2018 and BRSpeech- MOS
datasets. Among the supervised and self-supervised learning models using
BRSpeechMOS, Whisper-Small achieved the best linear correlation of 0.6980, and
the speaker verification model, SpeakerNet, had linear correlation of 0.6963.
Using VCC2018, the best supervised and self-supervised learning model,
Whisper-Large, achieved linear correlation of 0.7274, and the best model
speaker verification, TitaNet, achieved a linear correlation of 0.6933.
Although the results of the speaker verification models are slightly lower, the
SpeakerNet model has only 5M parameters, making it suitable for real-time
applications, and the TitaNet model produces an embedding of size 192, the
smallest among all the evaluated models. The experiment results are
reproducible with publicly available source-code1 .
|
2306.09987 | Albert Larson | Albert Larson and Ali Shafqat Akanda | Transforming Observations of Ocean Temperature with a Deep Convolutional
Residual Regressive Neural Network | null | null | null | null | physics.ao-ph cs.LG | http://creativecommons.org/licenses/by/4.0/ | Sea surface temperature (SST) is an essential climate variable that can be
measured via ground truth, remote sensing, or hybrid model methodologies. Here,
we celebrate SST surveillance progress via the application of a few relevant
technological advances from the late 20th and early 21st century. We further
develop our existing water cycle observation framework, Flux to Flow (F2F), to
fuse AMSR-E and MODIS into a higher resolution product with the goal of
capturing gradients and filling cloud gaps that are otherwise unavailable. Our
neural network architecture is constrained to a deep convolutional residual
regressive neural network. We utilize three snapshots of twelve monthly SST
measurements in 2010 as measured by the passive microwave radiometer AMSR-E,
the visible and infrared monitoring MODIS instrument, and the in situ Argo
dataset ISAS. The performance of the platform and success of this approach is
evaluated using the root mean squared error (RMSE) metric. We determine that
the 1:1 configuration of input and output data and a large observation region
is too challenging for the single compute node and dcrrnn structure as is. When
constrained to a single 100 x 100 pixel region and a small training dataset,
the algorithm improves from the baseline experiment covering a much larger
geography. For next discrete steps, we envision the consideration of a large
input range with a very small output range. Furthermore, we see the need to
integrate land and sea variables before performing computer vision tasks like
those within. Finally, we see parallelization as necessary to overcome the
compute obstacles we encountered.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 17:35:11 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Larson",
"Albert",
""
],
[
"Akanda",
"Ali Shafqat",
""
]
] | TITLE: Transforming Observations of Ocean Temperature with a Deep Convolutional
Residual Regressive Neural Network
ABSTRACT: Sea surface temperature (SST) is an essential climate variable that can be
measured via ground truth, remote sensing, or hybrid model methodologies. Here,
we celebrate SST surveillance progress via the application of a few relevant
technological advances from the late 20th and early 21st century. We further
develop our existing water cycle observation framework, Flux to Flow (F2F), to
fuse AMSR-E and MODIS into a higher resolution product with the goal of
capturing gradients and filling cloud gaps that are otherwise unavailable. Our
neural network architecture is constrained to a deep convolutional residual
regressive neural network. We utilize three snapshots of twelve monthly SST
measurements in 2010 as measured by the passive microwave radiometer AMSR-E,
the visible and infrared monitoring MODIS instrument, and the in situ Argo
dataset ISAS. The performance of the platform and success of this approach is
evaluated using the root mean squared error (RMSE) metric. We determine that
the 1:1 configuration of input and output data and a large observation region
is too challenging for the single compute node and dcrrnn structure as is. When
constrained to a single 100 x 100 pixel region and a small training dataset,
the algorithm improves from the baseline experiment covering a much larger
geography. For next discrete steps, we envision the consideration of a large
input range with a very small output range. Furthermore, we see the need to
integrate land and sea variables before performing computer vision tasks like
those within. Finally, we see parallelization as necessary to overcome the
compute obstacles we encountered.
|
2306.09991 | Andrei Kucharavy | Andrei Kucharavy, Rachid Guerraoui and Ljiljana Dolamic | Evolutionary Algorithms in the Light of SGD: Limit Equivalence, Minima
Flatness, and Transfer Learning | To be published in ALIFE 2023; 16 pages, 10 figures, 1 listing | null | null | null | cs.NE cs.LG q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Whenever applicable, the Stochastic Gradient Descent (SGD) has shown itself
to be unreasonably effective. Instead of underperforming and getting trapped in
local minima due to the batch noise, SGD leverages it to learn to generalize
better and find minima that are good enough for the entire dataset. This led to
numerous theoretical and experimental investigations, especially in the context
of Artificial Neural Networks (ANNs), leading to better machine learning
algorithms. However, SGD is not applicable in a non-differentiable setting,
leaving all that prior research off the table.
In this paper, we show that a class of evolutionary algorithms (EAs) inspired
by the Gillespie-Orr Mutational Landscapes model for natural evolution is
formally equivalent to SGD in certain settings and, in practice, is well
adapted to large ANNs. We refer to such EAs as Gillespie-Orr EA class (GO-EAs)
and empirically show how an insight transfer from SGD can work for them. We
then show that for ANNs trained to near-optimality or in the transfer learning
setting, the equivalence also allows transferring the insights from the
Mutational Landscapes model to SGD.
We then leverage this equivalence to experimentally show how SGD and GO-EAs
can provide mutual insight through examples of minima flatness, transfer
learning, and mixing of individuals in EAs applied to large models.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 22:26:44 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Kucharavy",
"Andrei",
""
],
[
"Guerraoui",
"Rachid",
""
],
[
"Dolamic",
"Ljiljana",
""
]
] | TITLE: Evolutionary Algorithms in the Light of SGD: Limit Equivalence, Minima
Flatness, and Transfer Learning
ABSTRACT: Whenever applicable, the Stochastic Gradient Descent (SGD) has shown itself
to be unreasonably effective. Instead of underperforming and getting trapped in
local minima due to the batch noise, SGD leverages it to learn to generalize
better and find minima that are good enough for the entire dataset. This led to
numerous theoretical and experimental investigations, especially in the context
of Artificial Neural Networks (ANNs), leading to better machine learning
algorithms. However, SGD is not applicable in a non-differentiable setting,
leaving all that prior research off the table.
In this paper, we show that a class of evolutionary algorithms (EAs) inspired
by the Gillespie-Orr Mutational Landscapes model for natural evolution is
formally equivalent to SGD in certain settings and, in practice, is well
adapted to large ANNs. We refer to such EAs as Gillespie-Orr EA class (GO-EAs)
and empirically show how an insight transfer from SGD can work for them. We
then show that for ANNs trained to near-optimality or in the transfer learning
setting, the equivalence also allows transferring the insights from the
Mutational Landscapes model to SGD.
We then leverage this equivalence to experimentally show how SGD and GO-EAs
can provide mutual insight through examples of minima flatness, transfer
learning, and mixing of individuals in EAs applied to large models.
|
2306.09998 | Juliette Marrie | Juliette Marrie, Michael Arbel, Diane Larlus, Julien Mairal | SLACK: Stable Learning of Augmentations with Cold-start and KL
regularization | Accepted to CVPR 2023 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data augmentation is known to improve the generalization capabilities of
neural networks, provided that the set of transformations is chosen with care,
a selection often performed manually. Automatic data augmentation aims at
automating this process. However, most recent approaches still rely on some
prior information; they start from a small pool of manually-selected default
transformations that are either used to pretrain the network or forced to be
part of the policy learned by the automatic data augmentation algorithm. In
this paper, we propose to directly learn the augmentation policy without
leveraging such prior knowledge. The resulting bilevel optimization problem
becomes more challenging due to the larger search space and the inherent
instability of bilevel optimization algorithms. To mitigate these issues (i) we
follow a successive cold-start strategy with a Kullback-Leibler regularization,
and (ii) we parameterize magnitudes as continuous distributions. Our approach
leads to competitive results on standard benchmarks despite a more challenging
setting, and generalizes beyond natural images.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 17:51:07 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Marrie",
"Juliette",
""
],
[
"Arbel",
"Michael",
""
],
[
"Larlus",
"Diane",
""
],
[
"Mairal",
"Julien",
""
]
] | TITLE: SLACK: Stable Learning of Augmentations with Cold-start and KL
regularization
ABSTRACT: Data augmentation is known to improve the generalization capabilities of
neural networks, provided that the set of transformations is chosen with care,
a selection often performed manually. Automatic data augmentation aims at
automating this process. However, most recent approaches still rely on some
prior information; they start from a small pool of manually-selected default
transformations that are either used to pretrain the network or forced to be
part of the policy learned by the automatic data augmentation algorithm. In
this paper, we propose to directly learn the augmentation policy without
leveraging such prior knowledge. The resulting bilevel optimization problem
becomes more challenging due to the larger search space and the inherent
instability of bilevel optimization algorithms. To mitigate these issues (i) we
follow a successive cold-start strategy with a Kullback-Leibler regularization,
and (ii) we parameterize magnitudes as continuous distributions. Our approach
leads to competitive results on standard benchmarks despite a more challenging
setting, and generalizes beyond natural images.
|
2306.10013 | Yuqi Wang | Yuqi Wang, Yuntao Chen, Xingyu Liao, Lue Fan and Zhaoxiang Zhang | PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic
Segmentation | technical report | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | Comprehensive modeling of the surrounding 3D world is key to the success of
autonomous driving. However, existing perception tasks like object detection,
road structure segmentation, depth & elevation estimation, and open-set object
localization each only focus on a small facet of the holistic 3D scene
understanding task. This divide-and-conquer strategy simplifies the algorithm
development procedure at the cost of losing an end-to-end unified solution to
the problem. In this work, we address this limitation by studying camera-based
3D panoptic segmentation, aiming to achieve a unified occupancy representation
for camera-only 3D scene understanding. To achieve this, we introduce a novel
method called PanoOcc, which utilizes voxel queries to aggregate spatiotemporal
information from multi-frame and multi-view images in a coarse-to-fine scheme,
integrating feature learning and scene representation into a unified occupancy
representation. We have conducted extensive ablation studies to verify the
effectiveness and efficiency of the proposed method. Our approach achieves new
state-of-the-art results for camera-based semantic segmentation and panoptic
segmentation on the nuScenes dataset. Furthermore, our method can be easily
extended to dense occupancy prediction and has shown promising performance on
the Occ3D benchmark. The code will be released at
https://github.com/Robertwyq/PanoOcc.
| [
{
"version": "v1",
"created": "Fri, 16 Jun 2023 17:59:33 GMT"
}
] | 2023-06-19T00:00:00 | [
[
"Wang",
"Yuqi",
""
],
[
"Chen",
"Yuntao",
""
],
[
"Liao",
"Xingyu",
""
],
[
"Fan",
"Lue",
""
],
[
"Zhang",
"Zhaoxiang",
""
]
] | TITLE: PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic
Segmentation
ABSTRACT: Comprehensive modeling of the surrounding 3D world is key to the success of
autonomous driving. However, existing perception tasks like object detection,
road structure segmentation, depth & elevation estimation, and open-set object
localization each only focus on a small facet of the holistic 3D scene
understanding task. This divide-and-conquer strategy simplifies the algorithm
development procedure at the cost of losing an end-to-end unified solution to
the problem. In this work, we address this limitation by studying camera-based
3D panoptic segmentation, aiming to achieve a unified occupancy representation
for camera-only 3D scene understanding. To achieve this, we introduce a novel
method called PanoOcc, which utilizes voxel queries to aggregate spatiotemporal
information from multi-frame and multi-view images in a coarse-to-fine scheme,
integrating feature learning and scene representation into a unified occupancy
representation. We have conducted extensive ablation studies to verify the
effectiveness and efficiency of the proposed method. Our approach achieves new
state-of-the-art results for camera-based semantic segmentation and panoptic
segmentation on the nuScenes dataset. Furthermore, our method can be easily
extended to dense occupancy prediction and has shown promising performance on
the Occ3D benchmark. The code will be released at
https://github.com/Robertwyq/PanoOcc.
|
1906.06521 | Hongsong Wang | Hongsong Wang and Jiashi Feng | Delving into 3D Action Anticipation from Streaming Videos | We will modify this method | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Action anticipation, which aims to recognize the action with a partial
observation, becomes increasingly popular due to a wide range of applications.
In this paper, we investigate the problem of 3D action anticipation from
streaming videos with the target of understanding best practices for solving
this problem. We first introduce several complementary evaluation metrics and
present a basic model based on frame-wise action classification. To achieve
better performance, we then investigate two important factors, i.e., the length
of the training clip and clip sampling method. We also explore multi-task
learning strategies by incorporating auxiliary information from two aspects:
the full action representation and the class-agnostic action label. Our
comprehensive experiments uncover the best practices for 3D action
anticipation, and accordingly we propose a novel method with a multi-task loss.
The proposed method considerably outperforms the recent methods and exhibits
the state-of-the-art performance on standard benchmarks.
| [
{
"version": "v1",
"created": "Sat, 15 Jun 2019 10:30:29 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 00:09:45 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Wang",
"Hongsong",
""
],
[
"Feng",
"Jiashi",
""
]
] | TITLE: Delving into 3D Action Anticipation from Streaming Videos
ABSTRACT: Action anticipation, which aims to recognize the action with a partial
observation, becomes increasingly popular due to a wide range of applications.
In this paper, we investigate the problem of 3D action anticipation from
streaming videos with the target of understanding best practices for solving
this problem. We first introduce several complementary evaluation metrics and
present a basic model based on frame-wise action classification. To achieve
better performance, we then investigate two important factors, i.e., the length
of the training clip and clip sampling method. We also explore multi-task
learning strategies by incorporating auxiliary information from two aspects:
the full action representation and the class-agnostic action label. Our
comprehensive experiments uncover the best practices for 3D action
anticipation, and accordingly we propose a novel method with a multi-task loss.
The proposed method considerably outperforms the recent methods and exhibits
the state-of-the-art performance on standard benchmarks.
|
1910.09616 | Siddharth Roheda | Siddharth Roheda, Hamid Krim | Volterra Neural Networks (VNNs) | Conference Version published in AAAI 2020. Journal version under
review at JMLR | null | 10.1609/aaai.v34i07.6870 | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The importance of inference in Machine Learning (ML) has led to an explosive
number of different proposals in ML, and particularly in Deep Learning. In an
attempt to reduce the complexity of Convolutional Neural Networks, we propose a
Volterra filter-inspired Network architecture. This architecture introduces
controlled non-linearities in the form of interactions between the delayed
input samples of data. We propose a cascaded implementation of Volterra
Filtering so as to significantly reduce the number of parameters required to
carry out the same classification task as that of a conventional Neural
Network. We demonstrate an efficient parallel implementation of this Volterra
Neural Network (VNN), along with its remarkable performance while retaining a
relatively simpler and potentially more tractable structure. Furthermore, we
show a rather sophisticated adaptation of this network to nonlinearly fuse the
RGB (spatial) information and the Optical Flow (temporal) information of a
video sequence for action recognition. The proposed approach is evaluated on
UCF-101 and HMDB-51 datasets for action recognition, and is shown to outperform
state of the art CNN approaches.
| [
{
"version": "v1",
"created": "Mon, 21 Oct 2019 19:22:38 GMT"
},
{
"version": "v2",
"created": "Fri, 27 Dec 2019 21:48:35 GMT"
},
{
"version": "v3",
"created": "Thu, 12 Nov 2020 15:10:55 GMT"
},
{
"version": "v4",
"created": "Wed, 10 Nov 2021 14:39:13 GMT"
},
{
"version": "v5",
"created": "Thu, 15 Jun 2023 16:06:40 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Roheda",
"Siddharth",
""
],
[
"Krim",
"Hamid",
""
]
] | TITLE: Volterra Neural Networks (VNNs)
ABSTRACT: The importance of inference in Machine Learning (ML) has led to an explosive
number of different proposals in ML, and particularly in Deep Learning. In an
attempt to reduce the complexity of Convolutional Neural Networks, we propose a
Volterra filter-inspired Network architecture. This architecture introduces
controlled non-linearities in the form of interactions between the delayed
input samples of data. We propose a cascaded implementation of Volterra
Filtering so as to significantly reduce the number of parameters required to
carry out the same classification task as that of a conventional Neural
Network. We demonstrate an efficient parallel implementation of this Volterra
Neural Network (VNN), along with its remarkable performance while retaining a
relatively simpler and potentially more tractable structure. Furthermore, we
show a rather sophisticated adaptation of this network to nonlinearly fuse the
RGB (spatial) information and the Optical Flow (temporal) information of a
video sequence for action recognition. The proposed approach is evaluated on
UCF-101 and HMDB-51 datasets for action recognition, and is shown to outperform
state of the art CNN approaches.
|
2006.16955 | Stanislaw Jastrzebski | Tobiasz Cieplinski, Tomasz Danel, Sabina Podlewska, Stanislaw
Jastrzebski | We Should at Least Be Able to Design Molecules That Dock Well | Published in Journal of Chemical Information and Modeling | null | 10.1021/acs.jcim.2c01355 | null | q-bio.BM cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Designing compounds with desired properties is a key element of the drug
discovery process. However, measuring progress in the field has been
challenging due to the lack of realistic retrospective benchmarks, and the
large cost of prospective validation. To close this gap, we propose a benchmark
based on docking, a popular computational method for assessing molecule binding
to a protein. Concretely, the goal is to generate drug-like molecules that are
scored highly by SMINA, a popular docking software. We observe that popular
graph-based generative models fail to generate molecules with a high docking
score when trained using a realistically sized training set. This suggests a
limitation of the current incarnation of models for de novo drug design.
Finally, we propose a simplified version of the benchmark based on a simpler
scoring function, and show that the tested models are able to partially solve
it. We release the benchmark as an easy to use package available at
https://github.com/cieplinski-tobiasz/smina-docking-benchmark. We hope that our
benchmark will serve as a stepping stone towards the goal of automatically
generating promising drug candidates.
| [
{
"version": "v1",
"created": "Sat, 20 Jun 2020 16:40:56 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Jul 2020 00:30:07 GMT"
},
{
"version": "v3",
"created": "Mon, 28 Dec 2020 08:10:50 GMT"
},
{
"version": "v4",
"created": "Mon, 28 Jun 2021 08:21:45 GMT"
},
{
"version": "v5",
"created": "Tue, 13 Jun 2023 18:15:18 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Cieplinski",
"Tobiasz",
""
],
[
"Danel",
"Tomasz",
""
],
[
"Podlewska",
"Sabina",
""
],
[
"Jastrzebski",
"Stanislaw",
""
]
] | TITLE: We Should at Least Be Able to Design Molecules That Dock Well
ABSTRACT: Designing compounds with desired properties is a key element of the drug
discovery process. However, measuring progress in the field has been
challenging due to the lack of realistic retrospective benchmarks, and the
large cost of prospective validation. To close this gap, we propose a benchmark
based on docking, a popular computational method for assessing molecule binding
to a protein. Concretely, the goal is to generate drug-like molecules that are
scored highly by SMINA, a popular docking software. We observe that popular
graph-based generative models fail to generate molecules with a high docking
score when trained using a realistically sized training set. This suggests a
limitation of the current incarnation of models for de novo drug design.
Finally, we propose a simplified version of the benchmark based on a simpler
scoring function, and show that the tested models are able to partially solve
it. We release the benchmark as an easy to use package available at
https://github.com/cieplinski-tobiasz/smina-docking-benchmark. We hope that our
benchmark will serve as a stepping stone towards the goal of automatically
generating promising drug candidates.
|
2012.15030 | Zhiyuan Chen Dr | Chen ZhiYuan, Olugbenro. O. Selere and Nicholas Lu Chee Seng | Equipment Failure Analysis for Oil and Gas Industry with an Ensemble
Predictive Model | This paper is published at the Fifth International Conference on
Computer Science & Computational Mathematics (ICCSCM 2016) | Lecture Notes in Electrical Engineering,2023 | 10.1007/978-981-19-8406-8_45 | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper aims at improving the classification accuracy of a Support Vector
Machine (SVM) classifier with Sequential Minimal Optimization (SMO) training
algorithm in order to properly classify failure and normal instances from oil
and gas equipment data. Recent applications of failure analysis have made use
of the SVM technique without implementing SMO training algorithm, while in our
study we show that the proposed solution can perform much better when using the
SMO training algorithm. Furthermore, we implement the ensemble approach, which
is a hybrid rule based and neural network classifier to improve the performance
of the SVM classifier (with SMO training algorithm). The optimization study is
as a result of the underperformance of the classifier when dealing with
imbalanced dataset. The selected best performing classifiers are combined
together with SVM classifier (with SMO training algorithm) by using the
stacking ensemble method which is to create an efficient ensemble predictive
model that can handle the issue of imbalanced data. The classification
performance of this predictive model is considerably better than the SVM with
and without SMO training algorithm and many other conventional classifiers.
| [
{
"version": "v1",
"created": "Wed, 30 Dec 2020 04:14:15 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"ZhiYuan",
"Chen",
""
],
[
"Selere",
"Olugbenro. O.",
""
],
[
"Seng",
"Nicholas Lu Chee",
""
]
] | TITLE: Equipment Failure Analysis for Oil and Gas Industry with an Ensemble
Predictive Model
ABSTRACT: This paper aims at improving the classification accuracy of a Support Vector
Machine (SVM) classifier with Sequential Minimal Optimization (SMO) training
algorithm in order to properly classify failure and normal instances from oil
and gas equipment data. Recent applications of failure analysis have made use
of the SVM technique without implementing SMO training algorithm, while in our
study we show that the proposed solution can perform much better when using the
SMO training algorithm. Furthermore, we implement the ensemble approach, which
is a hybrid rule based and neural network classifier to improve the performance
of the SVM classifier (with SMO training algorithm). The optimization study is
as a result of the underperformance of the classifier when dealing with
imbalanced dataset. The selected best performing classifiers are combined
together with SVM classifier (with SMO training algorithm) by using the
stacking ensemble method which is to create an efficient ensemble predictive
model that can handle the issue of imbalanced data. The classification
performance of this predictive model is considerably better than the SVM with
and without SMO training algorithm and many other conventional classifiers.
|
2102.08788 | Ali Burak \"Unal | Ali Burak \"Unal, Nico Pfeifer, Mete Akg\"un | ppAURORA: Privacy Preserving Area Under Receiver Operating
Characteristic and Precision-Recall Curves | Accepted in NSS-SocialSec 2023 | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by/4.0/ | Computing an AUC as a performance measure to compare the quality of different
machine learning models is one of the final steps of many research projects.
Many of these methods are trained on privacy-sensitive data and there are
several different approaches like $\epsilon$-differential privacy, federated
machine learning and cryptography if the datasets cannot be shared or used
jointly at one place for training and/or testing. In this setting, it can also
be a problem to compute the global AUC, since the labels might also contain
privacy-sensitive information. There have been approaches based on
$\epsilon$-differential privacy to address this problem, but to the best of our
knowledge, no exact privacy preserving solution has been introduced. In this
paper, we propose an MPC-based solution, called ppAURORA, with private merging
of individually sorted lists from multiple sources to compute the exact AUC as
one could obtain on the pooled original test samples. With ppAURORA, the
computation of the exact area under precision-recall and receiver operating
characteristic curves is possible even when ties between prediction confidence
values exist. We use ppAURORA to evaluate two different models predicting acute
myeloid leukemia therapy response and heart disease, respectively. We also
assess its scalability via synthetic data experiments. All these experiments
show that we efficiently and privately compute the exact same AUC with both
evaluation metrics as one can obtain on the pooled test samples in plaintext
according to the semi-honest adversary setting.
| [
{
"version": "v1",
"created": "Wed, 17 Feb 2021 14:30:22 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jun 2021 12:17:28 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Jun 2023 16:09:19 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Ünal",
"Ali Burak",
""
],
[
"Pfeifer",
"Nico",
""
],
[
"Akgün",
"Mete",
""
]
] | TITLE: ppAURORA: Privacy Preserving Area Under Receiver Operating
Characteristic and Precision-Recall Curves
ABSTRACT: Computing an AUC as a performance measure to compare the quality of different
machine learning models is one of the final steps of many research projects.
Many of these methods are trained on privacy-sensitive data and there are
several different approaches like $\epsilon$-differential privacy, federated
machine learning and cryptography if the datasets cannot be shared or used
jointly at one place for training and/or testing. In this setting, it can also
be a problem to compute the global AUC, since the labels might also contain
privacy-sensitive information. There have been approaches based on
$\epsilon$-differential privacy to address this problem, but to the best of our
knowledge, no exact privacy preserving solution has been introduced. In this
paper, we propose an MPC-based solution, called ppAURORA, with private merging
of individually sorted lists from multiple sources to compute the exact AUC as
one could obtain on the pooled original test samples. With ppAURORA, the
computation of the exact area under precision-recall and receiver operating
characteristic curves is possible even when ties between prediction confidence
values exist. We use ppAURORA to evaluate two different models predicting acute
myeloid leukemia therapy response and heart disease, respectively. We also
assess its scalability via synthetic data experiments. All these experiments
show that we efficiently and privately compute the exact same AUC with both
evaluation metrics as one can obtain on the pooled test samples in plaintext
according to the semi-honest adversary setting.
|
2104.06722 | Oishik Chatterjee | Oishik Chatterjee, Isha Pandey, Aashish Waikar, Vishwajeet Kumar,
Ganesh Ramakrishnan | WARM: A Weakly (+Semi) Supervised Model for Solving Math word Problems | Accepted in COLING'22 | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Solving math word problems (MWPs) is an important and challenging problem in
natural language processing. Existing approaches to solve MWPs require full
supervision in the form of intermediate equations. However, labeling every MWP
with its corresponding equations is a time-consuming and expensive task. In
order to address this challenge of equation annotation, we propose a weakly
supervised model for solving MWPs by requiring only the final answer as
supervision. We approach this problem by first learning to generate the
equation using the problem description and the final answer, which we
subsequently use to train a supervised MWP solver. We propose and compare
various weakly supervised techniques to learn to generate equations directly
from the problem description and answer. Through extensive experiments, we
demonstrate that without using equations for supervision, our approach achieves
accuracy gains of 4.5% and 32% over the state-of-the-art weakly supervised
approach, on the standard Math23K and AllArith datasets respectively.
Additionally, we curate and release new datasets of roughly 10k MWPs each in
English and in Hindi (a low resource language).These datasets are suitable for
training weakly supervised models. We also present an extension of WARMM to
semi-supervised learning and present further improvements on results, along
with insights.
| [
{
"version": "v1",
"created": "Wed, 14 Apr 2021 09:25:38 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 19:26:15 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Chatterjee",
"Oishik",
""
],
[
"Pandey",
"Isha",
""
],
[
"Waikar",
"Aashish",
""
],
[
"Kumar",
"Vishwajeet",
""
],
[
"Ramakrishnan",
"Ganesh",
""
]
] | TITLE: WARM: A Weakly (+Semi) Supervised Model for Solving Math word Problems
ABSTRACT: Solving math word problems (MWPs) is an important and challenging problem in
natural language processing. Existing approaches to solve MWPs require full
supervision in the form of intermediate equations. However, labeling every MWP
with its corresponding equations is a time-consuming and expensive task. In
order to address this challenge of equation annotation, we propose a weakly
supervised model for solving MWPs by requiring only the final answer as
supervision. We approach this problem by first learning to generate the
equation using the problem description and the final answer, which we
subsequently use to train a supervised MWP solver. We propose and compare
various weakly supervised techniques to learn to generate equations directly
from the problem description and answer. Through extensive experiments, we
demonstrate that without using equations for supervision, our approach achieves
accuracy gains of 4.5% and 32% over the state-of-the-art weakly supervised
approach, on the standard Math23K and AllArith datasets respectively.
Additionally, we curate and release new datasets of roughly 10k MWPs each in
English and in Hindi (a low resource language).These datasets are suitable for
training weakly supervised models. We also present an extension of WARMM to
semi-supervised learning and present further improvements on results, along
with insights.
|
2108.10831 | Chuang Zhu | Xinyu Jia, Chuang Zhu, Minzhen Li, Wenqi Tang, Shengjie Liu, Wenli
Zhou | LLVIP: A Visible-infrared Paired Dataset for Low-light Vision | 10 pages, 11 figures, ICCV workshop | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | It is very challenging for various visual tasks such as image fusion,
pedestrian detection and image-to-image translation in low light conditions due
to the loss of effective target areas. In this case, infrared and visible
images can be used together to provide both rich detail information and
effective target areas. In this paper, we present LLVIP, a visible-infrared
paired dataset for low-light vision. This dataset contains 30976 images, or
15488 pairs, most of which were taken at very dark scenes, and all of the
images are strictly aligned in time and space. Pedestrians in the dataset are
labeled. We compare the dataset with other visible-infrared datasets and
evaluate the performance of some popular visual algorithms including image
fusion, pedestrian detection and image-to-image translation on the dataset. The
experimental results demonstrate the complementary effect of fusion on image
information, and find the deficiency of existing algorithms of the three visual
tasks in very low-light conditions. We believe the LLVIP dataset will
contribute to the community of computer vision by promoting image fusion,
pedestrian detection and image-to-image translation in very low-light
applications. The dataset is being released in
https://bupt-ai-cz.github.io/LLVIP. Raw data is also provided for further
research such as image registration.
| [
{
"version": "v1",
"created": "Tue, 24 Aug 2021 16:29:17 GMT"
},
{
"version": "v2",
"created": "Sun, 17 Oct 2021 14:07:00 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Jun 2022 13:10:34 GMT"
},
{
"version": "v4",
"created": "Wed, 14 Jun 2023 12:14:17 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Jia",
"Xinyu",
""
],
[
"Zhu",
"Chuang",
""
],
[
"Li",
"Minzhen",
""
],
[
"Tang",
"Wenqi",
""
],
[
"Liu",
"Shengjie",
""
],
[
"Zhou",
"Wenli",
""
]
] | TITLE: LLVIP: A Visible-infrared Paired Dataset for Low-light Vision
ABSTRACT: It is very challenging for various visual tasks such as image fusion,
pedestrian detection and image-to-image translation in low light conditions due
to the loss of effective target areas. In this case, infrared and visible
images can be used together to provide both rich detail information and
effective target areas. In this paper, we present LLVIP, a visible-infrared
paired dataset for low-light vision. This dataset contains 30976 images, or
15488 pairs, most of which were taken at very dark scenes, and all of the
images are strictly aligned in time and space. Pedestrians in the dataset are
labeled. We compare the dataset with other visible-infrared datasets and
evaluate the performance of some popular visual algorithms including image
fusion, pedestrian detection and image-to-image translation on the dataset. The
experimental results demonstrate the complementary effect of fusion on image
information, and find the deficiency of existing algorithms of the three visual
tasks in very low-light conditions. We believe the LLVIP dataset will
contribute to the community of computer vision by promoting image fusion,
pedestrian detection and image-to-image translation in very low-light
applications. The dataset is being released in
https://bupt-ai-cz.github.io/LLVIP. Raw data is also provided for further
research such as image registration.
|
2109.10794 | Anthony Caterini | Anthony L. Caterini, Gabriel Loaiza-Ganem | Entropic Issues in Likelihood-Based OOD Detection | Accepted to the NeurIPS 2021 ICBINB Workshop. Extended the result in
expectation to a result with high probability | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep generative models trained by maximum likelihood remain very popular
methods for reasoning about data probabilistically. However, it has been
observed that they can assign higher likelihoods to out-of-distribution (OOD)
data than in-distribution data, thus calling into question the meaning of these
likelihood values. In this work we provide a novel perspective on this
phenomenon, decomposing the average likelihood into a KL divergence term and an
entropy term. We argue that the latter can explain the curious OOD behaviour
mentioned above, suppressing likelihood values on datasets with higher entropy.
Although our idea is simple, we have not seen it explored yet in the
literature. This analysis provides further explanation for the success of OOD
detection methods based on likelihood ratios, as the problematic entropy term
cancels out in expectation. Finally, we discuss how this observation relates to
recent success in OOD detection with manifold-supported models, for which the
above decomposition does not hold directly.
| [
{
"version": "v1",
"created": "Wed, 22 Sep 2021 15:32:57 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 18:02:07 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Caterini",
"Anthony L.",
""
],
[
"Loaiza-Ganem",
"Gabriel",
""
]
] | TITLE: Entropic Issues in Likelihood-Based OOD Detection
ABSTRACT: Deep generative models trained by maximum likelihood remain very popular
methods for reasoning about data probabilistically. However, it has been
observed that they can assign higher likelihoods to out-of-distribution (OOD)
data than in-distribution data, thus calling into question the meaning of these
likelihood values. In this work we provide a novel perspective on this
phenomenon, decomposing the average likelihood into a KL divergence term and an
entropy term. We argue that the latter can explain the curious OOD behaviour
mentioned above, suppressing likelihood values on datasets with higher entropy.
Although our idea is simple, we have not seen it explored yet in the
literature. This analysis provides further explanation for the success of OOD
detection methods based on likelihood ratios, as the problematic entropy term
cancels out in expectation. Finally, we discuss how this observation relates to
recent success in OOD detection with manifold-supported models, for which the
above decomposition does not hold directly.
|
2110.13409 | Jinting Zhu | Jinting Zhu, Julian Jang-Jaccard, Amardeep Singh, Paul A. Watters,
Seyit Camtepe | Task-Aware Meta Learning-based Siamese Neural Network for Classifying
Obfuscated Malware | null | Future Internet 2023, 15(6), 214 | 10.3390/fi15060214 | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | Malware authors apply different techniques of control flow obfuscation, in
order to create new malware variants to avoid detection. Existing Siamese
neural network (SNN)-based malware detection methods fail to correctly classify
different malware families when such obfuscated malware samples are present in
the training dataset, resulting in high false-positive rates. To address this
issue, we propose a novel task-aware few-shot-learning-based Siamese Neural
Network that is resilient against the presence of malware variants affected by
such control flow obfuscation techniques. Using the average entropy features of
each malware family as inputs, in addition to the image features, our model
generates the parameters for the feature layers, to more accurately adjust the
feature embedding for different malware families, each of which has obfuscated
malware variants. In addition, our proposed method can classify malware
classes, even if there are only one or a few training samples available. Our
model utilizes few-shot learning with the extracted features of a pre-trained
network (e.g., VGG-16), to avoid the bias typically associated with a model
trained with a limited number of training samples. Our proposed approach is
highly effective in recognizing unique malware signatures, thus correctly
classifying malware samples that belong to the same malware family, even in the
presence of obfuscated malware variants. Our experimental results, validated by
N-way on N-shot learning, show that our model is highly effective in
classification accuracy, exceeding a rate \textgreater 91\%, compared to other
similar methods.
| [
{
"version": "v1",
"created": "Tue, 26 Oct 2021 04:44:13 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Mar 2022 02:43:01 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Jun 2023 02:30:53 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Zhu",
"Jinting",
""
],
[
"Jang-Jaccard",
"Julian",
""
],
[
"Singh",
"Amardeep",
""
],
[
"Watters",
"Paul A.",
""
],
[
"Camtepe",
"Seyit",
""
]
] | TITLE: Task-Aware Meta Learning-based Siamese Neural Network for Classifying
Obfuscated Malware
ABSTRACT: Malware authors apply different techniques of control flow obfuscation, in
order to create new malware variants to avoid detection. Existing Siamese
neural network (SNN)-based malware detection methods fail to correctly classify
different malware families when such obfuscated malware samples are present in
the training dataset, resulting in high false-positive rates. To address this
issue, we propose a novel task-aware few-shot-learning-based Siamese Neural
Network that is resilient against the presence of malware variants affected by
such control flow obfuscation techniques. Using the average entropy features of
each malware family as inputs, in addition to the image features, our model
generates the parameters for the feature layers, to more accurately adjust the
feature embedding for different malware families, each of which has obfuscated
malware variants. In addition, our proposed method can classify malware
classes, even if there are only one or a few training samples available. Our
model utilizes few-shot learning with the extracted features of a pre-trained
network (e.g., VGG-16), to avoid the bias typically associated with a model
trained with a limited number of training samples. Our proposed approach is
highly effective in recognizing unique malware signatures, thus correctly
classifying malware samples that belong to the same malware family, even in the
presence of obfuscated malware variants. Our experimental results, validated by
N-way on N-shot learning, show that our model is highly effective in
classification accuracy, exceeding a rate \textgreater 91\%, compared to other
similar methods.
|
2111.03837 | Ali Osman Berk \c{S}apc{\i} | Ali Osman Berk Sapci, Oznur Tastan, Reyyan Yeniterzi | Focusing on Potential Named Entities During Active Label Acquisition | 20 pages, 8 figures | Natural Language Engineering, pp. 1-23, 2023 | 10.1017/S1351324923000165 | null | cs.CL cs.IR cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Named entity recognition (NER) aims to identify mentions of named entities in
an unstructured text and classify them into predefined named entity classes.
While deep learning-based pre-trained language models help to achieve good
predictive performances in NER, many domain-specific NER applications still
call for a substantial amount of labeled data. Active learning (AL), a general
framework for the label acquisition problem, has been used for NER tasks to
minimize the annotation cost without sacrificing model performance. However,
the heavily imbalanced class distribution of tokens introduces challenges in
designing effective AL querying methods for NER. We propose several AL sentence
query evaluation functions that pay more attention to potential positive
tokens, and evaluate these proposed functions with both sentence-based and
token-based cost evaluation strategies. We also propose a better data-driven
normalization approach to penalize sentences that are too long or too short.
Our experiments on three datasets from different domains reveal that the
proposed approach reduces the number of annotated tokens while achieving better
or comparable prediction performance with conventional methods.
| [
{
"version": "v1",
"created": "Sat, 6 Nov 2021 09:04:16 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Mar 2023 22:18:01 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Jun 2023 00:32:17 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Sapci",
"Ali Osman Berk",
""
],
[
"Tastan",
"Oznur",
""
],
[
"Yeniterzi",
"Reyyan",
""
]
] | TITLE: Focusing on Potential Named Entities During Active Label Acquisition
ABSTRACT: Named entity recognition (NER) aims to identify mentions of named entities in
an unstructured text and classify them into predefined named entity classes.
While deep learning-based pre-trained language models help to achieve good
predictive performances in NER, many domain-specific NER applications still
call for a substantial amount of labeled data. Active learning (AL), a general
framework for the label acquisition problem, has been used for NER tasks to
minimize the annotation cost without sacrificing model performance. However,
the heavily imbalanced class distribution of tokens introduces challenges in
designing effective AL querying methods for NER. We propose several AL sentence
query evaluation functions that pay more attention to potential positive
tokens, and evaluate these proposed functions with both sentence-based and
token-based cost evaluation strategies. We also propose a better data-driven
normalization approach to penalize sentences that are too long or too short.
Our experiments on three datasets from different domains reveal that the
proposed approach reduces the number of annotated tokens while achieving better
or comparable prediction performance with conventional methods.
|
2111.05850 | Tanja Tornede | Tanja Tornede and Alexander Tornede and Jonas Hanselle and Marcel
Wever and Felix Mohr and Eyke H\"ullermeier | Towards Green Automated Machine Learning: Status Quo and Future
Directions | Published in Journal of Artificial Intelligence Research | null | 10.1613/jair.1.14340 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automated machine learning (AutoML) strives for the automatic configuration
of machine learning algorithms and their composition into an overall (software)
solution - a machine learning pipeline - tailored to the learning task
(dataset) at hand. Over the last decade, AutoML has developed into an
independent research field with hundreds of contributions. At the same time,
AutoML is being criticised for its high resource consumption as many approaches
rely on the (costly) evaluation of many machine learning pipelines, as well as
the expensive large scale experiments across many datasets and approaches. In
the spirit of recent work on Green AI, this paper proposes Green AutoML, a
paradigm to make the whole AutoML process more environmentally friendly.
Therefore, we first elaborate on how to quantify the environmental footprint of
an AutoML tool. Afterward, different strategies on how to design and benchmark
an AutoML tool wrt. their "greenness", i.e. sustainability, are summarized.
Finally, we elaborate on how to be transparent about the environmental
footprint and what kind of research incentives could direct the community into
a more sustainable AutoML research direction. Additionally, we propose a
sustainability checklist to be attached to every AutoML paper featuring all
core aspects of Green AutoML.
| [
{
"version": "v1",
"created": "Wed, 10 Nov 2021 18:57:27 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Jan 2022 08:14:35 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Oct 2022 14:50:01 GMT"
},
{
"version": "v4",
"created": "Tue, 13 Jun 2023 19:49:17 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Tornede",
"Tanja",
""
],
[
"Tornede",
"Alexander",
""
],
[
"Hanselle",
"Jonas",
""
],
[
"Wever",
"Marcel",
""
],
[
"Mohr",
"Felix",
""
],
[
"Hüllermeier",
"Eyke",
""
]
] | TITLE: Towards Green Automated Machine Learning: Status Quo and Future
Directions
ABSTRACT: Automated machine learning (AutoML) strives for the automatic configuration
of machine learning algorithms and their composition into an overall (software)
solution - a machine learning pipeline - tailored to the learning task
(dataset) at hand. Over the last decade, AutoML has developed into an
independent research field with hundreds of contributions. At the same time,
AutoML is being criticised for its high resource consumption as many approaches
rely on the (costly) evaluation of many machine learning pipelines, as well as
the expensive large scale experiments across many datasets and approaches. In
the spirit of recent work on Green AI, this paper proposes Green AutoML, a
paradigm to make the whole AutoML process more environmentally friendly.
Therefore, we first elaborate on how to quantify the environmental footprint of
an AutoML tool. Afterward, different strategies on how to design and benchmark
an AutoML tool wrt. their "greenness", i.e. sustainability, are summarized.
Finally, we elaborate on how to be transparent about the environmental
footprint and what kind of research incentives could direct the community into
a more sustainable AutoML research direction. Additionally, we propose a
sustainability checklist to be attached to every AutoML paper featuring all
core aspects of Green AutoML.
|
2202.02947 | Seyyedali Hosseinalipour | Seyyedali Hosseinalipour, Su Wang, Nicolo Michelusi, Vaneet Aggarwal,
Christopher G. Brinton, David J. Love, Mung Chiang | Parallel Successive Learning for Dynamic Distributed Model Training over
Heterogeneous Wireless Networks | null | null | null | null | cs.LG cs.AI cs.NI cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning (FedL) has emerged as a popular technique for distributing
model training over a set of wireless devices, via iterative local updates (at
devices) and global aggregations (at the server). In this paper, we develop
parallel successive learning (PSL), which expands the FedL architecture along
three dimensions: (i) Network, allowing decentralized cooperation among the
devices via device-to-device (D2D) communications. (ii) Heterogeneity,
interpreted at three levels: (ii-a) Learning: PSL considers heterogeneous
number of stochastic gradient descent iterations with different mini-batch
sizes at the devices; (ii-b) Data: PSL presumes a dynamic environment with data
arrival and departure, where the distributions of local datasets evolve over
time, captured via a new metric for model/concept drift. (ii-c) Device: PSL
considers devices with different computation and communication capabilities.
(iii) Proximity, where devices have different distances to each other and the
access point. PSL considers the realistic scenario where global aggregations
are conducted with idle times in-between them for resource efficiency
improvements, and incorporates data dispersion and model dispersion with local
model condensation into FedL. Our analysis sheds light on the notion of cold
vs. warmed up models, and model inertia in distributed machine learning. We
then propose network-aware dynamic model tracking to optimize the model
learning vs. resource efficiency tradeoff, which we show is an NP-hard
signomial programming problem. We finally solve this problem through proposing
a general optimization solver. Our numerical results reveal new findings on the
interdependencies between the idle times in-between the global aggregations,
model/concept drift, and D2D cooperation configuration.
| [
{
"version": "v1",
"created": "Mon, 7 Feb 2022 05:11:01 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Feb 2022 20:00:55 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Feb 2022 03:54:59 GMT"
},
{
"version": "v4",
"created": "Thu, 17 Feb 2022 19:41:25 GMT"
},
{
"version": "v5",
"created": "Sat, 22 Oct 2022 19:02:30 GMT"
},
{
"version": "v6",
"created": "Wed, 14 Jun 2023 16:50:03 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Hosseinalipour",
"Seyyedali",
""
],
[
"Wang",
"Su",
""
],
[
"Michelusi",
"Nicolo",
""
],
[
"Aggarwal",
"Vaneet",
""
],
[
"Brinton",
"Christopher G.",
""
],
[
"Love",
"David J.",
""
],
[
"Chiang",
"Mung",
""
]
] | TITLE: Parallel Successive Learning for Dynamic Distributed Model Training over
Heterogeneous Wireless Networks
ABSTRACT: Federated learning (FedL) has emerged as a popular technique for distributing
model training over a set of wireless devices, via iterative local updates (at
devices) and global aggregations (at the server). In this paper, we develop
parallel successive learning (PSL), which expands the FedL architecture along
three dimensions: (i) Network, allowing decentralized cooperation among the
devices via device-to-device (D2D) communications. (ii) Heterogeneity,
interpreted at three levels: (ii-a) Learning: PSL considers heterogeneous
number of stochastic gradient descent iterations with different mini-batch
sizes at the devices; (ii-b) Data: PSL presumes a dynamic environment with data
arrival and departure, where the distributions of local datasets evolve over
time, captured via a new metric for model/concept drift. (ii-c) Device: PSL
considers devices with different computation and communication capabilities.
(iii) Proximity, where devices have different distances to each other and the
access point. PSL considers the realistic scenario where global aggregations
are conducted with idle times in-between them for resource efficiency
improvements, and incorporates data dispersion and model dispersion with local
model condensation into FedL. Our analysis sheds light on the notion of cold
vs. warmed up models, and model inertia in distributed machine learning. We
then propose network-aware dynamic model tracking to optimize the model
learning vs. resource efficiency tradeoff, which we show is an NP-hard
signomial programming problem. We finally solve this problem through proposing
a general optimization solver. Our numerical results reveal new findings on the
interdependencies between the idle times in-between the global aggregations,
model/concept drift, and D2D cooperation configuration.
|
2202.05797 | Christopher Jung | Pranjal Awasthi, Christopher Jung, Jamie Morgenstern | Distributionally Robust Data Join | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Suppose we are given two datasets: a labeled dataset and unlabeled dataset
which also has additional auxiliary features not present in the first dataset.
What is the most principled way to use these datasets together to construct a
predictor?
The answer should depend upon whether these datasets are generated by the
same or different distributions over their mutual feature sets, and how similar
the test distribution will be to either of those distributions. In many
applications, the two datasets will likely follow different distributions, but
both may be close to the test distribution. We introduce the problem of
building a predictor which minimizes the maximum loss over all probability
distributions over the original features, auxiliary features, and binary
labels, whose Wasserstein distance is $r_1$ away from the empirical
distribution over the labeled dataset and $r_2$ away from that of the unlabeled
dataset. This can be thought of as a generalization of distributionally robust
optimization (DRO), which allows for two data sources, one of which is
unlabeled and may contain auxiliary features.
| [
{
"version": "v1",
"created": "Fri, 11 Feb 2022 17:46:24 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 00:02:37 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Awasthi",
"Pranjal",
""
],
[
"Jung",
"Christopher",
""
],
[
"Morgenstern",
"Jamie",
""
]
] | TITLE: Distributionally Robust Data Join
ABSTRACT: Suppose we are given two datasets: a labeled dataset and unlabeled dataset
which also has additional auxiliary features not present in the first dataset.
What is the most principled way to use these datasets together to construct a
predictor?
The answer should depend upon whether these datasets are generated by the
same or different distributions over their mutual feature sets, and how similar
the test distribution will be to either of those distributions. In many
applications, the two datasets will likely follow different distributions, but
both may be close to the test distribution. We introduce the problem of
building a predictor which minimizes the maximum loss over all probability
distributions over the original features, auxiliary features, and binary
labels, whose Wasserstein distance is $r_1$ away from the empirical
distribution over the labeled dataset and $r_2$ away from that of the unlabeled
dataset. This can be thought of as a generalization of distributionally robust
optimization (DRO), which allows for two data sources, one of which is
unlabeled and may contain auxiliary features.
|
2202.07135 | Jinyang Yuan | Jinyang Yuan, Tonglin Chen, Bin Li, Xiangyang Xue | Compositional Scene Representation Learning via Reconstruction: A Survey | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual scenes are composed of visual concepts and have the property of
combinatorial explosion. An important reason for humans to efficiently learn
from diverse visual scenes is the ability of compositional perception, and it
is desirable for artificial intelligence to have similar abilities.
Compositional scene representation learning is a task that enables such
abilities. In recent years, various methods have been proposed to apply deep
neural networks, which have been proven to be advantageous in representation
learning, to learn compositional scene representations via reconstruction,
advancing this research direction into the deep learning era. Learning via
reconstruction is advantageous because it may utilize massive unlabeled data
and avoid costly and laborious data annotation. In this survey, we first
outline the current progress on reconstruction-based compositional scene
representation learning with deep neural networks, including development
history and categorizations of existing methods from the perspectives of the
modeling of visual scenes and the inference of scene representations; then
provide benchmarks, including an open source toolbox to reproduce the benchmark
experiments, of representative methods that consider the most extensively
studied problem setting and form the foundation for other methods; and finally
discuss the limitations of existing methods and future directions of this
research topic.
| [
{
"version": "v1",
"created": "Tue, 15 Feb 2022 02:14:05 GMT"
},
{
"version": "v2",
"created": "Fri, 3 Jun 2022 04:02:33 GMT"
},
{
"version": "v3",
"created": "Thu, 9 Feb 2023 12:50:18 GMT"
},
{
"version": "v4",
"created": "Wed, 14 Jun 2023 16:25:03 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Yuan",
"Jinyang",
""
],
[
"Chen",
"Tonglin",
""
],
[
"Li",
"Bin",
""
],
[
"Xue",
"Xiangyang",
""
]
] | TITLE: Compositional Scene Representation Learning via Reconstruction: A Survey
ABSTRACT: Visual scenes are composed of visual concepts and have the property of
combinatorial explosion. An important reason for humans to efficiently learn
from diverse visual scenes is the ability of compositional perception, and it
is desirable for artificial intelligence to have similar abilities.
Compositional scene representation learning is a task that enables such
abilities. In recent years, various methods have been proposed to apply deep
neural networks, which have been proven to be advantageous in representation
learning, to learn compositional scene representations via reconstruction,
advancing this research direction into the deep learning era. Learning via
reconstruction is advantageous because it may utilize massive unlabeled data
and avoid costly and laborious data annotation. In this survey, we first
outline the current progress on reconstruction-based compositional scene
representation learning with deep neural networks, including development
history and categorizations of existing methods from the perspectives of the
modeling of visual scenes and the inference of scene representations; then
provide benchmarks, including an open source toolbox to reproduce the benchmark
experiments, of representative methods that consider the most extensively
studied problem setting and form the foundation for other methods; and finally
discuss the limitations of existing methods and future directions of this
research topic.
|
2203.15380 | Wei Li | Wei Li, Xing Wang, Xin Xia, Jie Wu, Jiashi Li, Xuefeng Xiao, Min
Zheng, Shiping Wen | SepViT: Separable Vision Transformer | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision Transformers have witnessed prevailing success in a series of vision
tasks. However, these Transformers often rely on extensive computational costs
to achieve high performance, which is burdensome to deploy on
resource-constrained devices. To alleviate this issue, we draw lessons from
depthwise separable convolution and imitate its ideology to design an efficient
Transformer backbone, i.e., Separable Vision Transformer, abbreviated as
SepViT. SepViT helps to carry out the local-global information interaction
within and among the windows in sequential order via a depthwise separable
self-attention. The novel window token embedding and grouped self-attention are
employed to compute the attention relationship among windows with negligible
cost and establish long-range visual interactions across multiple windows,
respectively. Extensive experiments on general-purpose vision benchmarks
demonstrate that SepViT can achieve a state-of-the-art trade-off between
performance and latency. Among them, SepViT achieves 84.2% top-1 accuracy on
ImageNet-1K classification while decreasing the latency by 40%, compared to the
ones with similar accuracy (e.g., CSWin). Furthermore, SepViT achieves 51.0%
mIoU on ADE20K semantic segmentation task, 47.9 AP on the RetinaNet-based COCO
detection task, 49.4 box AP and 44.6 mask AP on Mask R-CNN-based COCO object
detection and instance segmentation tasks.
| [
{
"version": "v1",
"created": "Tue, 29 Mar 2022 09:20:01 GMT"
},
{
"version": "v2",
"created": "Sun, 3 Apr 2022 01:36:46 GMT"
},
{
"version": "v3",
"created": "Sat, 7 May 2022 08:20:10 GMT"
},
{
"version": "v4",
"created": "Thu, 15 Jun 2023 16:37:26 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Li",
"Wei",
""
],
[
"Wang",
"Xing",
""
],
[
"Xia",
"Xin",
""
],
[
"Wu",
"Jie",
""
],
[
"Li",
"Jiashi",
""
],
[
"Xiao",
"Xuefeng",
""
],
[
"Zheng",
"Min",
""
],
[
"Wen",
"Shiping",
""
]
] | TITLE: SepViT: Separable Vision Transformer
ABSTRACT: Vision Transformers have witnessed prevailing success in a series of vision
tasks. However, these Transformers often rely on extensive computational costs
to achieve high performance, which is burdensome to deploy on
resource-constrained devices. To alleviate this issue, we draw lessons from
depthwise separable convolution and imitate its ideology to design an efficient
Transformer backbone, i.e., Separable Vision Transformer, abbreviated as
SepViT. SepViT helps to carry out the local-global information interaction
within and among the windows in sequential order via a depthwise separable
self-attention. The novel window token embedding and grouped self-attention are
employed to compute the attention relationship among windows with negligible
cost and establish long-range visual interactions across multiple windows,
respectively. Extensive experiments on general-purpose vision benchmarks
demonstrate that SepViT can achieve a state-of-the-art trade-off between
performance and latency. Among them, SepViT achieves 84.2% top-1 accuracy on
ImageNet-1K classification while decreasing the latency by 40%, compared to the
ones with similar accuracy (e.g., CSWin). Furthermore, SepViT achieves 51.0%
mIoU on ADE20K semantic segmentation task, 47.9 AP on the RetinaNet-based COCO
detection task, 49.4 box AP and 44.6 mask AP on Mask R-CNN-based COCO object
detection and instance segmentation tasks.
|
2205.09256 | Liangke Gui | Liangke Gui, Yingshan Chang, Qiuyuan Huang, Subhojit Som, Alex
Hauptmann, Jianfeng Gao, Yonatan Bisk | Training Vision-Language Transformers from Captions | null | null | null | null | cs.CV cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-Language Transformers can be learned without low-level human labels
(e.g. class labels, bounding boxes, etc). Existing work, whether explicitly
utilizing bounding boxes or patches, assumes that the visual backbone must
first be trained on ImageNet class prediction before being integrated into a
multimodal linguistic pipeline. We show that this is not necessary and
introduce a new model Vision-Language from Captions (VLC) built on top of
Masked Auto-Encoders that does not require this supervision. In fact, in a
head-to-head comparison between ViLT, the current state-of-the-art patch-based
vision-language transformer which is pretrained with supervised object
classification, and our model, VLC, we find that our approach 1. outperforms
ViLT on standard benchmarks, 2. provides more interpretable and intuitive patch
visualizations, and 3. is competitive with many larger models that utilize ROIs
trained on annotated bounding-boxes.
| [
{
"version": "v1",
"created": "Thu, 19 May 2022 00:19:48 GMT"
},
{
"version": "v2",
"created": "Tue, 3 Jan 2023 17:46:37 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Jun 2023 17:37:46 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Gui",
"Liangke",
""
],
[
"Chang",
"Yingshan",
""
],
[
"Huang",
"Qiuyuan",
""
],
[
"Som",
"Subhojit",
""
],
[
"Hauptmann",
"Alex",
""
],
[
"Gao",
"Jianfeng",
""
],
[
"Bisk",
"Yonatan",
""
]
] | TITLE: Training Vision-Language Transformers from Captions
ABSTRACT: Vision-Language Transformers can be learned without low-level human labels
(e.g. class labels, bounding boxes, etc). Existing work, whether explicitly
utilizing bounding boxes or patches, assumes that the visual backbone must
first be trained on ImageNet class prediction before being integrated into a
multimodal linguistic pipeline. We show that this is not necessary and
introduce a new model Vision-Language from Captions (VLC) built on top of
Masked Auto-Encoders that does not require this supervision. In fact, in a
head-to-head comparison between ViLT, the current state-of-the-art patch-based
vision-language transformer which is pretrained with supervised object
classification, and our model, VLC, we find that our approach 1. outperforms
ViLT on standard benchmarks, 2. provides more interpretable and intuitive patch
visualizations, and 3. is competitive with many larger models that utilize ROIs
trained on annotated bounding-boxes.
|
2205.09864 | Nigel Fernandez | Nigel Fernandez, Aritra Ghosh, Naiming Liu, Zichao Wang, Beno\^it
Choffin, Richard Baraniuk, Andrew Lan | Automated Scoring for Reading Comprehension via In-context BERT Tuning | Published as a conference paper at AIED 2022. A grand prize-winner
for the NAEP AS Challenge. Code available at:
https://github.com/ni9elf/automated-scoring | null | null | null | cs.LG cs.AI cs.CY | http://creativecommons.org/licenses/by/4.0/ | Automated scoring of open-ended student responses has the potential to
significantly reduce human grader effort. Recent advances in automated scoring
often leverage textual representations based on pre-trained language models
such as BERT and GPT as input to scoring models. Most existing approaches train
a separate model for each item/question, which is suitable for scenarios such
as essay scoring where items can be quite different from one another. However,
these approaches have two limitations: 1) they fail to leverage item linkage
for scenarios such as reading comprehension where multiple items may share a
reading passage; 2) they are not scalable since storing one model per item
becomes difficult when models have a large number of parameters. In this paper,
we report our (grand prize-winning) solution to the National Assessment of
Education Progress (NAEP) automated scoring challenge for reading
comprehension. Our approach, in-context BERT fine-tuning, produces a single
shared scoring model for all items with a carefully-designed input structure to
provide contextual information on each item. We demonstrate the effectiveness
of our approach via local evaluations using the training dataset provided by
the challenge. We also discuss the biases, common error types, and limitations
of our approach.
| [
{
"version": "v1",
"created": "Thu, 19 May 2022 21:16:15 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 04:37:17 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Fernandez",
"Nigel",
""
],
[
"Ghosh",
"Aritra",
""
],
[
"Liu",
"Naiming",
""
],
[
"Wang",
"Zichao",
""
],
[
"Choffin",
"Benoît",
""
],
[
"Baraniuk",
"Richard",
""
],
[
"Lan",
"Andrew",
""
]
] | TITLE: Automated Scoring for Reading Comprehension via In-context BERT Tuning
ABSTRACT: Automated scoring of open-ended student responses has the potential to
significantly reduce human grader effort. Recent advances in automated scoring
often leverage textual representations based on pre-trained language models
such as BERT and GPT as input to scoring models. Most existing approaches train
a separate model for each item/question, which is suitable for scenarios such
as essay scoring where items can be quite different from one another. However,
these approaches have two limitations: 1) they fail to leverage item linkage
for scenarios such as reading comprehension where multiple items may share a
reading passage; 2) they are not scalable since storing one model per item
becomes difficult when models have a large number of parameters. In this paper,
we report our (grand prize-winning) solution to the National Assessment of
Education Progress (NAEP) automated scoring challenge for reading
comprehension. Our approach, in-context BERT fine-tuning, produces a single
shared scoring model for all items with a carefully-designed input structure to
provide contextual information on each item. We demonstrate the effectiveness
of our approach via local evaluations using the training dataset provided by
the challenge. We also discuss the biases, common error types, and limitations
of our approach.
|
2205.12386 | Aidan San | Aidan San, Yuan Zhuang, Jan Bakus, Colin Lockard, David Ciemiewicz,
Sandeep Atluri, Yangfeng Ji, Kevin Small, Heba Elfardy | PLAtE: A Large-scale Dataset for List Page Web Extraction | Accepted to ACL Industry Track 2023 | null | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, neural models have been leveraged to significantly improve the
performance of information extraction from semi-structured websites. However, a
barrier for continued progress is the small number of datasets large enough to
train these models. In this work, we introduce the PLAtE (Pages of Lists
Attribute Extraction) benchmark dataset as a challenging new web extraction
task. PLAtE focuses on shopping data, specifically extractions from product
review pages with multiple items encompassing the tasks of: (1) finding
product-list segmentation boundaries and (2) extracting attributes for each
product. PLAtE is composed of 52, 898 items collected from 6, 694 pages and
156, 014 attributes, making it the first largescale list page web extraction
dataset. We use a multi-stage approach to collect and annotate the dataset and
adapt three state-of-the-art web extraction models to the two tasks comparing
their strengths and weaknesses both quantitatively and qualitatively.
| [
{
"version": "v1",
"created": "Tue, 24 May 2022 22:26:58 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 17:06:49 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"San",
"Aidan",
""
],
[
"Zhuang",
"Yuan",
""
],
[
"Bakus",
"Jan",
""
],
[
"Lockard",
"Colin",
""
],
[
"Ciemiewicz",
"David",
""
],
[
"Atluri",
"Sandeep",
""
],
[
"Ji",
"Yangfeng",
""
],
[
"Small",
"Kevin",
""
],
[
"Elfardy",
"Heba",
""
]
] | TITLE: PLAtE: A Large-scale Dataset for List Page Web Extraction
ABSTRACT: Recently, neural models have been leveraged to significantly improve the
performance of information extraction from semi-structured websites. However, a
barrier for continued progress is the small number of datasets large enough to
train these models. In this work, we introduce the PLAtE (Pages of Lists
Attribute Extraction) benchmark dataset as a challenging new web extraction
task. PLAtE focuses on shopping data, specifically extractions from product
review pages with multiple items encompassing the tasks of: (1) finding
product-list segmentation boundaries and (2) extracting attributes for each
product. PLAtE is composed of 52, 898 items collected from 6, 694 pages and
156, 014 attributes, making it the first largescale list page web extraction
dataset. We use a multi-stage approach to collect and annotate the dataset and
adapt three state-of-the-art web extraction models to the two tasks comparing
their strengths and weaknesses both quantitatively and qualitatively.
|
2205.14377 | Kai Hu | Kai Hu, Yu Liu, Renhe Liu, Wei Lu, Gang Yu, Bin Fu | Enhancing Quality of Pose-varied Face Restoration with Local Weak
Feature Sensing and GAN Prior | pdfLaTeX 2021, 11 pages with 15 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Facial semantic guidance (including facial landmarks, facial heatmaps, and
facial parsing maps) and facial generative adversarial networks (GAN) prior
have been widely used in blind face restoration (BFR) in recent years. Although
existing BFR methods have achieved good performance in ordinary cases, these
solutions have limited resilience when applied to face images with serious
degradation and pose-varied (e.g., looking right, looking left, laughing, etc.)
in real-world scenarios. In this work, we propose a well-designed blind face
restoration network with generative facial prior. The proposed network is
mainly comprised of an asymmetric codec and a StyleGAN2 prior network. In the
asymmetric codec, we adopt a mixed multi-path residual block (MMRB) to
gradually extract weak texture features of input images, which can better
preserve the original facial features and avoid excessive fantasy. The MMRB can
also be plug-and-play in other networks. Furthermore, thanks to the affluent
and diverse facial priors of the StyleGAN2 model, we adopt it as the primary
generator network in our proposed method and specially design a novel
self-supervised training strategy to fit the distribution closer to the target
and flexibly restore natural and realistic facial details. Extensive
experiments on synthetic and real-world datasets demonstrate that our model
performs superior to the prior art for face restoration and face
super-resolution tasks.
| [
{
"version": "v1",
"created": "Sat, 28 May 2022 09:23:48 GMT"
},
{
"version": "v2",
"created": "Fri, 24 Jun 2022 08:03:57 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Jun 2023 02:24:58 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Hu",
"Kai",
""
],
[
"Liu",
"Yu",
""
],
[
"Liu",
"Renhe",
""
],
[
"Lu",
"Wei",
""
],
[
"Yu",
"Gang",
""
],
[
"Fu",
"Bin",
""
]
] | TITLE: Enhancing Quality of Pose-varied Face Restoration with Local Weak
Feature Sensing and GAN Prior
ABSTRACT: Facial semantic guidance (including facial landmarks, facial heatmaps, and
facial parsing maps) and facial generative adversarial networks (GAN) prior
have been widely used in blind face restoration (BFR) in recent years. Although
existing BFR methods have achieved good performance in ordinary cases, these
solutions have limited resilience when applied to face images with serious
degradation and pose-varied (e.g., looking right, looking left, laughing, etc.)
in real-world scenarios. In this work, we propose a well-designed blind face
restoration network with generative facial prior. The proposed network is
mainly comprised of an asymmetric codec and a StyleGAN2 prior network. In the
asymmetric codec, we adopt a mixed multi-path residual block (MMRB) to
gradually extract weak texture features of input images, which can better
preserve the original facial features and avoid excessive fantasy. The MMRB can
also be plug-and-play in other networks. Furthermore, thanks to the affluent
and diverse facial priors of the StyleGAN2 model, we adopt it as the primary
generator network in our proposed method and specially design a novel
self-supervised training strategy to fit the distribution closer to the target
and flexibly restore natural and realistic facial details. Extensive
experiments on synthetic and real-world datasets demonstrate that our model
performs superior to the prior art for face restoration and face
super-resolution tasks.
|
2206.02998 | Cheng-Yu Fang | Cheng-Yu Fang, Xian-Feng Han | Joint Geometric-Semantic Driven Character Line Drawing Generation | Published in ICMR '23: Proceedings of the 2023 ACM International
Conference on Multimedia Retrieval, June 2023 | In Proceedings of ICMR '23. Association for Computing Machinery,
New York, NY, USA, 226-233 (2023) | 10.1145/3591106.3592216 | null | cs.MM | http://creativecommons.org/licenses/by/4.0/ | Character line drawing synthesis can be formulated as a special case of
image-to-image translation problem that automatically manipulates the
photo-to-line drawing style transformation. In this paper, we present the first
generative adversarial network based end-to-end trainable translation
architecture, dubbed P2LDGAN, for automatic generation of high-quality
character drawings from input photos/images. The core component of our approach
is the joint geometric-semantic driven generator, which uses our well-designed
cross-scale dense skip connections framework to embed learned geometric and
semantic information for generating delicate line drawings. In order to support
the evaluation of our model, we release a new dataset including 1,532
well-matched pairs of freehand character line drawings as well as corresponding
character images/photos, where these line drawings with diverse styles are
manually drawn by skilled artists. Extensive experiments on our introduced
dataset demonstrate the superior performance of our proposed models against the
state-of-the-art approaches in terms of quantitative, qualitative and human
evaluations. Our code, models and dataset will be available at Github.
| [
{
"version": "v1",
"created": "Tue, 7 Jun 2022 04:16:36 GMT"
},
{
"version": "v2",
"created": "Tue, 20 Sep 2022 03:41:34 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Jun 2023 09:32:24 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Fang",
"Cheng-Yu",
""
],
[
"Han",
"Xian-Feng",
""
]
] | TITLE: Joint Geometric-Semantic Driven Character Line Drawing Generation
ABSTRACT: Character line drawing synthesis can be formulated as a special case of
image-to-image translation problem that automatically manipulates the
photo-to-line drawing style transformation. In this paper, we present the first
generative adversarial network based end-to-end trainable translation
architecture, dubbed P2LDGAN, for automatic generation of high-quality
character drawings from input photos/images. The core component of our approach
is the joint geometric-semantic driven generator, which uses our well-designed
cross-scale dense skip connections framework to embed learned geometric and
semantic information for generating delicate line drawings. In order to support
the evaluation of our model, we release a new dataset including 1,532
well-matched pairs of freehand character line drawings as well as corresponding
character images/photos, where these line drawings with diverse styles are
manually drawn by skilled artists. Extensive experiments on our introduced
dataset demonstrate the superior performance of our proposed models against the
state-of-the-art approaches in terms of quantitative, qualitative and human
evaluations. Our code, models and dataset will be available at Github.
|
2206.03617 | Virendra Marathe | Virendra J. Marathe and Pallika Kanani and Daniel W. Peterson and Guy
Steele Jr | Subject Granular Differential Privacy in Federated Learning | null | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper considers subject level privacy in the FL setting, where a subject
is an individual whose private information is embodied by several data items
either confined within a single federation user or distributed across multiple
federation users. We propose two new algorithms that enforce subject level DP
at each federation user locally. Our first algorithm, called LocalGroupDP, is a
straightforward application of group differential privacy in the popular DP-SGD
algorithm. Our second algorithm is based on a novel idea of hierarchical
gradient averaging (HiGradAvgDP) for subjects participating in a training
mini-batch. We also show that user level Local Differential Privacy (LDP)
naturally guarantees subject level DP. We observe the problem of horizontal
composition of subject level privacy loss in FL - subject level privacy loss
incurred at individual users composes across the federation. We formally prove
the subject level DP guarantee for our algorithms, and also show their effect
on model utility loss. Our empirical evaluation on FEMNIST and Shakespeare
datasets shows that LocalGroupDP delivers the best performance among our
algorithms. However, its model utility lags behind that of models trained using
a DP-SGD based algorithm that provides a weaker item level privacy guarantee.
Privacy loss amplification due to subject sampling fractions and horizontal
composition remain key challenges for model utility.
| [
{
"version": "v1",
"created": "Tue, 7 Jun 2022 23:54:36 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 15:37:11 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Marathe",
"Virendra J.",
""
],
[
"Kanani",
"Pallika",
""
],
[
"Peterson",
"Daniel W.",
""
],
[
"Steele",
"Guy",
"Jr"
]
] | TITLE: Subject Granular Differential Privacy in Federated Learning
ABSTRACT: This paper considers subject level privacy in the FL setting, where a subject
is an individual whose private information is embodied by several data items
either confined within a single federation user or distributed across multiple
federation users. We propose two new algorithms that enforce subject level DP
at each federation user locally. Our first algorithm, called LocalGroupDP, is a
straightforward application of group differential privacy in the popular DP-SGD
algorithm. Our second algorithm is based on a novel idea of hierarchical
gradient averaging (HiGradAvgDP) for subjects participating in a training
mini-batch. We also show that user level Local Differential Privacy (LDP)
naturally guarantees subject level DP. We observe the problem of horizontal
composition of subject level privacy loss in FL - subject level privacy loss
incurred at individual users composes across the federation. We formally prove
the subject level DP guarantee for our algorithms, and also show their effect
on model utility loss. Our empirical evaluation on FEMNIST and Shakespeare
datasets shows that LocalGroupDP delivers the best performance among our
algorithms. However, its model utility lags behind that of models trained using
a DP-SGD based algorithm that provides a weaker item level privacy guarantee.
Privacy loss amplification due to subject sampling fractions and horizontal
composition remain key challenges for model utility.
|
2207.03128 | Qijian Zhang | Qijian Zhang, Junhui Hou, Yue Qian | PointMCD: Boosting Deep Point Cloud Encoders via Multi-view Cross-modal
Distillation for 3D Shape Recognition | Accepted to TMM | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As two fundamental representation modalities of 3D objects, 3D point clouds
and multi-view 2D images record shape information from different domains of
geometric structures and visual appearances. In the current deep learning era,
remarkable progress in processing such two data modalities has been achieved
through respectively customizing compatible 3D and 2D network architectures.
However, unlike multi-view image-based 2D visual modeling paradigms, which have
shown leading performance in several common 3D shape recognition benchmarks,
point cloud-based 3D geometric modeling paradigms are still highly limited by
insufficient learning capacity, due to the difficulty of extracting
discriminative features from irregular geometric signals. In this paper, we
explore the possibility of boosting deep 3D point cloud encoders by
transferring visual knowledge extracted from deep 2D image encoders under a
standard teacher-student distillation workflow. Generally, we propose PointMCD,
a unified multi-view cross-modal distillation architecture, including a
pretrained deep image encoder as the teacher and a deep point encoder as the
student. To perform heterogeneous feature alignment between 2D visual and 3D
geometric domains, we further investigate visibility-aware feature projection
(VAFP), by which point-wise embeddings are reasonably aggregated into
view-specific geometric descriptors. By pair-wisely aligning multi-view visual
and geometric descriptors, we can obtain more powerful deep point encoders
without exhausting and complicated network modification. Experiments on 3D
shape classification, part segmentation, and unsupervised learning strongly
validate the effectiveness of our method. The code and data will be publicly
available at https://github.com/keeganhk/PointMCD.
| [
{
"version": "v1",
"created": "Thu, 7 Jul 2022 07:23:20 GMT"
},
{
"version": "v2",
"created": "Thu, 25 Aug 2022 04:04:04 GMT"
},
{
"version": "v3",
"created": "Thu, 13 Apr 2023 09:44:16 GMT"
},
{
"version": "v4",
"created": "Thu, 15 Jun 2023 06:21:09 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Zhang",
"Qijian",
""
],
[
"Hou",
"Junhui",
""
],
[
"Qian",
"Yue",
""
]
] | TITLE: PointMCD: Boosting Deep Point Cloud Encoders via Multi-view Cross-modal
Distillation for 3D Shape Recognition
ABSTRACT: As two fundamental representation modalities of 3D objects, 3D point clouds
and multi-view 2D images record shape information from different domains of
geometric structures and visual appearances. In the current deep learning era,
remarkable progress in processing such two data modalities has been achieved
through respectively customizing compatible 3D and 2D network architectures.
However, unlike multi-view image-based 2D visual modeling paradigms, which have
shown leading performance in several common 3D shape recognition benchmarks,
point cloud-based 3D geometric modeling paradigms are still highly limited by
insufficient learning capacity, due to the difficulty of extracting
discriminative features from irregular geometric signals. In this paper, we
explore the possibility of boosting deep 3D point cloud encoders by
transferring visual knowledge extracted from deep 2D image encoders under a
standard teacher-student distillation workflow. Generally, we propose PointMCD,
a unified multi-view cross-modal distillation architecture, including a
pretrained deep image encoder as the teacher and a deep point encoder as the
student. To perform heterogeneous feature alignment between 2D visual and 3D
geometric domains, we further investigate visibility-aware feature projection
(VAFP), by which point-wise embeddings are reasonably aggregated into
view-specific geometric descriptors. By pair-wisely aligning multi-view visual
and geometric descriptors, we can obtain more powerful deep point encoders
without exhausting and complicated network modification. Experiments on 3D
shape classification, part segmentation, and unsupervised learning strongly
validate the effectiveness of our method. The code and data will be publicly
available at https://github.com/keeganhk/PointMCD.
|
2208.09717 | Jian Peng Yang | Jianpeng Yang, Yuhang Niu, Xuemei Xie, Guangming Shi | Learning Primitive-aware Discriminative Representations for Few-shot
Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-shot learning (FSL) aims to learn a classifier that can be easily adapted
to recognize novel classes with only a few labeled examples. Some recent work
about FSL has yielded promising classification performance, where the
image-level feature is used to calculate the similarity among samples for
classification. However, the image-level feature ignores abundant fine-grained
and structural in-formation of objects that may be transferable and consistent
between seen and unseen classes. How can humans easily identify novel classes
with several sam-ples? Some study from cognitive science argues that humans can
recognize novel categories through primitives. Although base and novel
categories are non-overlapping, they can share some primitives in common.
Inspired by above re-search, we propose a Primitive Mining and Reasoning
Network (PMRN) to learn primitive-aware representations based on metric-based
FSL model. Concretely, we first add Self-supervision Jigsaw task (SSJ) for
feature extractor parallelly, guiding the model to encode visual pattern
corresponding to object parts into fea-ture channels. To further mine
discriminative representations, an Adaptive Chan-nel Grouping (ACG) method is
applied to cluster and weight spatially and se-mantically related visual
patterns to generate a group of visual primitives. To fur-ther enhance the
discriminability and transferability of primitives, we propose a visual
primitive Correlation Reasoning Network (CRN) based on graph convolu-tional
network to learn abundant structural information and internal correlation among
primitives. Finally, a primitive-level metric is conducted for classification
in a meta-task based on episodic training strategy. Extensive experiments show
that our method achieves state-of-the-art results on six standard benchmarks.
| [
{
"version": "v1",
"created": "Sat, 20 Aug 2022 16:22:22 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 16:54:31 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Yang",
"Jianpeng",
""
],
[
"Niu",
"Yuhang",
""
],
[
"Xie",
"Xuemei",
""
],
[
"Shi",
"Guangming",
""
]
] | TITLE: Learning Primitive-aware Discriminative Representations for Few-shot
Learning
ABSTRACT: Few-shot learning (FSL) aims to learn a classifier that can be easily adapted
to recognize novel classes with only a few labeled examples. Some recent work
about FSL has yielded promising classification performance, where the
image-level feature is used to calculate the similarity among samples for
classification. However, the image-level feature ignores abundant fine-grained
and structural in-formation of objects that may be transferable and consistent
between seen and unseen classes. How can humans easily identify novel classes
with several sam-ples? Some study from cognitive science argues that humans can
recognize novel categories through primitives. Although base and novel
categories are non-overlapping, they can share some primitives in common.
Inspired by above re-search, we propose a Primitive Mining and Reasoning
Network (PMRN) to learn primitive-aware representations based on metric-based
FSL model. Concretely, we first add Self-supervision Jigsaw task (SSJ) for
feature extractor parallelly, guiding the model to encode visual pattern
corresponding to object parts into fea-ture channels. To further mine
discriminative representations, an Adaptive Chan-nel Grouping (ACG) method is
applied to cluster and weight spatially and se-mantically related visual
patterns to generate a group of visual primitives. To fur-ther enhance the
discriminability and transferability of primitives, we propose a visual
primitive Correlation Reasoning Network (CRN) based on graph convolu-tional
network to learn abundant structural information and internal correlation among
primitives. Finally, a primitive-level metric is conducted for classification
in a meta-task based on episodic training strategy. Extensive experiments show
that our method achieves state-of-the-art results on six standard benchmarks.
|
2209.04502 | Christopher Bellman | David Barrera, Christopher Bellman, Paul C. van Oorschot | A Close Look at a Systematic Method for Analyzing Sets of Security
Advice | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We carry out a detailed analysis of the security advice coding method
(SAcoding) of Barrera et al. (2021), which is designed to analyze security
advice in the sense of measuring actionability and categorizing advice items as
practices, policies, principles, or outcomes. The main part of our analysis
explores the extent to which a second coder's assignment of codes to advice
items agrees with that of a first, for a dataset of 1013 security advice items
nominally addressing Internet of Things devices. More broadly, we seek a deeper
understanding of the soundness and utility of the SAcoding method, and the
degree to which it meets the design goal of reducing subjectivity in assigning
codes to security advice items. Our analysis results in suggestions for
modifications to the coding tree methodology, and some recommendations. We
believe the coding tree approach may be of interest for analysis of qualitative
data beyond security advice datasets alone.
| [
{
"version": "v1",
"created": "Fri, 9 Sep 2022 19:59:02 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 18:16:35 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Barrera",
"David",
""
],
[
"Bellman",
"Christopher",
""
],
[
"van Oorschot",
"Paul C.",
""
]
] | TITLE: A Close Look at a Systematic Method for Analyzing Sets of Security
Advice
ABSTRACT: We carry out a detailed analysis of the security advice coding method
(SAcoding) of Barrera et al. (2021), which is designed to analyze security
advice in the sense of measuring actionability and categorizing advice items as
practices, policies, principles, or outcomes. The main part of our analysis
explores the extent to which a second coder's assignment of codes to advice
items agrees with that of a first, for a dataset of 1013 security advice items
nominally addressing Internet of Things devices. More broadly, we seek a deeper
understanding of the soundness and utility of the SAcoding method, and the
degree to which it meets the design goal of reducing subjectivity in assigning
codes to security advice items. Our analysis results in suggestions for
modifications to the coding tree methodology, and some recommendations. We
believe the coding tree approach may be of interest for analysis of qualitative
data beyond security advice datasets alone.
|
2209.10326 | Hao Li | Hao Li, Jinfa Huang, Peng Jin, Guoli Song, Qi Wu, Jie Chen | Toward 3D Spatial Reasoning for Human-like Text-based Visual Question
Answering | Accepted by TIP2023, The Arxiv version of "Weakly-Supervised 3D
Spatial Reasoning for Text-based Visual Question Answering" | null | 10.1109/TIP.2023.3276570 | null | cs.CV cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-based Visual Question Answering~(TextVQA) aims to produce correct
answers for given questions about the images with multiple scene texts. In most
cases, the texts naturally attach to the surface of the objects. Therefore,
spatial reasoning between texts and objects is crucial in TextVQA. However,
existing approaches are constrained within 2D spatial information learned from
the input images and rely on transformer-based architectures to reason
implicitly during the fusion process. Under this setting, these 2D spatial
reasoning approaches cannot distinguish the fine-grain spatial relations
between visual objects and scene texts on the same image plane, thereby
impairing the interpretability and performance of TextVQA models. In this
paper, we introduce 3D geometric information into a human-like spatial
reasoning process to capture the contextual knowledge of key objects
step-by-step. %we formulate a human-like spatial reasoning process by
introducing 3D geometric information for capturing key objects' contextual
knowledge. To enhance the model's understanding of 3D spatial relationships,
Specifically, (i)~we propose a relation prediction module for accurately
locating the region of interest of critical objects; (ii)~we design a
depth-aware attention calibration module for calibrating the OCR tokens'
attention according to critical objects. Extensive experiments show that our
method achieves state-of-the-art performance on TextVQA and ST-VQA datasets.
More encouragingly, our model surpasses others by clear margins of 5.7\% and
12.1\% on questions that involve spatial reasoning in TextVQA and ST-VQA valid
split. Besides, we also verify the generalizability of our model on the
text-based image captioning task.
| [
{
"version": "v1",
"created": "Wed, 21 Sep 2022 12:49:14 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 02:38:25 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Li",
"Hao",
""
],
[
"Huang",
"Jinfa",
""
],
[
"Jin",
"Peng",
""
],
[
"Song",
"Guoli",
""
],
[
"Wu",
"Qi",
""
],
[
"Chen",
"Jie",
""
]
] | TITLE: Toward 3D Spatial Reasoning for Human-like Text-based Visual Question
Answering
ABSTRACT: Text-based Visual Question Answering~(TextVQA) aims to produce correct
answers for given questions about the images with multiple scene texts. In most
cases, the texts naturally attach to the surface of the objects. Therefore,
spatial reasoning between texts and objects is crucial in TextVQA. However,
existing approaches are constrained within 2D spatial information learned from
the input images and rely on transformer-based architectures to reason
implicitly during the fusion process. Under this setting, these 2D spatial
reasoning approaches cannot distinguish the fine-grain spatial relations
between visual objects and scene texts on the same image plane, thereby
impairing the interpretability and performance of TextVQA models. In this
paper, we introduce 3D geometric information into a human-like spatial
reasoning process to capture the contextual knowledge of key objects
step-by-step. %we formulate a human-like spatial reasoning process by
introducing 3D geometric information for capturing key objects' contextual
knowledge. To enhance the model's understanding of 3D spatial relationships,
Specifically, (i)~we propose a relation prediction module for accurately
locating the region of interest of critical objects; (ii)~we design a
depth-aware attention calibration module for calibrating the OCR tokens'
attention according to critical objects. Extensive experiments show that our
method achieves state-of-the-art performance on TextVQA and ST-VQA datasets.
More encouragingly, our model surpasses others by clear margins of 5.7\% and
12.1\% on questions that involve spatial reasoning in TextVQA and ST-VQA valid
split. Besides, we also verify the generalizability of our model on the
text-based image captioning task.
|
2210.01298 | Hanzhe Teng | Hanzhe Teng, Dimitrios Chatziparaschis, Xinyue Kan, Amit K.
Roy-Chowdhury, Konstantinos Karydis | Centroid Distance Keypoint Detector for Colored Point Clouds | Accepted to IEEE/CVF Winter Conference on Applications of Computer
Vision (WACV) 2023; copyright will be transferred to IEEE upon publication | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Keypoint detection serves as the basis for many computer vision and robotics
applications. Despite the fact that colored point clouds can be readily
obtained, most existing keypoint detectors extract only geometry-salient
keypoints, which can impede the overall performance of systems that intend to
(or have the potential to) leverage color information. To promote advances in
such systems, we propose an efficient multi-modal keypoint detector that can
extract both geometry-salient and color-salient keypoints in colored point
clouds. The proposed CEntroid Distance (CED) keypoint detector comprises an
intuitive and effective saliency measure, the centroid distance, that can be
used in both 3D space and color space, and a multi-modal non-maximum
suppression algorithm that can select keypoints with high saliency in two or
more modalities. The proposed saliency measure leverages directly the
distribution of points in a local neighborhood and does not require normal
estimation or eigenvalue decomposition. We evaluate the proposed method in
terms of repeatability and computational efficiency (i.e. running time) against
state-of-the-art keypoint detectors on both synthetic and real-world datasets.
Results demonstrate that our proposed CED keypoint detector requires minimal
computational time while attaining high repeatability. To showcase one of the
potential applications of the proposed method, we further investigate the task
of colored point cloud registration. Results suggest that our proposed CED
detector outperforms state-of-the-art handcrafted and learning-based keypoint
detectors in the evaluated scenes. The C++ implementation of the proposed
method is made publicly available at
https://github.com/UCR-Robotics/CED_Detector.
| [
{
"version": "v1",
"created": "Tue, 4 Oct 2022 00:55:51 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 04:43:24 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Teng",
"Hanzhe",
""
],
[
"Chatziparaschis",
"Dimitrios",
""
],
[
"Kan",
"Xinyue",
""
],
[
"Roy-Chowdhury",
"Amit K.",
""
],
[
"Karydis",
"Konstantinos",
""
]
] | TITLE: Centroid Distance Keypoint Detector for Colored Point Clouds
ABSTRACT: Keypoint detection serves as the basis for many computer vision and robotics
applications. Despite the fact that colored point clouds can be readily
obtained, most existing keypoint detectors extract only geometry-salient
keypoints, which can impede the overall performance of systems that intend to
(or have the potential to) leverage color information. To promote advances in
such systems, we propose an efficient multi-modal keypoint detector that can
extract both geometry-salient and color-salient keypoints in colored point
clouds. The proposed CEntroid Distance (CED) keypoint detector comprises an
intuitive and effective saliency measure, the centroid distance, that can be
used in both 3D space and color space, and a multi-modal non-maximum
suppression algorithm that can select keypoints with high saliency in two or
more modalities. The proposed saliency measure leverages directly the
distribution of points in a local neighborhood and does not require normal
estimation or eigenvalue decomposition. We evaluate the proposed method in
terms of repeatability and computational efficiency (i.e. running time) against
state-of-the-art keypoint detectors on both synthetic and real-world datasets.
Results demonstrate that our proposed CED keypoint detector requires minimal
computational time while attaining high repeatability. To showcase one of the
potential applications of the proposed method, we further investigate the task
of colored point cloud registration. Results suggest that our proposed CED
detector outperforms state-of-the-art handcrafted and learning-based keypoint
detectors in the evaluated scenes. The C++ implementation of the proposed
method is made publicly available at
https://github.com/UCR-Robotics/CED_Detector.
|
2210.02834 | Matteo Sodano | Matteo Sodano, Federico Magistri, Tiziano Guadagnino, Jens Behley,
Cyrill Stachniss | Robust Double-Encoder Network for RGB-D Panoptic Segmentation | null | ICRA 2023 | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Perception is crucial for robots that act in real-world environments, as
autonomous systems need to see and understand the world around them to act
properly. Panoptic segmentation provides an interpretation of the scene by
computing a pixelwise semantic label together with instance IDs. In this paper,
we address panoptic segmentation using RGB-D data of indoor scenes. We propose
a novel encoder-decoder neural network that processes RGB and depth separately
through two encoders. The features of the individual encoders are progressively
merged at different resolutions, such that the RGB features are enhanced using
complementary depth information. We propose a novel merging approach called
ResidualExcite, which reweighs each entry of the feature map according to its
importance. With our double-encoder architecture, we are robust to missing
cues. In particular, the same model can train and infer on RGB-D, RGB-only, and
depth-only input data, without the need to train specialized models. We
evaluate our method on publicly available datasets and show that our approach
achieves superior results compared to other common approaches for panoptic
segmentation.
| [
{
"version": "v1",
"created": "Thu, 6 Oct 2022 11:46:37 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 14:29:19 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Sodano",
"Matteo",
""
],
[
"Magistri",
"Federico",
""
],
[
"Guadagnino",
"Tiziano",
""
],
[
"Behley",
"Jens",
""
],
[
"Stachniss",
"Cyrill",
""
]
] | TITLE: Robust Double-Encoder Network for RGB-D Panoptic Segmentation
ABSTRACT: Perception is crucial for robots that act in real-world environments, as
autonomous systems need to see and understand the world around them to act
properly. Panoptic segmentation provides an interpretation of the scene by
computing a pixelwise semantic label together with instance IDs. In this paper,
we address panoptic segmentation using RGB-D data of indoor scenes. We propose
a novel encoder-decoder neural network that processes RGB and depth separately
through two encoders. The features of the individual encoders are progressively
merged at different resolutions, such that the RGB features are enhanced using
complementary depth information. We propose a novel merging approach called
ResidualExcite, which reweighs each entry of the feature map according to its
importance. With our double-encoder architecture, we are robust to missing
cues. In particular, the same model can train and infer on RGB-D, RGB-only, and
depth-only input data, without the need to train specialized models. We
evaluate our method on publicly available datasets and show that our approach
achieves superior results compared to other common approaches for panoptic
segmentation.
|
2210.04296 | Chieh-Hsin Lai | Chieh-Hsin Lai, Yuhta Takida, Naoki Murata, Toshimitsu Uesaka, Yuki
Mitsufuji, Stefano Ermon | FP-Diffusion: Improving Score-based Diffusion Models by Enforcing the
Underlying Score Fokker-Planck Equation | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Score-based generative models (SGMs) learn a family of noise-conditional
score functions corresponding to the data density perturbed with increasingly
large amounts of noise. These perturbed data densities are linked together by
the Fokker-Planck equation (FPE), a partial differential equation (PDE)
governing the spatial-temporal evolution of a density undergoing a diffusion
process. In this work, we derive a corresponding equation called the score FPE
that characterizes the noise-conditional scores of the perturbed data densities
(i.e., their gradients). Surprisingly, despite the impressive empirical
performance, we observe that scores learned through denoising score matching
(DSM) fail to fulfill the underlying score FPE, which is an inherent
self-consistency property of the ground truth score. We prove that satisfying
the score FPE is desirable as it improves the likelihood and the degree of
conservativity. Hence, we propose to regularize the DSM objective to enforce
satisfaction of the score FPE, and we show the effectiveness of this approach
across various datasets.
| [
{
"version": "v1",
"created": "Sun, 9 Oct 2022 16:27:25 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Nov 2022 00:37:09 GMT"
},
{
"version": "v3",
"created": "Tue, 31 Jan 2023 02:30:34 GMT"
},
{
"version": "v4",
"created": "Wed, 14 Jun 2023 05:26:28 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Lai",
"Chieh-Hsin",
""
],
[
"Takida",
"Yuhta",
""
],
[
"Murata",
"Naoki",
""
],
[
"Uesaka",
"Toshimitsu",
""
],
[
"Mitsufuji",
"Yuki",
""
],
[
"Ermon",
"Stefano",
""
]
] | TITLE: FP-Diffusion: Improving Score-based Diffusion Models by Enforcing the
Underlying Score Fokker-Planck Equation
ABSTRACT: Score-based generative models (SGMs) learn a family of noise-conditional
score functions corresponding to the data density perturbed with increasingly
large amounts of noise. These perturbed data densities are linked together by
the Fokker-Planck equation (FPE), a partial differential equation (PDE)
governing the spatial-temporal evolution of a density undergoing a diffusion
process. In this work, we derive a corresponding equation called the score FPE
that characterizes the noise-conditional scores of the perturbed data densities
(i.e., their gradients). Surprisingly, despite the impressive empirical
performance, we observe that scores learned through denoising score matching
(DSM) fail to fulfill the underlying score FPE, which is an inherent
self-consistency property of the ground truth score. We prove that satisfying
the score FPE is desirable as it improves the likelihood and the degree of
conservativity. Hence, we propose to regularize the DSM objective to enforce
satisfaction of the score FPE, and we show the effectiveness of this approach
across various datasets.
|
2210.08188 | Haiyun He | Haiyun He, Gholamali Aminian, Yuheng Bu, Miguel Rodrigues, Vincent Y.
F. Tan | How Does Pseudo-Labeling Affect the Generalization Error of the
Semi-Supervised Gibbs Algorithm? | 30 pages, 4 figures | null | null | null | cs.IT cs.LG math.IT | http://creativecommons.org/publicdomain/zero/1.0/ | We provide an exact characterization of the expected generalization error
(gen-error) for semi-supervised learning (SSL) with pseudo-labeling via the
Gibbs algorithm. The gen-error is expressed in terms of the symmetrized KL
information between the output hypothesis, the pseudo-labeled dataset, and the
labeled dataset. Distribution-free upper and lower bounds on the gen-error can
also be obtained. Our findings offer new insights that the generalization
performance of SSL with pseudo-labeling is affected not only by the information
between the output hypothesis and input training data but also by the
information {\em shared} between the {\em labeled} and {\em pseudo-labeled}
data samples. This serves as a guideline to choose an appropriate
pseudo-labeling method from a given family of methods. To deepen our
understanding, we further explore two examples -- mean estimation and logistic
regression. In particular, we analyze how the ratio of the number of unlabeled
to labeled data $\lambda$ affects the gen-error under both scenarios. As
$\lambda$ increases, the gen-error for mean estimation decreases and then
saturates at a value larger than when all the samples are labeled, and the gap
can be quantified {\em exactly} with our analysis, and is dependent on the
\emph{cross-covariance} between the labeled and pseudo-labeled data samples.
For logistic regression, the gen-error and the variance component of the excess
risk also decrease as $\lambda$ increases.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2022 04:11:56 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 17:22:45 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"He",
"Haiyun",
""
],
[
"Aminian",
"Gholamali",
""
],
[
"Bu",
"Yuheng",
""
],
[
"Rodrigues",
"Miguel",
""
],
[
"Tan",
"Vincent Y. F.",
""
]
] | TITLE: How Does Pseudo-Labeling Affect the Generalization Error of the
Semi-Supervised Gibbs Algorithm?
ABSTRACT: We provide an exact characterization of the expected generalization error
(gen-error) for semi-supervised learning (SSL) with pseudo-labeling via the
Gibbs algorithm. The gen-error is expressed in terms of the symmetrized KL
information between the output hypothesis, the pseudo-labeled dataset, and the
labeled dataset. Distribution-free upper and lower bounds on the gen-error can
also be obtained. Our findings offer new insights that the generalization
performance of SSL with pseudo-labeling is affected not only by the information
between the output hypothesis and input training data but also by the
information {\em shared} between the {\em labeled} and {\em pseudo-labeled}
data samples. This serves as a guideline to choose an appropriate
pseudo-labeling method from a given family of methods. To deepen our
understanding, we further explore two examples -- mean estimation and logistic
regression. In particular, we analyze how the ratio of the number of unlabeled
to labeled data $\lambda$ affects the gen-error under both scenarios. As
$\lambda$ increases, the gen-error for mean estimation decreases and then
saturates at a value larger than when all the samples are labeled, and the gap
can be quantified {\em exactly} with our analysis, and is dependent on the
\emph{cross-covariance} between the labeled and pseudo-labeled data samples.
For logistic regression, the gen-error and the variance component of the excess
risk also decrease as $\lambda$ increases.
|
2210.09759 | Nikolaos Dimitriadis | Nikolaos Dimitriadis, Pascal Frossard, Fran\c{c}ois Fleuret | Pareto Manifold Learning: Tackling multiple tasks via ensembles of
single-task models | Accepted ICML 2023 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | In Multi-Task Learning (MTL), tasks may compete and limit the performance
achieved on each other, rather than guiding the optimization to a solution,
superior to all its single-task trained counterparts. Since there is often not
a unique solution optimal for all tasks, practitioners have to balance
tradeoffs between tasks' performance, and resort to optimality in the Pareto
sense. Most MTL methodologies either completely neglect this aspect, and
instead of aiming at learning a Pareto Front, produce one solution predefined
by their optimization schemes, or produce diverse but discrete solutions.
Recent approaches parameterize the Pareto Front via neural networks, leading to
complex mappings from tradeoff to objective space. In this paper, we conjecture
that the Pareto Front admits a linear parameterization in parameter space,
which leads us to propose \textit{Pareto Manifold Learning}, an ensembling
method in weight space. Our approach produces a continuous Pareto Front in a
single training run, that allows to modulate the performance on each task
during inference. Experiments on multi-task learning benchmarks, ranging from
image classification to tabular datasets and scene understanding, show that
\textit{Pareto Manifold Learning} outperforms state-of-the-art single-point
algorithms, while learning a better Pareto parameterization than multi-point
baselines.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2022 11:20:54 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 11:45:45 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Dimitriadis",
"Nikolaos",
""
],
[
"Frossard",
"Pascal",
""
],
[
"Fleuret",
"François",
""
]
] | TITLE: Pareto Manifold Learning: Tackling multiple tasks via ensembles of
single-task models
ABSTRACT: In Multi-Task Learning (MTL), tasks may compete and limit the performance
achieved on each other, rather than guiding the optimization to a solution,
superior to all its single-task trained counterparts. Since there is often not
a unique solution optimal for all tasks, practitioners have to balance
tradeoffs between tasks' performance, and resort to optimality in the Pareto
sense. Most MTL methodologies either completely neglect this aspect, and
instead of aiming at learning a Pareto Front, produce one solution predefined
by their optimization schemes, or produce diverse but discrete solutions.
Recent approaches parameterize the Pareto Front via neural networks, leading to
complex mappings from tradeoff to objective space. In this paper, we conjecture
that the Pareto Front admits a linear parameterization in parameter space,
which leads us to propose \textit{Pareto Manifold Learning}, an ensembling
method in weight space. Our approach produces a continuous Pareto Front in a
single training run, that allows to modulate the performance on each task
during inference. Experiments on multi-task learning benchmarks, ranging from
image classification to tabular datasets and scene understanding, show that
\textit{Pareto Manifold Learning} outperforms state-of-the-art single-point
algorithms, while learning a better Pareto parameterization than multi-point
baselines.
|
2210.10636 | Parikshit Bansal | Parikshit Bansal, Yashoteja Prabhu, Emre Kiciman, Amit Sharma | Using Interventions to Improve Out-of-Distribution Generalization of
Text-Matching Recommendation Systems | NeurIPS 2022 CML4Impact Workshop, NeurIPS 2022 DistShift Workshop | null | null | null | cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Given a user's input text, text-matching recommender systems output relevant
items by comparing the input text to available items' description, such as
product-to-product recommendation on e-commerce platforms. As users' interests
and item inventory are expected to change, it is important for a text-matching
system to generalize to data shifts, a task known as out-of-distribution (OOD)
generalization. However, we find that the popular approach of fine-tuning a
large, base language model on paired item relevance data (e.g., user clicks)
can be counter-productive for OOD generalization. For a product recommendation
task, fine-tuning obtains worse accuracy than the base model when recommending
items in a new category or for a future time period. To explain this
generalization failure, we consider an intervention-based importance metric,
which shows that a fine-tuned model captures spurious correlations and fails to
learn the causal features that determine the relevance between any two text
inputs. Moreover, standard methods for causal regularization do not apply in
this setting, because unlike in images, there exist no universally spurious
features in a text-matching task (the same token may be spurious or causal
depending on the text it is being matched to). For OOD generalization on text
inputs, therefore, we highlight a different goal: avoiding high importance
scores for certain features. We do so using an intervention-based regularizer
that constraints the causal effect of any token on the model's relevance score
to be similar to the base model. Results on Amazon product and 3 question
recommendation datasets show that our proposed regularizer improves
generalization for both in-distribution and OOD evaluation, especially in
difficult scenarios when the base model is not accurate.
| [
{
"version": "v1",
"created": "Fri, 7 Oct 2022 11:16:45 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 10:17:56 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Bansal",
"Parikshit",
""
],
[
"Prabhu",
"Yashoteja",
""
],
[
"Kiciman",
"Emre",
""
],
[
"Sharma",
"Amit",
""
]
] | TITLE: Using Interventions to Improve Out-of-Distribution Generalization of
Text-Matching Recommendation Systems
ABSTRACT: Given a user's input text, text-matching recommender systems output relevant
items by comparing the input text to available items' description, such as
product-to-product recommendation on e-commerce platforms. As users' interests
and item inventory are expected to change, it is important for a text-matching
system to generalize to data shifts, a task known as out-of-distribution (OOD)
generalization. However, we find that the popular approach of fine-tuning a
large, base language model on paired item relevance data (e.g., user clicks)
can be counter-productive for OOD generalization. For a product recommendation
task, fine-tuning obtains worse accuracy than the base model when recommending
items in a new category or for a future time period. To explain this
generalization failure, we consider an intervention-based importance metric,
which shows that a fine-tuned model captures spurious correlations and fails to
learn the causal features that determine the relevance between any two text
inputs. Moreover, standard methods for causal regularization do not apply in
this setting, because unlike in images, there exist no universally spurious
features in a text-matching task (the same token may be spurious or causal
depending on the text it is being matched to). For OOD generalization on text
inputs, therefore, we highlight a different goal: avoiding high importance
scores for certain features. We do so using an intervention-based regularizer
that constraints the causal effect of any token on the model's relevance score
to be similar to the base model. Results on Amazon product and 3 question
recommendation datasets show that our proposed regularizer improves
generalization for both in-distribution and OOD evaluation, especially in
difficult scenarios when the base model is not accurate.
|
2211.04054 | Juan Pablo Zuluaga-Gomez | Juan Zuluaga-Gomez and Karel Vesel\'y and Igor Sz\"oke and Alexander
Blatt and Petr Motlicek and Martin Kocour and Mickael Rigault and Khalid
Choukri and Amrutha Prasad and Seyyed Saeed Sarfjoo and Iuliia Nigmatulina
and Claudia Cevenini and Pavel Kol\v{c}\'arek and Allan Tart and Jan
\v{C}ernock\'y and Dietrich Klakow | ATCO2 corpus: A Large-Scale Dataset for Research on Automatic Speech
Recognition and Natural Language Understanding of Air Traffic Control
Communications | Manuscript under review; The code is available at:
https://github.com/idiap/atco2-corpus | null | null | null | cs.CL cs.AI cs.SD eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Personal assistants, automatic speech recognizers and dialogue understanding
systems are becoming more critical in our interconnected digital world. A clear
example is air traffic control (ATC) communications. ATC aims at guiding
aircraft and controlling the airspace in a safe and optimal manner. These
voice-based dialogues are carried between an air traffic controller (ATCO) and
pilots via very-high frequency radio channels. In order to incorporate these
novel technologies into ATC (low-resource domain), large-scale annotated
datasets are required to develop the data-driven AI systems. Two examples are
automatic speech recognition (ASR) and natural language understanding (NLU). In
this paper, we introduce the ATCO2 corpus, a dataset that aims at fostering
research on the challenging ATC field, which has lagged behind due to lack of
annotated data. The ATCO2 corpus covers 1) data collection and pre-processing,
2) pseudo-annotations of speech data, and 3) extraction of ATC-related named
entities. The ATCO2 corpus is split into three subsets. 1) ATCO2-test-set
corpus contains 4 hours of ATC speech with manual transcripts and a subset with
gold annotations for named-entity recognition (callsign, command, value). 2)
The ATCO2-PL-set corpus consists of 5281 hours of unlabeled ATC data enriched
with automatic transcripts from an in-domain speech recognizer, contextual
information, speaker turn information, signal-to-noise ratio estimate and
English language detection score per sample. Both available for purchase
through ELDA at http://catalog.elra.info/en-us/repository/browse/ELRA-S0484. 3)
The ATCO2-test-set-1h corpus is a one-hour subset from the original test set
corpus, that we are offering for free at https://www.atco2.org/data. We expect
the ATCO2 corpus will foster research on robust ASR and NLU not only in the
field of ATC communications but also in the general research community.
| [
{
"version": "v1",
"created": "Tue, 8 Nov 2022 07:26:45 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 13:53:05 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Zuluaga-Gomez",
"Juan",
""
],
[
"Veselý",
"Karel",
""
],
[
"Szöke",
"Igor",
""
],
[
"Blatt",
"Alexander",
""
],
[
"Motlicek",
"Petr",
""
],
[
"Kocour",
"Martin",
""
],
[
"Rigault",
"Mickael",
""
],
[
"Choukri",
"Khalid",
""
],
[
"Prasad",
"Amrutha",
""
],
[
"Sarfjoo",
"Seyyed Saeed",
""
],
[
"Nigmatulina",
"Iuliia",
""
],
[
"Cevenini",
"Claudia",
""
],
[
"Kolčárek",
"Pavel",
""
],
[
"Tart",
"Allan",
""
],
[
"Černocký",
"Jan",
""
],
[
"Klakow",
"Dietrich",
""
]
] | TITLE: ATCO2 corpus: A Large-Scale Dataset for Research on Automatic Speech
Recognition and Natural Language Understanding of Air Traffic Control
Communications
ABSTRACT: Personal assistants, automatic speech recognizers and dialogue understanding
systems are becoming more critical in our interconnected digital world. A clear
example is air traffic control (ATC) communications. ATC aims at guiding
aircraft and controlling the airspace in a safe and optimal manner. These
voice-based dialogues are carried between an air traffic controller (ATCO) and
pilots via very-high frequency radio channels. In order to incorporate these
novel technologies into ATC (low-resource domain), large-scale annotated
datasets are required to develop the data-driven AI systems. Two examples are
automatic speech recognition (ASR) and natural language understanding (NLU). In
this paper, we introduce the ATCO2 corpus, a dataset that aims at fostering
research on the challenging ATC field, which has lagged behind due to lack of
annotated data. The ATCO2 corpus covers 1) data collection and pre-processing,
2) pseudo-annotations of speech data, and 3) extraction of ATC-related named
entities. The ATCO2 corpus is split into three subsets. 1) ATCO2-test-set
corpus contains 4 hours of ATC speech with manual transcripts and a subset with
gold annotations for named-entity recognition (callsign, command, value). 2)
The ATCO2-PL-set corpus consists of 5281 hours of unlabeled ATC data enriched
with automatic transcripts from an in-domain speech recognizer, contextual
information, speaker turn information, signal-to-noise ratio estimate and
English language detection score per sample. Both available for purchase
through ELDA at http://catalog.elra.info/en-us/repository/browse/ELRA-S0484. 3)
The ATCO2-test-set-1h corpus is a one-hour subset from the original test set
corpus, that we are offering for free at https://www.atco2.org/data. We expect
the ATCO2 corpus will foster research on robust ASR and NLU not only in the
field of ATC communications but also in the general research community.
|
2211.09715 | Micka\"el Delcey | Micka\"el Delcey, Yoann Cheny, S\'ebastien Kiesgen de Richter | Physics-informed neural networks for gravity currents reconstruction
from limited data | null | null | 10.1063/5.0136886 | null | physics.flu-dyn cs.LG | http://creativecommons.org/licenses/by/4.0/ | The present work investigates the use of physics-informed neural networks
(PINNs) for the 3D reconstruction of unsteady gravity currents from limited
data. In the PINN context, the flow fields are reconstructed by training a
neural network whose objective function penalizes the mismatch between the
network predictions and the observed data and embeds the underlying equations
using automatic differentiation. This study relies on a high-fidelity numerical
experiment of the canonical lock-exchange configuration. This allows us to
benchmark quantitatively the PINNs reconstruction capabilities on several
training databases that mimic state-of-the-art experimental measurement
techniques for density and velocity. Notably, spatially averaged density
measurements by light attenuation technique (LAT) are employed for the training
procedure. An optimal experimental setup for flow reconstruction by PINNs is
proposed according to two criteria : the implementation complexity and the
accuracy of the inferred fields.
| [
{
"version": "v1",
"created": "Thu, 3 Nov 2022 11:27:29 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 07:33:22 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Delcey",
"Mickaël",
""
],
[
"Cheny",
"Yoann",
""
],
[
"de Richter",
"Sébastien Kiesgen",
""
]
] | TITLE: Physics-informed neural networks for gravity currents reconstruction
from limited data
ABSTRACT: The present work investigates the use of physics-informed neural networks
(PINNs) for the 3D reconstruction of unsteady gravity currents from limited
data. In the PINN context, the flow fields are reconstructed by training a
neural network whose objective function penalizes the mismatch between the
network predictions and the observed data and embeds the underlying equations
using automatic differentiation. This study relies on a high-fidelity numerical
experiment of the canonical lock-exchange configuration. This allows us to
benchmark quantitatively the PINNs reconstruction capabilities on several
training databases that mimic state-of-the-art experimental measurement
techniques for density and velocity. Notably, spatially averaged density
measurements by light attenuation technique (LAT) are employed for the training
procedure. An optimal experimental setup for flow reconstruction by PINNs is
proposed according to two criteria : the implementation complexity and the
accuracy of the inferred fields.
|
2211.09878 | Jo\"el Tang | Jo\"el Tang, Marina Fomicheva, Lucia Specia | Reducing Hallucinations in Neural Machine Translation with Feature
Attribution | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Neural conditional language generation models achieve the state-of-the-art in
Neural Machine Translation (NMT) but are highly dependent on the quality of
parallel training dataset. When trained on low-quality datasets, these models
are prone to various error types, including hallucinations, i.e. outputs that
are fluent, but unrelated to the source sentences. These errors are
particularly dangerous, because on the surface the translation can be perceived
as a correct output, especially if the reader does not understand the source
language. We present a case study focusing on model understanding and
regularisation to reduce hallucinations in NMT. We first use feature
attribution methods to study the behaviour of an NMT model that produces
hallucinations. We then leverage these methods to propose a novel loss function
that substantially helps reduce hallucinations and does not require retraining
the model from scratch.
| [
{
"version": "v1",
"created": "Thu, 17 Nov 2022 20:33:56 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 19:36:10 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Tang",
"Joël",
""
],
[
"Fomicheva",
"Marina",
""
],
[
"Specia",
"Lucia",
""
]
] | TITLE: Reducing Hallucinations in Neural Machine Translation with Feature
Attribution
ABSTRACT: Neural conditional language generation models achieve the state-of-the-art in
Neural Machine Translation (NMT) but are highly dependent on the quality of
parallel training dataset. When trained on low-quality datasets, these models
are prone to various error types, including hallucinations, i.e. outputs that
are fluent, but unrelated to the source sentences. These errors are
particularly dangerous, because on the surface the translation can be perceived
as a correct output, especially if the reader does not understand the source
language. We present a case study focusing on model understanding and
regularisation to reduce hallucinations in NMT. We first use feature
attribution methods to study the behaviour of an NMT model that produces
hallucinations. We then leverage these methods to propose a novel loss function
that substantially helps reduce hallucinations and does not require retraining
the model from scratch.
|
2211.14051 | Jan Egger | Jianning Li, Andr\'e Ferreira, Behrus Puladi, Victor Alves, Michael
Kamp, Moon-Sung Kim, Felix Nensa, Jens Kleesiek, Seyed-Ahmad Ahmadi, Jan
Egger | Open-Source Skull Reconstruction with MONAI | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a deep learning-based approach for skull reconstruction for MONAI,
which has been pre-trained on the MUG500+ skull dataset. The implementation
follows the MONAI contribution guidelines, hence, it can be easily tried out
and used, and extended by MONAI users. The primary goal of this paper lies in
the investigation of open-sourcing codes and pre-trained deep learning models
under the MONAI framework. Nowadays, open-sourcing software, especially
(pre-trained) deep learning models, has become increasingly important. Over the
years, medical image analysis experienced a tremendous transformation. Over a
decade ago, algorithms had to be implemented and optimized with low-level
programming languages, like C or C++, to run in a reasonable time on a desktop
PC, which was not as powerful as today's computers. Nowadays, users have
high-level scripting languages like Python, and frameworks like PyTorch and
TensorFlow, along with a sea of public code repositories at hand. As a result,
implementations that had thousands of lines of C or C++ code in the past, can
now be scripted with a few lines and in addition executed in a fraction of the
time. To put this even on a higher level, the Medical Open Network for
Artificial Intelligence (MONAI) framework tailors medical imaging research to
an even more convenient process, which can boost and push the whole field. The
MONAI framework is a freely available, community-supported, open-source and
PyTorch-based framework, that also enables to provide research contributions
with pre-trained models to others. Codes and pre-trained weights for skull
reconstruction are publicly available at:
https://github.com/Project-MONAI/research-contributions/tree/master/SkullRec
| [
{
"version": "v1",
"created": "Fri, 25 Nov 2022 12:12:23 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 09:37:12 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Li",
"Jianning",
""
],
[
"Ferreira",
"André",
""
],
[
"Puladi",
"Behrus",
""
],
[
"Alves",
"Victor",
""
],
[
"Kamp",
"Michael",
""
],
[
"Kim",
"Moon-Sung",
""
],
[
"Nensa",
"Felix",
""
],
[
"Kleesiek",
"Jens",
""
],
[
"Ahmadi",
"Seyed-Ahmad",
""
],
[
"Egger",
"Jan",
""
]
] | TITLE: Open-Source Skull Reconstruction with MONAI
ABSTRACT: We present a deep learning-based approach for skull reconstruction for MONAI,
which has been pre-trained on the MUG500+ skull dataset. The implementation
follows the MONAI contribution guidelines, hence, it can be easily tried out
and used, and extended by MONAI users. The primary goal of this paper lies in
the investigation of open-sourcing codes and pre-trained deep learning models
under the MONAI framework. Nowadays, open-sourcing software, especially
(pre-trained) deep learning models, has become increasingly important. Over the
years, medical image analysis experienced a tremendous transformation. Over a
decade ago, algorithms had to be implemented and optimized with low-level
programming languages, like C or C++, to run in a reasonable time on a desktop
PC, which was not as powerful as today's computers. Nowadays, users have
high-level scripting languages like Python, and frameworks like PyTorch and
TensorFlow, along with a sea of public code repositories at hand. As a result,
implementations that had thousands of lines of C or C++ code in the past, can
now be scripted with a few lines and in addition executed in a fraction of the
time. To put this even on a higher level, the Medical Open Network for
Artificial Intelligence (MONAI) framework tailors medical imaging research to
an even more convenient process, which can boost and push the whole field. The
MONAI framework is a freely available, community-supported, open-source and
PyTorch-based framework, that also enables to provide research contributions
with pre-trained models to others. Codes and pre-trained weights for skull
reconstruction are publicly available at:
https://github.com/Project-MONAI/research-contributions/tree/master/SkullRec
|
2211.14208 | Jeongwhan Choi | Jeongwhan Choi, Seoyoung Hong, Noseong Park, Sung-Bae Cho | GREAD: Graph Neural Reaction-Diffusion Networks | Accepted by ICML 2023 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Graph neural networks (GNNs) are one of the most popular research topics for
deep learning. GNN methods typically have been designed on top of the graph
signal processing theory. In particular, diffusion equations have been widely
used for designing the core processing layer of GNNs, and therefore they are
inevitably vulnerable to the notorious oversmoothing problem. Recently, a
couple of papers paid attention to reaction equations in conjunctions with
diffusion equations. However, they all consider limited forms of reaction
equations. To this end, we present a reaction-diffusion equation-based GNN
method that considers all popular types of reaction equations in addition to
one special reaction equation designed by us. To our knowledge, our paper is
one of the most comprehensive studies on reaction-diffusion equation-based
GNNs. In our experiments with 9 datasets and 28 baselines, our method, called
GREAD, outperforms them in a majority of cases. Further synthetic data
experiments show that it mitigates the oversmoothing problem and works well for
various homophily rates.
| [
{
"version": "v1",
"created": "Fri, 25 Nov 2022 16:16:25 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 15:34:42 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Jun 2023 22:53:38 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Choi",
"Jeongwhan",
""
],
[
"Hong",
"Seoyoung",
""
],
[
"Park",
"Noseong",
""
],
[
"Cho",
"Sung-Bae",
""
]
] | TITLE: GREAD: Graph Neural Reaction-Diffusion Networks
ABSTRACT: Graph neural networks (GNNs) are one of the most popular research topics for
deep learning. GNN methods typically have been designed on top of the graph
signal processing theory. In particular, diffusion equations have been widely
used for designing the core processing layer of GNNs, and therefore they are
inevitably vulnerable to the notorious oversmoothing problem. Recently, a
couple of papers paid attention to reaction equations in conjunctions with
diffusion equations. However, they all consider limited forms of reaction
equations. To this end, we present a reaction-diffusion equation-based GNN
method that considers all popular types of reaction equations in addition to
one special reaction equation designed by us. To our knowledge, our paper is
one of the most comprehensive studies on reaction-diffusion equation-based
GNNs. In our experiments with 9 datasets and 28 baselines, our method, called
GREAD, outperforms them in a majority of cases. Further synthetic data
experiments show that it mitigates the oversmoothing problem and works well for
various homophily rates.
|
2211.15631 | Alexey Melnikov | Mohammad Kordzanganeh, Markus Buchberger, Basil Kyriacou, Maxim
Povolotskii, Wilhelm Fischer, Andrii Kurkin, Wilfrid Somogyi, Asel
Sagingalieva, Markus Pflitsch, Alexey Melnikov | Benchmarking simulated and physical quantum processing units using
quantum and hybrid algorithms | 21 pages, 6 figures, 12 tables | Adv. Quantum Technol. 6, 2300043 (2023) | 10.1002/qute.202300043 | null | quant-ph cs.LG cs.PF | http://creativecommons.org/licenses/by/4.0/ | Powerful hardware services and software libraries are vital tools for quickly
and affordably designing, testing, and executing quantum algorithms. A robust
large-scale study of how the performance of these platforms scales with the
number of qubits is key to providing quantum solutions to challenging industry
problems. This work benchmarks the runtime and accuracy for a representative
sample of specialized high-performance simulated and physical quantum
processing units. Results show the QMware simulator can reduce the runtime for
executing a quantum circuit by up to 78% compared to the next fastest option
for algorithms with fewer than 27 qubits. The AWS SV1 simulator offers a
runtime advantage for larger circuits, up to the maximum 34 qubits available
with SV1. Beyond this limit, QMware can execute circuits as large as 40 qubits.
Physical quantum devices, such as Rigetti's Aspen-M2, can provide an
exponential runtime advantage for circuits with more than 30 qubits. However,
the high financial cost of physical quantum processing units presents a serious
barrier to practical use. Moreover, only IonQ's Harmony quantum device achieves
high fidelity with more than four qubits. This study paves the way to
understanding the optimal combination of available software and hardware for
executing practical quantum algorithms.
| [
{
"version": "v1",
"created": "Mon, 28 Nov 2022 18:35:18 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 17:36:40 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Kordzanganeh",
"Mohammad",
""
],
[
"Buchberger",
"Markus",
""
],
[
"Kyriacou",
"Basil",
""
],
[
"Povolotskii",
"Maxim",
""
],
[
"Fischer",
"Wilhelm",
""
],
[
"Kurkin",
"Andrii",
""
],
[
"Somogyi",
"Wilfrid",
""
],
[
"Sagingalieva",
"Asel",
""
],
[
"Pflitsch",
"Markus",
""
],
[
"Melnikov",
"Alexey",
""
]
] | TITLE: Benchmarking simulated and physical quantum processing units using
quantum and hybrid algorithms
ABSTRACT: Powerful hardware services and software libraries are vital tools for quickly
and affordably designing, testing, and executing quantum algorithms. A robust
large-scale study of how the performance of these platforms scales with the
number of qubits is key to providing quantum solutions to challenging industry
problems. This work benchmarks the runtime and accuracy for a representative
sample of specialized high-performance simulated and physical quantum
processing units. Results show the QMware simulator can reduce the runtime for
executing a quantum circuit by up to 78% compared to the next fastest option
for algorithms with fewer than 27 qubits. The AWS SV1 simulator offers a
runtime advantage for larger circuits, up to the maximum 34 qubits available
with SV1. Beyond this limit, QMware can execute circuits as large as 40 qubits.
Physical quantum devices, such as Rigetti's Aspen-M2, can provide an
exponential runtime advantage for circuits with more than 30 qubits. However,
the high financial cost of physical quantum processing units presents a serious
barrier to practical use. Moreover, only IonQ's Harmony quantum device achieves
high fidelity with more than four qubits. This study paves the way to
understanding the optimal combination of available software and hardware for
executing practical quantum algorithms.
|
2212.00239 | Hanzhi Yin | Ruibin Yuan, Hanzhi Yin, Yi Wang, Yifan He, Yushi Ye, Lei Zhang,
Zhizheng Wu | Inconsistency Ranking-based Noisy Label Detection for High-quality Data | 5 pages | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The success of deep learning requires high-quality annotated and massive
data. However, the size and the quality of a dataset are usually a trade-off in
practice, as data collection and cleaning are expensive and time-consuming. In
real-world applications, especially those using crowdsourcing datasets, it is
important to exclude noisy labels. To address this, this paper proposes an
automatic noisy label detection (NLD) technique with inconsistency ranking for
high-quality data. We apply this technique to the automatic speaker
verification (ASV) task as a proof of concept. We investigate both inter-class
and intra-class inconsistency ranking and compare several metric learning loss
functions under different noise settings. Experimental results confirm that the
proposed solution could increase both the efficient and effective cleaning of
large-scale speaker recognition datasets.
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2022 03:09:33 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 14:08:55 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Yuan",
"Ruibin",
""
],
[
"Yin",
"Hanzhi",
""
],
[
"Wang",
"Yi",
""
],
[
"He",
"Yifan",
""
],
[
"Ye",
"Yushi",
""
],
[
"Zhang",
"Lei",
""
],
[
"Wu",
"Zhizheng",
""
]
] | TITLE: Inconsistency Ranking-based Noisy Label Detection for High-quality Data
ABSTRACT: The success of deep learning requires high-quality annotated and massive
data. However, the size and the quality of a dataset are usually a trade-off in
practice, as data collection and cleaning are expensive and time-consuming. In
real-world applications, especially those using crowdsourcing datasets, it is
important to exclude noisy labels. To address this, this paper proposes an
automatic noisy label detection (NLD) technique with inconsistency ranking for
high-quality data. We apply this technique to the automatic speaker
verification (ASV) task as a proof of concept. We investigate both inter-class
and intra-class inconsistency ranking and compare several metric learning loss
functions under different noise settings. Experimental results confirm that the
proposed solution could increase both the efficient and effective cleaning of
large-scale speaker recognition datasets.
|
2212.05679 | Peiran Gong | Wu Yue, Peiran Gong, Maoguo Gong, Hangqi Ding, Zedong Tang, Yibo Liu,
Wenping Ma, Qiguang Miao | Evolutionary Multitasking with Solution Space Cutting for Point Cloud
Registration | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Point cloud registration (PCR) is a popular research topic in computer
vision. Recently, the registration method in an evolutionary way has received
continuous attention because of its robustness to the initial pose and
flexibility in objective function design. However, most evolving registration
methods cannot tackle the local optimum well and they have rarely investigated
the success ratio, which implies the probability of not falling into local
optima and is closely related to the practicality of the algorithm.
Evolutionary multi-task optimization (EMTO) is a widely used paradigm, which
can boost exploration capability through knowledge transfer among related
tasks. Inspired by this concept, this study proposes a novel evolving
registration algorithm via EMTO, where the multi-task configuration is based on
the idea of solution space cutting. Concretely, one task searching in cut space
assists another task with complex function landscape in escaping from local
optima and enhancing successful registration ratio. To reduce unnecessary
computational cost, a sparse-to-dense strategy is proposed. In addition, a
novel fitness function robust to various overlap rates as well as a
problem-specific metric of computational cost is introduced. Compared with 8
evolving approaches, 4 traditional approaches and 3 deep learning approaches on
the object-scale and scene-scale registration datasets, experimental results
demonstrate that the proposed method has superior performances in terms of
precision and tackling local optima.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 03:32:05 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 14:05:58 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Yue",
"Wu",
""
],
[
"Gong",
"Peiran",
""
],
[
"Gong",
"Maoguo",
""
],
[
"Ding",
"Hangqi",
""
],
[
"Tang",
"Zedong",
""
],
[
"Liu",
"Yibo",
""
],
[
"Ma",
"Wenping",
""
],
[
"Miao",
"Qiguang",
""
]
] | TITLE: Evolutionary Multitasking with Solution Space Cutting for Point Cloud
Registration
ABSTRACT: Point cloud registration (PCR) is a popular research topic in computer
vision. Recently, the registration method in an evolutionary way has received
continuous attention because of its robustness to the initial pose and
flexibility in objective function design. However, most evolving registration
methods cannot tackle the local optimum well and they have rarely investigated
the success ratio, which implies the probability of not falling into local
optima and is closely related to the practicality of the algorithm.
Evolutionary multi-task optimization (EMTO) is a widely used paradigm, which
can boost exploration capability through knowledge transfer among related
tasks. Inspired by this concept, this study proposes a novel evolving
registration algorithm via EMTO, where the multi-task configuration is based on
the idea of solution space cutting. Concretely, one task searching in cut space
assists another task with complex function landscape in escaping from local
optima and enhancing successful registration ratio. To reduce unnecessary
computational cost, a sparse-to-dense strategy is proposed. In addition, a
novel fitness function robust to various overlap rates as well as a
problem-specific metric of computational cost is introduced. Compared with 8
evolving approaches, 4 traditional approaches and 3 deep learning approaches on
the object-scale and scene-scale registration datasets, experimental results
demonstrate that the proposed method has superior performances in terms of
precision and tackling local optima.
|
2212.05749 | Nicklas Hansen | Nicklas Hansen and Zhecheng Yuan and Yanjie Ze and Tongzhou Mu and
Aravind Rajeswaran and Hao Su and Huazhe Xu and Xiaolong Wang | On Pre-Training for Visuo-Motor Control: Revisiting a
Learning-from-Scratch Baseline | Code: https://github.com/gemcollector/learning-from-scratch | null | null | null | cs.LG cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | In this paper, we examine the effectiveness of pre-training for visuo-motor
control tasks. We revisit a simple Learning-from-Scratch (LfS) baseline that
incorporates data augmentation and a shallow ConvNet, and find that this
baseline is surprisingly competitive with recent approaches (PVR, MVP, R3M)
that leverage frozen visual representations trained on large-scale vision
datasets -- across a variety of algorithms, task domains, and metrics in
simulation and on a real robot. Our results demonstrate that these methods are
hindered by a significant domain gap between the pre-training datasets and
current benchmarks for visuo-motor control, which is alleviated by finetuning.
Based on our findings, we provide recommendations for future research in
pre-training for control and hope that our simple yet strong baseline will aid
in accurately benchmarking progress in this area.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 07:59:31 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 17:57:21 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Hansen",
"Nicklas",
""
],
[
"Yuan",
"Zhecheng",
""
],
[
"Ze",
"Yanjie",
""
],
[
"Mu",
"Tongzhou",
""
],
[
"Rajeswaran",
"Aravind",
""
],
[
"Su",
"Hao",
""
],
[
"Xu",
"Huazhe",
""
],
[
"Wang",
"Xiaolong",
""
]
] | TITLE: On Pre-Training for Visuo-Motor Control: Revisiting a
Learning-from-Scratch Baseline
ABSTRACT: In this paper, we examine the effectiveness of pre-training for visuo-motor
control tasks. We revisit a simple Learning-from-Scratch (LfS) baseline that
incorporates data augmentation and a shallow ConvNet, and find that this
baseline is surprisingly competitive with recent approaches (PVR, MVP, R3M)
that leverage frozen visual representations trained on large-scale vision
datasets -- across a variety of algorithms, task domains, and metrics in
simulation and on a real robot. Our results demonstrate that these methods are
hindered by a significant domain gap between the pre-training datasets and
current benchmarks for visuo-motor control, which is alleviated by finetuning.
Based on our findings, we provide recommendations for future research in
pre-training for control and hope that our simple yet strong baseline will aid
in accurately benchmarking progress in this area.
|
2212.06331 | Xinhao Liu | Chao Chen, Xinhao Liu, Yiming Li, Li Ding, Chen Feng | DeepMapping2: Self-Supervised Large-Scale LiDAR Map Optimization | accepted by CVPR 2023 | Proceedings of the IEEE/CVF Conference on Computer Vision and
Pattern Recognition, pages 9306--9316 (2023) | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | LiDAR mapping is important yet challenging in self-driving and mobile
robotics. To tackle such a global point cloud registration problem, DeepMapping
converts the complex map estimation into a self-supervised training of simple
deep networks. Despite its broad convergence range on small datasets,
DeepMapping still cannot produce satisfactory results on large-scale datasets
with thousands of frames. This is due to the lack of loop closures and exact
cross-frame point correspondences, and the slow convergence of its global
localization network. We propose DeepMapping2 by adding two novel techniques to
address these issues: (1) organization of training batch based on map topology
from loop closing, and (2) self-supervised local-to-global point consistency
loss leveraging pairwise registration. Our experiments and ablation studies on
public datasets (KITTI, NCLT, and Nebula) demonstrate the effectiveness of our
method.
| [
{
"version": "v1",
"created": "Tue, 13 Dec 2022 02:22:04 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 06:08:57 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Chen",
"Chao",
""
],
[
"Liu",
"Xinhao",
""
],
[
"Li",
"Yiming",
""
],
[
"Ding",
"Li",
""
],
[
"Feng",
"Chen",
""
]
] | TITLE: DeepMapping2: Self-Supervised Large-Scale LiDAR Map Optimization
ABSTRACT: LiDAR mapping is important yet challenging in self-driving and mobile
robotics. To tackle such a global point cloud registration problem, DeepMapping
converts the complex map estimation into a self-supervised training of simple
deep networks. Despite its broad convergence range on small datasets,
DeepMapping still cannot produce satisfactory results on large-scale datasets
with thousands of frames. This is due to the lack of loop closures and exact
cross-frame point correspondences, and the slow convergence of its global
localization network. We propose DeepMapping2 by adding two novel techniques to
address these issues: (1) organization of training batch based on map topology
from loop closing, and (2) self-supervised local-to-global point consistency
loss leveraging pairwise registration. Our experiments and ablation studies on
public datasets (KITTI, NCLT, and Nebula) demonstrate the effectiveness of our
method.
|
2212.10873 | Hyunsoo Cho | Hyunsoo Cho, Hyuhng Joon Kim, Junyeob Kim, Sang-Woo Lee, Sang-goo Lee,
Kang Min Yoo, Taeuk Kim | Prompt-Augmented Linear Probing: Scaling beyond the Limit of Few-shot
In-Context Learners | AAAI 2023 | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Through in-context learning (ICL), large-scale language models are effective
few-shot learners without additional model fine-tuning. However, the ICL
performance does not scale well with the number of available training samples
as it is limited by the inherent input length constraint of the underlying
language model. Meanwhile, many studies have revealed that language models are
also powerful feature extractors, allowing them to be utilized in a black-box
manner and enabling the linear probing paradigm, where lightweight
discriminators are trained on top of the pre-extracted input representations.
This paper proposes prompt-augmented linear probing (PALP), a hybrid of linear
probing and ICL, which leverages the best of both worlds. PALP inherits the
scalability of linear probing and the capability of enforcing language models
to derive more meaningful representations via tailoring input into a more
conceivable form. Throughout in-depth investigations on various datasets, we
verified that PALP significantly enhances the input representations closing the
gap between ICL in the data-hungry scenario and fine-tuning in the
data-abundant scenario with little training overhead, potentially making PALP a
strong alternative in a black-box scenario.
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2022 09:37:05 GMT"
},
{
"version": "v2",
"created": "Wed, 28 Dec 2022 01:08:35 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Jun 2023 03:23:44 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Cho",
"Hyunsoo",
""
],
[
"Kim",
"Hyuhng Joon",
""
],
[
"Kim",
"Junyeob",
""
],
[
"Lee",
"Sang-Woo",
""
],
[
"Lee",
"Sang-goo",
""
],
[
"Yoo",
"Kang Min",
""
],
[
"Kim",
"Taeuk",
""
]
] | TITLE: Prompt-Augmented Linear Probing: Scaling beyond the Limit of Few-shot
In-Context Learners
ABSTRACT: Through in-context learning (ICL), large-scale language models are effective
few-shot learners without additional model fine-tuning. However, the ICL
performance does not scale well with the number of available training samples
as it is limited by the inherent input length constraint of the underlying
language model. Meanwhile, many studies have revealed that language models are
also powerful feature extractors, allowing them to be utilized in a black-box
manner and enabling the linear probing paradigm, where lightweight
discriminators are trained on top of the pre-extracted input representations.
This paper proposes prompt-augmented linear probing (PALP), a hybrid of linear
probing and ICL, which leverages the best of both worlds. PALP inherits the
scalability of linear probing and the capability of enforcing language models
to derive more meaningful representations via tailoring input into a more
conceivable form. Throughout in-depth investigations on various datasets, we
verified that PALP significantly enhances the input representations closing the
gap between ICL in the data-hungry scenario and fine-tuning in the
data-abundant scenario with little training overhead, potentially making PALP a
strong alternative in a black-box scenario.
|
2301.06673 | Quoc-Huy Trinh | Trong-Hieu Nguyen Mau, Quoc-Huy Trinh, Nhat-Tan Bui, Minh-Triet Tran,
Hai-Dang Nguyen | Multi Kernel Positional Embedding ConvNeXt for Polyp Segmentation | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Medical image segmentation is the technique that helps doctor view and has a
precise diagnosis, particularly in Colorectal Cancer. Specifically, with the
increase in cases, the diagnosis and identification need to be faster and more
accurate for many patients; in endoscopic images, the segmentation task has
been vital to helping the doctor identify the position of the polyps or the
ache in the system correctly. As a result, many efforts have been made to apply
deep learning to automate polyp segmentation, mostly to ameliorate the U-shape
structure. However, the simple skip connection scheme in UNet leads to
deficient context information and the semantic gap between feature maps from
the encoder and decoder. To deal with this problem, we propose a novel
framework composed of ConvNeXt backbone and Multi Kernel Positional Embedding
block. Thanks to the suggested module, our method can attain better accuracy
and generalization in the polyps segmentation task. Extensive experiments show
that our model achieves the Dice coefficient of 0.8818 and the IOU score of
0.8163 on the Kvasir-SEG dataset. Furthermore, on various datasets, we make
competitive achievement results with other previous state-of-the-art methods.
| [
{
"version": "v1",
"created": "Tue, 17 Jan 2023 03:12:57 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 08:08:06 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Mau",
"Trong-Hieu Nguyen",
""
],
[
"Trinh",
"Quoc-Huy",
""
],
[
"Bui",
"Nhat-Tan",
""
],
[
"Tran",
"Minh-Triet",
""
],
[
"Nguyen",
"Hai-Dang",
""
]
] | TITLE: Multi Kernel Positional Embedding ConvNeXt for Polyp Segmentation
ABSTRACT: Medical image segmentation is the technique that helps doctor view and has a
precise diagnosis, particularly in Colorectal Cancer. Specifically, with the
increase in cases, the diagnosis and identification need to be faster and more
accurate for many patients; in endoscopic images, the segmentation task has
been vital to helping the doctor identify the position of the polyps or the
ache in the system correctly. As a result, many efforts have been made to apply
deep learning to automate polyp segmentation, mostly to ameliorate the U-shape
structure. However, the simple skip connection scheme in UNet leads to
deficient context information and the semantic gap between feature maps from
the encoder and decoder. To deal with this problem, we propose a novel
framework composed of ConvNeXt backbone and Multi Kernel Positional Embedding
block. Thanks to the suggested module, our method can attain better accuracy
and generalization in the polyps segmentation task. Extensive experiments show
that our model achieves the Dice coefficient of 0.8818 and the IOU score of
0.8163 on the Kvasir-SEG dataset. Furthermore, on various datasets, we make
competitive achievement results with other previous state-of-the-art methods.
|
2301.11660 | Hyunsoo Cho | Hyunsoo Cho, Choonghyun Park, Junyeop Kim, Hyuhng Joon Kim, Kang Min
Yoo, and Sang-goo Lee | Probing Out-of-Distribution Robustness of Language Models with
Parameter-Efficient Transfer Learning | *SEM 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | As the size of the pre-trained language model (PLM) continues to increase,
numerous parameter-efficient transfer learning methods have been proposed
recently to compensate for the tremendous cost of fine-tuning. Despite the
impressive results achieved by large pre-trained language models (PLMs) and
various parameter-efficient transfer learning (PETL) methods on sundry
benchmarks, it remains unclear if they can handle inputs that have been
distributionally shifted effectively. In this study, we systematically explore
how the ability to detect out-of-distribution (OOD) changes as the size of the
PLM grows or the transfer methods are altered. Specifically, we evaluated
various PETL techniques, including fine-tuning, Adapter, LoRA, and
prefix-tuning, on three different intention classification tasks, each
utilizing various language models with different scales.
| [
{
"version": "v1",
"created": "Fri, 27 Jan 2023 11:27:40 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Jan 2023 07:17:09 GMT"
},
{
"version": "v3",
"created": "Fri, 9 Jun 2023 05:06:58 GMT"
},
{
"version": "v4",
"created": "Wed, 14 Jun 2023 03:12:50 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Cho",
"Hyunsoo",
""
],
[
"Park",
"Choonghyun",
""
],
[
"Kim",
"Junyeop",
""
],
[
"Kim",
"Hyuhng Joon",
""
],
[
"Yoo",
"Kang Min",
""
],
[
"Lee",
"Sang-goo",
""
]
] | TITLE: Probing Out-of-Distribution Robustness of Language Models with
Parameter-Efficient Transfer Learning
ABSTRACT: As the size of the pre-trained language model (PLM) continues to increase,
numerous parameter-efficient transfer learning methods have been proposed
recently to compensate for the tremendous cost of fine-tuning. Despite the
impressive results achieved by large pre-trained language models (PLMs) and
various parameter-efficient transfer learning (PETL) methods on sundry
benchmarks, it remains unclear if they can handle inputs that have been
distributionally shifted effectively. In this study, we systematically explore
how the ability to detect out-of-distribution (OOD) changes as the size of the
PLM grows or the transfer methods are altered. Specifically, we evaluated
various PETL techniques, including fine-tuning, Adapter, LoRA, and
prefix-tuning, on three different intention classification tasks, each
utilizing various language models with different scales.
|
2302.00930 | Xingping Dong | Xingping Dong, Jianbing Shen, Fatih Porikli, Jiebo Luo, and Ling Shao | Adaptive Siamese Tracking with a Compact Latent Network | Accepted at TPAMI | IEEE Transactions on Pattern Analysis and Machine Intelligence,
2022 | 10.1109/TPAMI.2022.3230064 | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we provide an intuitive viewing to simplify the Siamese-based
trackers by converting the tracking task to a classification. Under this
viewing, we perform an in-depth analysis for them through visual simulations
and real tracking examples, and find that the failure cases in some challenging
situations can be regarded as the issue of missing decisive samples in offline
training. Since the samples in the initial (first) frame contain rich
sequence-specific information, we can regard them as the decisive samples to
represent the whole sequence. To quickly adapt the base model to new scenes, a
compact latent network is presented via fully using these decisive samples.
Specifically, we present a statistics-based compact latent feature for fast
adjustment by efficiently extracting the sequence-specific information.
Furthermore, a new diverse sample mining strategy is designed for training to
further improve the discrimination ability of the proposed compact latent
network. Finally, a conditional updating strategy is proposed to efficiently
update the basic models to handle scene variation during the tracking phase. To
evaluate the generalization ability and effectiveness and of our method, we
apply it to adjust three classical Siamese-based trackers, namely SiamRPN++,
SiamFC, and SiamBAN. Extensive experimental results on six recent datasets
demonstrate that all three adjusted trackers obtain the superior performance in
terms of the accuracy, while having high running speed.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2023 08:06:02 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 09:59:34 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Dong",
"Xingping",
""
],
[
"Shen",
"Jianbing",
""
],
[
"Porikli",
"Fatih",
""
],
[
"Luo",
"Jiebo",
""
],
[
"Shao",
"Ling",
""
]
] | TITLE: Adaptive Siamese Tracking with a Compact Latent Network
ABSTRACT: In this paper, we provide an intuitive viewing to simplify the Siamese-based
trackers by converting the tracking task to a classification. Under this
viewing, we perform an in-depth analysis for them through visual simulations
and real tracking examples, and find that the failure cases in some challenging
situations can be regarded as the issue of missing decisive samples in offline
training. Since the samples in the initial (first) frame contain rich
sequence-specific information, we can regard them as the decisive samples to
represent the whole sequence. To quickly adapt the base model to new scenes, a
compact latent network is presented via fully using these decisive samples.
Specifically, we present a statistics-based compact latent feature for fast
adjustment by efficiently extracting the sequence-specific information.
Furthermore, a new diverse sample mining strategy is designed for training to
further improve the discrimination ability of the proposed compact latent
network. Finally, a conditional updating strategy is proposed to efficiently
update the basic models to handle scene variation during the tracking phase. To
evaluate the generalization ability and effectiveness and of our method, we
apply it to adjust three classical Siamese-based trackers, namely SiamRPN++,
SiamFC, and SiamBAN. Extensive experimental results on six recent datasets
demonstrate that all three adjusted trackers obtain the superior performance in
terms of the accuracy, while having high running speed.
|
2302.01308 | Raja Marjieh | Raja Marjieh, Ilia Sucholutsky, Pol van Rijn, Nori Jacoby, Thomas L.
Griffiths | Large language models predict human sensory judgments across six
modalities | 9 pages, 3 figures | null | null | null | cs.CL cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Determining the extent to which the perceptual world can be recovered from
language is a longstanding problem in philosophy and cognitive science. We show
that state-of-the-art large language models can unlock new insights into this
problem by providing a lower bound on the amount of perceptual information that
can be extracted from language. Specifically, we elicit pairwise similarity
judgments from GPT models across six psychophysical datasets. We show that the
judgments are significantly correlated with human data across all domains,
recovering well-known representations like the color wheel and pitch spiral.
Surprisingly, we find that a model (GPT-4) co-trained on vision and language
does not necessarily lead to improvements specific to the visual modality. To
study the influence of specific languages on perception, we also apply the
models to a multilingual color-naming task. We find that GPT-4 replicates
cross-linguistic variation in English and Russian illuminating the interaction
of language and perception.
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2023 18:32:46 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 17:18:04 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Marjieh",
"Raja",
""
],
[
"Sucholutsky",
"Ilia",
""
],
[
"van Rijn",
"Pol",
""
],
[
"Jacoby",
"Nori",
""
],
[
"Griffiths",
"Thomas L.",
""
]
] | TITLE: Large language models predict human sensory judgments across six
modalities
ABSTRACT: Determining the extent to which the perceptual world can be recovered from
language is a longstanding problem in philosophy and cognitive science. We show
that state-of-the-art large language models can unlock new insights into this
problem by providing a lower bound on the amount of perceptual information that
can be extracted from language. Specifically, we elicit pairwise similarity
judgments from GPT models across six psychophysical datasets. We show that the
judgments are significantly correlated with human data across all domains,
recovering well-known representations like the color wheel and pitch spiral.
Surprisingly, we find that a model (GPT-4) co-trained on vision and language
does not necessarily lead to improvements specific to the visual modality. To
study the influence of specific languages on perception, we also apply the
models to a multilingual color-naming task. We find that GPT-4 replicates
cross-linguistic variation in English and Russian illuminating the interaction
of language and perception.
|
2302.03655 | Saro Passaro | Saro Passaro, C. Lawrence Zitnick | Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNs | 19 pages, 10 figures | null | null | null | cs.LG physics.chem-ph physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph neural networks that model 3D data, such as point clouds or atoms, are
typically desired to be $SO(3)$ equivariant, i.e., equivariant to 3D rotations.
Unfortunately equivariant convolutions, which are a fundamental operation for
equivariant networks, increase significantly in computational complexity as
higher-order tensors are used. In this paper, we address this issue by reducing
the $SO(3)$ convolutions or tensor products to mathematically equivalent
convolutions in $SO(2)$ . This is accomplished by aligning the node embeddings'
primary axis with the edge vectors, which sparsifies the tensor product and
reduces the computational complexity from $O(L^6)$ to $O(L^3)$, where $L$ is
the degree of the representation. We demonstrate the potential implications of
this improvement by proposing the Equivariant Spherical Channel Network (eSCN),
a graph neural network utilizing our novel approach to equivariant
convolutions, which achieves state-of-the-art results on the large-scale OC-20
and OC-22 datasets.
| [
{
"version": "v1",
"created": "Tue, 7 Feb 2023 18:16:13 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 14:07:05 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Passaro",
"Saro",
""
],
[
"Zitnick",
"C. Lawrence",
""
]
] | TITLE: Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNs
ABSTRACT: Graph neural networks that model 3D data, such as point clouds or atoms, are
typically desired to be $SO(3)$ equivariant, i.e., equivariant to 3D rotations.
Unfortunately equivariant convolutions, which are a fundamental operation for
equivariant networks, increase significantly in computational complexity as
higher-order tensors are used. In this paper, we address this issue by reducing
the $SO(3)$ convolutions or tensor products to mathematically equivalent
convolutions in $SO(2)$ . This is accomplished by aligning the node embeddings'
primary axis with the edge vectors, which sparsifies the tensor product and
reduces the computational complexity from $O(L^6)$ to $O(L^3)$, where $L$ is
the degree of the representation. We demonstrate the potential implications of
this improvement by proposing the Equivariant Spherical Channel Network (eSCN),
a graph neural network utilizing our novel approach to equivariant
convolutions, which achieves state-of-the-art results on the large-scale OC-20
and OC-22 datasets.
|
2302.05052 | Qing Zhang | Qing Zhang, Xiaoying Zhang, Yang Liu, Hongning Wang, Min Gao, Jiheng
Zhang, Ruocheng Guo | Debiasing Recommendation by Learning Identifiable Latent Confounders | null | null | null | null | cs.LG cs.IR | http://creativecommons.org/licenses/by/4.0/ | Recommendation systems aim to predict users' feedback on items not exposed to
them.
Confounding bias arises due to the presence of unmeasured variables (e.g.,
the socio-economic status of a user) that can affect both a user's exposure and
feedback. Existing methods either (1) make untenable assumptions about these
unmeasured variables or (2) directly infer latent confounders from users'
exposure. However, they cannot guarantee the identification of counterfactual
feedback, which can lead to biased predictions. In this work, we propose a
novel method, i.e., identifiable deconfounder (iDCF), which leverages a set of
proxy variables (e.g., observed user features) to resolve the aforementioned
non-identification issue. The proposed iDCF is a general deconfounded
recommendation framework that applies proximal causal inference to infer the
unmeasured confounders and identify the counterfactual feedback with
theoretical guarantees. Extensive experiments on various real-world and
synthetic datasets verify the proposed method's effectiveness and robustness.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2023 05:10:26 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 08:21:32 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Zhang",
"Qing",
""
],
[
"Zhang",
"Xiaoying",
""
],
[
"Liu",
"Yang",
""
],
[
"Wang",
"Hongning",
""
],
[
"Gao",
"Min",
""
],
[
"Zhang",
"Jiheng",
""
],
[
"Guo",
"Ruocheng",
""
]
] | TITLE: Debiasing Recommendation by Learning Identifiable Latent Confounders
ABSTRACT: Recommendation systems aim to predict users' feedback on items not exposed to
them.
Confounding bias arises due to the presence of unmeasured variables (e.g.,
the socio-economic status of a user) that can affect both a user's exposure and
feedback. Existing methods either (1) make untenable assumptions about these
unmeasured variables or (2) directly infer latent confounders from users'
exposure. However, they cannot guarantee the identification of counterfactual
feedback, which can lead to biased predictions. In this work, we propose a
novel method, i.e., identifiable deconfounder (iDCF), which leverages a set of
proxy variables (e.g., observed user features) to resolve the aforementioned
non-identification issue. The proposed iDCF is a general deconfounded
recommendation framework that applies proximal causal inference to infer the
unmeasured confounders and identify the counterfactual feedback with
theoretical guarantees. Extensive experiments on various real-world and
synthetic datasets verify the proposed method's effectiveness and robustness.
|
2302.06827 | Rahul Rathnakumar | Rahul Rathnakumar, Yutian Pang, Yongming Liu | B-BACN: Bayesian Boundary-Aware Convolutional Network for Crack
Characterization | 34 pages, 10 figures, 4 tables | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurately detecting crack boundaries is crucial for reliability assessment
and risk management of structures and materials, such as structural health
monitoring, diagnostics, prognostics, and maintenance scheduling. Uncertainty
quantification of crack detection is challenging due to various stochastic
factors, such as measurement noises, signal processing, and model
simplifications. A machine learning-based approach is proposed to quantify both
epistemic and aleatoric uncertainties concurrently. We introduce a Bayesian
Boundary-Aware Convolutional Network (B-BACN) that emphasizes uncertainty-aware
boundary refinement to generate precise and reliable crack boundary detections.
The proposed method employs a multi-task learning approach, where we use Monte
Carlo Dropout to learn the epistemic uncertainty and a Gaussian sampling
function to predict each sample's aleatoric uncertainty. Moreover, we include a
boundary refinement loss to B-BACN to enhance the determination of defect
boundaries. The proposed method is demonstrated with benchmark experimental
results and compared with several existing methods. The experimental results
illustrate the effectiveness of our proposed approach in uncertainty-aware
crack boundary detection, minimizing misclassification rate, and improving
model calibration capabilities.
| [
{
"version": "v1",
"created": "Tue, 14 Feb 2023 04:50:42 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Feb 2023 12:45:25 GMT"
},
{
"version": "v3",
"created": "Thu, 15 Jun 2023 04:37:25 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Rathnakumar",
"Rahul",
""
],
[
"Pang",
"Yutian",
""
],
[
"Liu",
"Yongming",
""
]
] | TITLE: B-BACN: Bayesian Boundary-Aware Convolutional Network for Crack
Characterization
ABSTRACT: Accurately detecting crack boundaries is crucial for reliability assessment
and risk management of structures and materials, such as structural health
monitoring, diagnostics, prognostics, and maintenance scheduling. Uncertainty
quantification of crack detection is challenging due to various stochastic
factors, such as measurement noises, signal processing, and model
simplifications. A machine learning-based approach is proposed to quantify both
epistemic and aleatoric uncertainties concurrently. We introduce a Bayesian
Boundary-Aware Convolutional Network (B-BACN) that emphasizes uncertainty-aware
boundary refinement to generate precise and reliable crack boundary detections.
The proposed method employs a multi-task learning approach, where we use Monte
Carlo Dropout to learn the epistemic uncertainty and a Gaussian sampling
function to predict each sample's aleatoric uncertainty. Moreover, we include a
boundary refinement loss to B-BACN to enhance the determination of defect
boundaries. The proposed method is demonstrated with benchmark experimental
results and compared with several existing methods. The experimental results
illustrate the effectiveness of our proposed approach in uncertainty-aware
crack boundary detection, minimizing misclassification rate, and improving
model calibration capabilities.
|
2302.08191 | Xuheng Cai | Xuheng Cai, Chao Huang, Lianghao Xia, Xubin Ren | LightGCL: Simple Yet Effective Graph Contrastive Learning for
Recommendation | Eleventh International Conference on Learning Representation (ICLR)
2023 | null | null | null | cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Graph neural network (GNN) is a powerful learning approach for graph-based
recommender systems. Recently, GNNs integrated with contrastive learning have
shown superior performance in recommendation with their data augmentation
schemes, aiming at dealing with highly sparse data. Despite their success, most
existing graph contrastive learning methods either perform stochastic
augmentation (e.g., node/edge perturbation) on the user-item interaction graph,
or rely on the heuristic-based augmentation techniques (e.g., user clustering)
for generating contrastive views. We argue that these methods cannot well
preserve the intrinsic semantic structures and are easily biased by the noise
perturbation. In this paper, we propose a simple yet effective graph
contrastive learning paradigm LightGCL that mitigates these issues impairing
the generality and robustness of CL-based recommenders. Our model exclusively
utilizes singular value decomposition for contrastive augmentation, which
enables the unconstrained structural refinement with global collaborative
relation modeling. Experiments conducted on several benchmark datasets
demonstrate the significant improvement in performance of our model over the
state-of-the-arts. Further analyses demonstrate the superiority of LightGCL's
robustness against data sparsity and popularity bias. The source code of our
model is available at https://github.com/HKUDS/LightGCL.
| [
{
"version": "v1",
"created": "Thu, 16 Feb 2023 10:16:21 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Feb 2023 04:08:47 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Jun 2023 14:25:15 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Cai",
"Xuheng",
""
],
[
"Huang",
"Chao",
""
],
[
"Xia",
"Lianghao",
""
],
[
"Ren",
"Xubin",
""
]
] | TITLE: LightGCL: Simple Yet Effective Graph Contrastive Learning for
Recommendation
ABSTRACT: Graph neural network (GNN) is a powerful learning approach for graph-based
recommender systems. Recently, GNNs integrated with contrastive learning have
shown superior performance in recommendation with their data augmentation
schemes, aiming at dealing with highly sparse data. Despite their success, most
existing graph contrastive learning methods either perform stochastic
augmentation (e.g., node/edge perturbation) on the user-item interaction graph,
or rely on the heuristic-based augmentation techniques (e.g., user clustering)
for generating contrastive views. We argue that these methods cannot well
preserve the intrinsic semantic structures and are easily biased by the noise
perturbation. In this paper, we propose a simple yet effective graph
contrastive learning paradigm LightGCL that mitigates these issues impairing
the generality and robustness of CL-based recommenders. Our model exclusively
utilizes singular value decomposition for contrastive augmentation, which
enables the unconstrained structural refinement with global collaborative
relation modeling. Experiments conducted on several benchmark datasets
demonstrate the significant improvement in performance of our model over the
state-of-the-arts. Further analyses demonstrate the superiority of LightGCL's
robustness against data sparsity and popularity bias. The source code of our
model is available at https://github.com/HKUDS/LightGCL.
|
2302.08582 | Tomek Korbak | Tomasz Korbak and Kejian Shi and Angelica Chen and Rasika Bhalerao and
Christopher L. Buckley and Jason Phang and Samuel R. Bowman and Ethan Perez | Pretraining Language Models with Human Preferences | ICML 2023 | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Language models (LMs) are pretrained to imitate internet text, including
content that would violate human preferences if generated by an LM: falsehoods,
offensive comments, personally identifiable information, low-quality or buggy
code, and more. Here, we explore alternative objectives for pretraining LMs in
a way that also guides them to generate text aligned with human preferences. We
benchmark five objectives for pretraining with human feedback across three
tasks and study how they affect the trade-off between alignment and
capabilities of pretrained LMs. We find a Pareto-optimal and simple approach
among those we explored: conditional training, or learning distribution over
tokens conditional on their human preference scores given by a reward model.
Conditional training reduces the rate of undesirable content by up to an order
of magnitude, both when generating without a prompt and with an
adversarially-chosen prompt. Moreover, conditional training maintains the
downstream task performance of standard LM pretraining, both before and after
task-specific finetuning. Pretraining with human feedback results in much
better preference satisfaction than standard LM pretraining followed by
finetuning with feedback, i.e., learning and then unlearning undesirable
behavior. Our results suggest that we should move beyond imitation learning
when pretraining LMs and incorporate human preferences from the start of
training.
| [
{
"version": "v1",
"created": "Thu, 16 Feb 2023 21:03:33 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Jun 2023 13:27:58 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Korbak",
"Tomasz",
""
],
[
"Shi",
"Kejian",
""
],
[
"Chen",
"Angelica",
""
],
[
"Bhalerao",
"Rasika",
""
],
[
"Buckley",
"Christopher L.",
""
],
[
"Phang",
"Jason",
""
],
[
"Bowman",
"Samuel R.",
""
],
[
"Perez",
"Ethan",
""
]
] | TITLE: Pretraining Language Models with Human Preferences
ABSTRACT: Language models (LMs) are pretrained to imitate internet text, including
content that would violate human preferences if generated by an LM: falsehoods,
offensive comments, personally identifiable information, low-quality or buggy
code, and more. Here, we explore alternative objectives for pretraining LMs in
a way that also guides them to generate text aligned with human preferences. We
benchmark five objectives for pretraining with human feedback across three
tasks and study how they affect the trade-off between alignment and
capabilities of pretrained LMs. We find a Pareto-optimal and simple approach
among those we explored: conditional training, or learning distribution over
tokens conditional on their human preference scores given by a reward model.
Conditional training reduces the rate of undesirable content by up to an order
of magnitude, both when generating without a prompt and with an
adversarially-chosen prompt. Moreover, conditional training maintains the
downstream task performance of standard LM pretraining, both before and after
task-specific finetuning. Pretraining with human feedback results in much
better preference satisfaction than standard LM pretraining followed by
finetuning with feedback, i.e., learning and then unlearning undesirable
behavior. Our results suggest that we should move beyond imitation learning
when pretraining LMs and incorporate human preferences from the start of
training.
|
2302.09178 | Jiaxi Tang | Jiaxi Tang, Yoel Drori, Daryl Chang, Maheswaran Sathiamoorthy, Justin
Gilmer, Li Wei, Xinyang Yi, Lichan Hong, Ed H. Chi | Improving Training Stability for Multitask Ranking Models in Recommender
Systems | Accepted at KDD 2023; 12 pages | null | 10.1145/3580305.3599846 | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recommender systems play an important role in many content platforms. While
most recommendation research is dedicated to designing better models to improve
user experience, we found that research on stabilizing the training for such
models is severely under-explored. As recommendation models become larger and
more sophisticated, they are more susceptible to training instability issues,
i.e., loss divergence, which can make the model unusable, waste significant
resources and block model developments. In this paper, we share our findings
and best practices we learned for improving the training stability of a
real-world multitask ranking model for YouTube recommendations. We show some
properties of the model that lead to unstable training and conjecture on the
causes. Furthermore, based on our observations of training dynamics near the
point of training instability, we hypothesize why existing solutions would
fail, and propose a new algorithm to mitigate the limitations of existing
solutions. Our experiments on YouTube production dataset show the proposed
algorithm can significantly improve training stability while not compromising
convergence, comparing with several commonly used baseline methods.
| [
{
"version": "v1",
"created": "Fri, 17 Feb 2023 23:04:56 GMT"
},
{
"version": "v2",
"created": "Thu, 15 Jun 2023 16:28:12 GMT"
}
] | 2023-06-16T00:00:00 | [
[
"Tang",
"Jiaxi",
""
],
[
"Drori",
"Yoel",
""
],
[
"Chang",
"Daryl",
""
],
[
"Sathiamoorthy",
"Maheswaran",
""
],
[
"Gilmer",
"Justin",
""
],
[
"Wei",
"Li",
""
],
[
"Yi",
"Xinyang",
""
],
[
"Hong",
"Lichan",
""
],
[
"Chi",
"Ed H.",
""
]
] | TITLE: Improving Training Stability for Multitask Ranking Models in Recommender
Systems
ABSTRACT: Recommender systems play an important role in many content platforms. While
most recommendation research is dedicated to designing better models to improve
user experience, we found that research on stabilizing the training for such
models is severely under-explored. As recommendation models become larger and
more sophisticated, they are more susceptible to training instability issues,
i.e., loss divergence, which can make the model unusable, waste significant
resources and block model developments. In this paper, we share our findings
and best practices we learned for improving the training stability of a
real-world multitask ranking model for YouTube recommendations. We show some
properties of the model that lead to unstable training and conjecture on the
causes. Furthermore, based on our observations of training dynamics near the
point of training instability, we hypothesize why existing solutions would
fail, and propose a new algorithm to mitigate the limitations of existing
solutions. Our experiments on YouTube production dataset show the proposed
algorithm can significantly improve training stability while not compromising
convergence, comparing with several commonly used baseline methods.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.