Search | arXiv e-print repository

<!DOCTYPE html> <html lang="en"> <head> <meta charset="utf-8"/> <meta name="viewport" content="width=device-width, initial-scale=1"/>  <link rel="apple-touch-icon" sizes="180x180" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/apple-touch-icon.png"> <link rel="icon" type="image/png" sizes="32x32" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/favicon-32x32.png"> <link rel="icon" type="image/png" sizes="16x16" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/favicon-16x16.png"> <link rel="manifest" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/site.webmanifest"> <link rel="mask-icon" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/safari-pinned-tab.svg" color="#b31b1b"> <link rel="shortcut icon" href="https://static.arxiv.org/static/base/1.0.0a5/images/icons/favicon.ico"> <meta name="msapplication-TileColor" content="#b31b1b"> <meta name="msapplication-config" content="images/icons/browserconfig.xml"> <meta name="theme-color" content="#b31b1b">  <title>Search | arXiv e-print repository</title> <script defer src="https://static.arxiv.org/static/base/1.0.0a5/fontawesome-free-5.11.2-web/js/all.js"></script> <link rel="stylesheet" href="https://static.arxiv.org/static/base/1.0.0a5/css/arxivstyle.css" /> <script type="text/x-mathjax-config"> MathJax.Hub.Config({ messageStyle: "none", extensions: ["tex2jax.js"], jax: ["input/TeX", "output/HTML-CSS"], tex2jax: { inlineMath: [ ['$','$'], ["\$","\$"] ], displayMath: [ ['$$','$$'], ["\\[","\\]"] ], processEscapes: true, ignoreClass: '.*', processClass: 'mathjax.*' }, TeX: { extensions: ["AMSmath.js", "AMSsymbols.js", "noErrors.js"], noErrors: { inlineDelimiters: ["$","$"], multiLine: false, style: { "font-size": "normal", "border": "" } } }, "HTML-CSS": { availableFonts: ["TeX"] } }); </script> <script src='//static.arxiv.org/MathJax-2.7.3/MathJax.js'></script> <script src="https://static.arxiv.org/static/base/1.0.0a5/js/notification.js"></script> <link rel="stylesheet" href="https://static.arxiv.org/static/search/0.5.6/css/bulma-tooltip.min.css" /> <link rel="stylesheet" href="https://static.arxiv.org/static/search/0.5.6/css/search.css" /> <script src="https://code.jquery.com/jquery-3.2.1.slim.min.js" integrity="sha256-k2WSCIexGzOj3Euiig+TlR8gA0EmPjuc79OEeY5L45g=" crossorigin="anonymous"></script> <script src="https://static.arxiv.org/static/search/0.5.6/js/fieldset.js"></script> <style> radio#cf-customfield_11400 { display: none; } </style> </head> <body> <header><a href="#main-container" class="is-sr-only">Skip to main content</a>  <div class="attribution level is-marginless" role="banner"> <div class="level-left"> <a class="level-item" href="https://cornell.edu/"><img src="https://static.arxiv.org/static/base/1.0.0a5/images/cornell-reduced-white-SMALL.svg" alt="Cornell University" width="200" aria-label="logo" /></a> </div> <div class="level-right is-marginless"><p class="sponsors level-item is-marginless"><span id="support-ack-url">We gratefully acknowledge support from<br /> the Simons Foundation, <a href="https://info.arxiv.org/about/ourmembers.html">member institutions</a>, and all contributors. <a href="https://info.arxiv.org/about/donate.html">Donate</a></span></p></div> </div>  <div class="identity level is-marginless"> <div class="level-left"> <div class="level-item"> <a class="arxiv" href="https://arxiv.org/" aria-label="arxiv-logo"> <img src="https://static.arxiv.org/static/base/1.0.0a5/images/arxiv-logo-one-color-white.svg" aria-label="logo" alt="arxiv logo" width="85" style="width:85px;"/> </a> </div> </div> <div class="search-block level-right"> <form class="level-item mini-search" method="GET" action="https://arxiv.org/search"> <div class="field has-addons"> <div class="control"> <input class="input is-small" type="text" name="query" placeholder="Search..." aria-label="Search term or terms" /> <p class="help"><a href="https://info.arxiv.org/help">Help</a> | <a href="https://arxiv.org/search/advanced">Advanced Search</a></p> </div> <div class="control"> <div class="select is-small"> <select name="searchtype" aria-label="Field to search"> <option value="all" selected="selected">All fields</option> <option value="title">Title</option> <option value="author">Author</option> <option value="abstract">Abstract</option> <option value="comments">Comments</option> <option value="journal_ref">Journal reference</option> <option value="acm_class">ACM classification</option> <option value="msc_class">MSC classification</option> <option value="report_num">Report number</option> <option value="paper_id">arXiv identifier</option> <option value="doi">DOI</option> <option value="orcid">ORCID</option> <option value="author_id">arXiv author ID</option> <option value="help">Help pages</option> <option value="full_text">Full text</option> </select> </div> </div> <input type="hidden" name="source" value="header"> <button class="button is-small is-cul-darker">Search</button> </div> </form> </div> </div>  <div class="container"> <div class="user-tools is-size-7 has-text-right has-text-weight-bold" role="navigation" aria-label="User menu"> <a href="https://arxiv.org/login">Login</a> </div> </div> </header> <main class="container" id="main-container"> <div class="level is-marginless"> <div class="level-left"> <h1 class="title is-clearfix"> Showing 1–50 of 136 results for author: <span class="mathjax">Woo, S</span> </h1> </div> <div class="level-right is-hidden-mobile">  <span class="help" style="display: inline-block;"><a href="https://github.com/arXiv/arxiv-search/releases">Search v0.5.6 released 2020-02-24</a>  </span> </div> </div> <div class="content"> <form method="GET" action="/search/cs" aria-role="search"> Searching in archive <strong>cs</strong>. <a href="/search/?searchtype=author&query=Woo%2C+S">Search in all archives.</a> <div class="field has-addons-tablet"> <div class="control is-expanded"> <label for="query" class="hidden-label">Search term or terms</label> <input class="input is-medium" id="query" name="query" placeholder="Search term..." type="text" value="Woo, S"> </div> <div class="select control is-medium"> <label class="is-hidden" for="searchtype">Field</label> <select class="is-medium" id="searchtype" name="searchtype"><option value="all">All fields</option><option value="title">Title</option><option selected value="author">Author(s)</option><option value="abstract">Abstract</option><option value="comments">Comments</option><option value="journal_ref">Journal reference</option><option value="acm_class">ACM classification</option><option value="msc_class">MSC classification</option><option value="report_num">Report number</option><option value="paper_id">arXiv identifier</option><option value="doi">DOI</option><option value="orcid">ORCID</option><option value="license">License (URI)</option><option value="author_id">arXiv author ID</option><option value="help">Help pages</option><option value="full_text">Full text</option></select> </div> <div class="control"> <button class="button is-link is-medium">Search</button> </div> </div> <div class="field"> <div class="control is-size-7"> <label class="radio"> <input checked id="abstracts-0" name="abstracts" type="radio" value="show"> Show abstracts </label> <label class="radio"> <input id="abstracts-1" name="abstracts" type="radio" value="hide"> Hide abstracts </label> </div> </div> <div class="is-clearfix" style="height: 2.5em"> <div class="is-pulled-right"> <a href="/search/advanced?terms-0-term=Woo%2C+S&terms-0-field=author&size=50&order=-announced_date_first">Advanced Search</a> </div> </div> <input type="hidden" name="order" value="-announced_date_first"> <input type="hidden" name="size" value="50"> </form> <div class="level breathe-horizontal"> <div class="level-left"> <form method="GET" action="/search/"> <div style="display: none;"> <select id="searchtype" name="searchtype"><option value="all">All fields</option><option value="title">Title</option><option selected value="author">Author(s)</option><option value="abstract">Abstract</option><option value="comments">Comments</option><option value="journal_ref">Journal reference</option><option value="acm_class">ACM classification</option><option value="msc_class">MSC classification</option><option value="report_num">Report number</option><option value="paper_id">arXiv identifier</option><option value="doi">DOI</option><option value="orcid">ORCID</option><option value="license">License (URI)</option><option value="author_id">arXiv author ID</option><option value="help">Help pages</option><option value="full_text">Full text</option></select> <input id="query" name="query" type="text" value="Woo, S"> <ul id="abstracts"><li><input checked id="abstracts-0" name="abstracts" type="radio" value="show"> <label for="abstracts-0">Show abstracts</label></li><li><input id="abstracts-1" name="abstracts" type="radio" value="hide"> <label for="abstracts-1">Hide abstracts</label></li></ul> </div> <div class="box field is-grouped is-grouped-multiline level-item"> <div class="control"> <span class="select is-small"> <select id="size" name="size"><option value="25">25</option><option selected value="50">50</option><option value="100">100</option><option value="200">200</option></select> </span> <label for="size">results per page</label>. </div> <div class="control"> <label for="order">Sort results by</label> <span class="select is-small"> <select id="order" name="order"><option selected value="-announced_date_first">Announcement date (newest first)</option><option value="announced_date_first">Announcement date (oldest first)</option><option value="-submitted_date">Submission date (newest first)</option><option value="submitted_date">Submission date (oldest first)</option><option value="">Relevance</option></select> </span> </div> <div class="control"> <button class="button is-small is-link">Go</button> </div> </div> </form> </div> </div> <nav class="pagination is-small is-centered breathe-horizontal" role="navigation" aria-label="pagination"> <a href="" class="pagination-previous is-invisible">Previous </a> <a href="/search/?searchtype=author&query=Woo%2C+S&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> </ul> </nav> <ol class="breathe-horizontal" start="1"> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.15224">arXiv:2411.15224</a> <span> [<a href="https://arxiv.org/pdf/2411.15224">pdf</a>, <a href="https://arxiv.org/format/2411.15224">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Parameter Efficient Mamba Tuning via Projector-targeted Diagonal-centric Linear Transformation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Ham%2C+S">Seokil Ham</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hee-Seon Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.15224v1-abstract-short" style="display: inline;"> Despite the growing interest in Mamba architecture as a potential replacement for Transformer architecture, parameter-efficient fine-tuning (PEFT) approaches for Mamba remain largely unexplored. In our study, we introduce two key insights-driven strategies for PEFT in Mamba architecture: (1) While state-space models (SSMs) have been regarded as the cornerstone of Mamba architecture, then expected… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.15224v1-abstract-full').style.display = 'inline'; document.getElementById('2411.15224v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.15224v1-abstract-full" style="display: none;"> Despite the growing interest in Mamba architecture as a potential replacement for Transformer architecture, parameter-efficient fine-tuning (PEFT) approaches for Mamba remain largely unexplored. In our study, we introduce two key insights-driven strategies for PEFT in Mamba architecture: (1) While state-space models (SSMs) have been regarded as the cornerstone of Mamba architecture, then expected to play a primary role in transfer learning, our findings reveal that Projectors -- not SSMs -- are the predominant contributors to transfer learning, and (2) Based on our observation that adapting pretrained Projectors to new tasks can be effectively approximated through a near-diagonal linear transformation, we propose a novel PEFT method specialized to Mamba architecture: Projector-targeted Diagonal-centric Linear Transformation (ProDiaL). ProDiaL focuses on optimizing only diagonal-centric linear transformation matrices, without directly fine-tuning the pretrained Projector weights. This targeted approach allows efficient task adaptation, utilizing less than 1% of the total parameters, and exhibits strong performance across both vision and language Mamba models, highlighting its versatility and effectiveness. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.15224v1-abstract-full').style.display = 'none'; document.getElementById('2411.15224v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.22735">arXiv:2410.22735</a> <span> [<a href="https://arxiv.org/pdf/2410.22735">pdf</a>, <a href="https://arxiv.org/format/2410.22735">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> MIXAD: Memory-Induced Explainable Time Series Anomaly Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+M">Minha Kim</a>, <a href="/search/cs?searchtype=author&query=Bhaumik%2C+K+K">Kishor Kumar Bhaumik</a>, <a href="/search/cs?searchtype=author&query=Ali%2C+A+A">Amin Ahsan Ali</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.22735v1-abstract-short" style="display: inline;"> For modern industrial applications, accurately detecting and diagnosing anomalies in multivariate time series data is essential. Despite such need, most state-of-the-art methods often prioritize detection performance over model interpretability. Addressing this gap, we introduce MIXAD (Memory-Induced Explainable Time Series Anomaly Detection), a model designed for interpretable anomaly detection.… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22735v1-abstract-full').style.display = 'inline'; document.getElementById('2410.22735v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.22735v1-abstract-full" style="display: none;"> For modern industrial applications, accurately detecting and diagnosing anomalies in multivariate time series data is essential. Despite such need, most state-of-the-art methods often prioritize detection performance over model interpretability. Addressing this gap, we introduce MIXAD (Memory-Induced Explainable Time Series Anomaly Detection), a model designed for interpretable anomaly detection. MIXAD leverages a memory network alongside spatiotemporal processing units to understand the intricate dynamics and topological structures inherent in sensor relationships. We also introduce a novel anomaly scoring method that detects significant shifts in memory activation patterns during anomalies. Our approach not only ensures decent detection performance but also outperforms state-of-the-art baselines by 34.30% and 34.51% in interpretability metrics. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22735v1-abstract-full').style.display = 'none'; document.getElementById('2410.22735v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 30 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ICPR 2024 (oral paper)</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.19341">arXiv:2410.19341</a> <span> [<a href="https://arxiv.org/pdf/2410.19341">pdf</a>, <a href="https://arxiv.org/format/2410.19341">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Context-Based Visual-Language Place Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Soojin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+S">Seong-Woo Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.19341v1-abstract-short" style="display: inline;"> In vision-based robot localization and SLAM, Visual Place Recognition (VPR) is essential. This paper addresses the problem of VPR, which involves accurately recognizing the location corresponding to a given query image. A popular approach to vision-based place recognition relies on low-level visual features. Despite significant progress in recent years, place recognition based on low-level visual… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.19341v1-abstract-full').style.display = 'inline'; document.getElementById('2410.19341v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.19341v1-abstract-full" style="display: none;"> In vision-based robot localization and SLAM, Visual Place Recognition (VPR) is essential. This paper addresses the problem of VPR, which involves accurately recognizing the location corresponding to a given query image. A popular approach to vision-based place recognition relies on low-level visual features. Despite significant progress in recent years, place recognition based on low-level visual features is challenging when there are changes in scene appearance. To address this, end-to-end training approaches have been proposed to overcome the limitations of hand-crafted features. However, these approaches still fail under drastic changes and require large amounts of labeled data to train models, presenting a significant limitation. Methods that leverage high-level semantic information, such as objects or categories, have been proposed to handle variations in appearance. In this paper, we introduce a novel VPR approach that remains robust to scene changes and does not require additional training. Our method constructs semantic image descriptors by extracting pixel-level embeddings using a zero-shot, language-driven semantic segmentation model. We validate our approach in challenging place recognition scenarios using real-world public dataset. The experiments demonstrate that our method outperforms non-learned image representation techniques and off-the-shelf convolutional neural network (CNN) descriptors. Our code is available at https: //github.com/woo-soojin/context-based-vlpr. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.19341v1-abstract-full').style.display = 'none'; document.getElementById('2410.19341v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 25 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.15589">arXiv:2410.15589</a> <span> [<a href="https://arxiv.org/pdf/2410.15589">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> SSMT: Few-Shot Traffic Forecasting with Single Source Meta-Transfer </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Bhaumik%2C+K+K">Kishor Kumar Bhaumik</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+M">Minha Kim</a>, <a href="/search/cs?searchtype=author&query=Niloy%2C+F+F">Fahim Faisal Niloy</a>, <a href="/search/cs?searchtype=author&query=Ali%2C+A+A">Amin Ahsan Ali</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.15589v1-abstract-short" style="display: inline;"> Traffic forecasting in Intelligent Transportation Systems (ITS) is vital for intelligent traffic prediction. Yet, ITS often relies on data from traffic sensors or vehicle devices, where certain cities might not have all those smart devices or enabling infrastructures. Also, recent studies have employed meta-learning to generalize spatial-temporal traffic networks, utilizing data from multiple citi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.15589v1-abstract-full').style.display = 'inline'; document.getElementById('2410.15589v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.15589v1-abstract-full" style="display: none;"> Traffic forecasting in Intelligent Transportation Systems (ITS) is vital for intelligent traffic prediction. Yet, ITS often relies on data from traffic sensors or vehicle devices, where certain cities might not have all those smart devices or enabling infrastructures. Also, recent studies have employed meta-learning to generalize spatial-temporal traffic networks, utilizing data from multiple cities for effective traffic forecasting for data-scarce target cities. However, collecting data from multiple cities can be costly and time-consuming. To tackle this challenge, we introduce Single Source Meta-Transfer Learning (SSMT) which relies only on a single source city for traffic prediction. Our method harnesses this transferred knowledge to enable few-shot traffic forecasting, particularly when the target city possesses limited data. Specifically, we use memory-augmented attention to store the heterogeneous spatial knowledge from the source city and selectively recall them for the data-scarce target city. We extend the idea of sinusoidal positional encoding to establish meta-learning tasks by leveraging diverse temporal traffic patterns from the source city. Moreover, to capture a more generalized representation of the positions we introduced a meta-positional encoding that learns the most optimal representation of the temporal pattern across all the tasks. We experiment on five real-world benchmark datasets to demonstrate that our method outperforms several existing methods in time series traffic prediction. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.15589v1-abstract-full').style.display = 'none'; document.getElementById('2410.15589v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ICPR 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.09831">arXiv:2410.09831</a> <span> [<a href="https://arxiv.org/pdf/2410.09831">pdf</a>, <a href="https://arxiv.org/format/2410.09831">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computational Engineering, Finance, and Science">cs.CE</span> </div> </div> <p class="title is-5 mathjax"> LoLI-Street: Benchmarking Low-Light Image Enhancement and Beyond </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Islam%2C+M+T">Md Tanvir Islam</a>, <a href="/search/cs?searchtype=author&query=Alam%2C+I">Inzamamul Alam</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a>, <a href="/search/cs?searchtype=author&query=Anwar%2C+S">Saeed Anwar</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+I+H">IK Hyun Lee</a>, <a href="/search/cs?searchtype=author&query=Muhammad%2C+K">Khan Muhammad</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.09831v1-abstract-short" style="display: inline;"> Low-light image enhancement (LLIE) is essential for numerous computer vision tasks, including object detection, tracking, segmentation, and scene understanding. Despite substantial research on improving low-quality images captured in underexposed conditions, clear vision remains critical for autonomous vehicles, which often struggle with low-light scenarios, signifying the need for continuous rese… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.09831v1-abstract-full').style.display = 'inline'; document.getElementById('2410.09831v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.09831v1-abstract-full" style="display: none;"> Low-light image enhancement (LLIE) is essential for numerous computer vision tasks, including object detection, tracking, segmentation, and scene understanding. Despite substantial research on improving low-quality images captured in underexposed conditions, clear vision remains critical for autonomous vehicles, which often struggle with low-light scenarios, signifying the need for continuous research. However, paired datasets for LLIE are scarce, particularly for street scenes, limiting the development of robust LLIE methods. Despite using advanced transformers and/or diffusion-based models, current LLIE methods struggle in real-world low-light conditions and lack training on street-scene datasets, limiting their effectiveness for autonomous vehicles. To bridge these gaps, we introduce a new dataset LoLI-Street (Low-Light Images of Streets) with 33k paired low-light and well-exposed images from street scenes in developed cities, covering 19k object classes for object detection. LoLI-Street dataset also features 1,000 real low-light test images for testing LLIE models under real-life conditions. Furthermore, we propose a transformer and diffusion-based LLIE model named "TriFuse". Leveraging the LoLI-Street dataset, we train and evaluate our TriFuse and SOTA models to benchmark on our dataset. Comparing various models, our dataset's generalization feasibility is evident in testing across different mainstream datasets by significantly enhancing images and object detection for practical applications in autonomous driving and surveillance systems. The complete code and dataset is available on https://github.com/tanvirnwu/TriFuse. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.09831v1-abstract-full').style.display = 'none'; document.getElementById('2410.09831v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 13 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by the Asian Conference on Computer Vision (ACCV 2024)</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.09529">arXiv:2410.09529</a> <span> [<a href="https://arxiv.org/pdf/2410.09529">pdf</a>, <a href="https://arxiv.org/format/2410.09529">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Preserving Old Memories in Vivid Detail: Human-Interactive Photo Restoration Framework </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Back%2C+S">Seung-Yeon Back</a>, <a href="/search/cs?searchtype=author&query=Son%2C+G">Geonho Son</a>, <a href="/search/cs?searchtype=author&query=Jeong%2C+D">Dahye Jeong</a>, <a href="/search/cs?searchtype=author&query=Park%2C+E">Eunil Park</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.09529v1-abstract-short" style="display: inline;"> Photo restoration technology enables preserving visual memories in photographs. However, physical prints are vulnerable to various forms of deterioration, ranging from physical damage to loss of image quality, etc. While restoration by human experts can improve the quality of outcomes, it often comes at a high price in terms of cost and time for restoration. In this work, we present the AI-based p… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.09529v1-abstract-full').style.display = 'inline'; document.getElementById('2410.09529v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.09529v1-abstract-full" style="display: none;"> Photo restoration technology enables preserving visual memories in photographs. However, physical prints are vulnerable to various forms of deterioration, ranging from physical damage to loss of image quality, etc. While restoration by human experts can improve the quality of outcomes, it often comes at a high price in terms of cost and time for restoration. In this work, we present the AI-based photo restoration framework composed of multiple stages, where each stage is tailored to enhance and restore specific types of photo damage, accelerating and automating the photo restoration process. By integrating these techniques into a unified architecture, our framework aims to offer a one-stop solution for restoring old and deteriorated photographs. Furthermore, we present a novel old photo restoration dataset because we lack a publicly available dataset for our evaluation. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.09529v1-abstract-full').style.display = 'none'; document.getElementById('2410.09529v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 12 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.04081">arXiv:2410.04081</a> <span> [<a href="https://arxiv.org/pdf/2410.04081">pdf</a>, <a href="https://arxiv.org/format/2410.04081">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Image and Video Processing">eess.IV</span> </div> </div> <p class="title is-5 mathjax"> $蔚$-VAE: Denoising as Visual Decoding </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhao%2C+L">Long Zhao</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sanghyun Woo</a>, <a href="/search/cs?searchtype=author&query=Wan%2C+Z">Ziyu Wan</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yandong Li</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+H">Han Zhang</a>, <a href="/search/cs?searchtype=author&query=Gong%2C+B">Boqing Gong</a>, <a href="/search/cs?searchtype=author&query=Adam%2C+H">Hartwig Adam</a>, <a href="/search/cs?searchtype=author&query=Jia%2C+X">Xuhui Jia</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+T">Ting Liu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.04081v1-abstract-short" style="display: inline;"> In generative modeling, tokenization simplifies complex data into compact, structured representations, creating a more efficient, learnable space. For high-dimensional visual data, it reduces redundancy and emphasizes key features for high-quality generation. Current visual tokenization methods rely on a traditional autoencoder framework, where the encoder compresses data into latent representatio… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.04081v1-abstract-full').style.display = 'inline'; document.getElementById('2410.04081v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.04081v1-abstract-full" style="display: none;"> In generative modeling, tokenization simplifies complex data into compact, structured representations, creating a more efficient, learnable space. For high-dimensional visual data, it reduces redundancy and emphasizes key features for high-quality generation. Current visual tokenization methods rely on a traditional autoencoder framework, where the encoder compresses data into latent representations, and the decoder reconstructs the original input. In this work, we offer a new perspective by proposing denoising as decoding, shifting from single-step reconstruction to iterative refinement. Specifically, we replace the decoder with a diffusion process that iteratively refines noise to recover the original image, guided by the latents provided by the encoder. We evaluate our approach by assessing both reconstruction (rFID) and generation quality (FID), comparing it to state-of-the-art autoencoding approach. We hope this work offers new insights into integrating iterative generation and autoencoding for improved compression and generation. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.04081v1-abstract-full').style.display = 'none'; document.getElementById('2410.04081v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 5 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2409.10027">arXiv:2409.10027</a> <span> [<a href="https://arxiv.org/pdf/2409.10027">pdf</a>, <a href="https://arxiv.org/format/2409.10027">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> E2Map: Experience-and-Emotion Map for Self-Reflective Robot Navigation with Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+C">Chan Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+K">Keonwoo Kim</a>, <a href="/search/cs?searchtype=author&query=Oh%2C+M">Mintaek Oh</a>, <a href="/search/cs?searchtype=author&query=Baek%2C+H">Hanbi Baek</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jiyang Lee</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+D">Donghwi Jung</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Soojin Woo</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+Y">Younkyung Woo</a>, <a href="/search/cs?searchtype=author&query=Tucker%2C+J">John Tucker</a>, <a href="/search/cs?searchtype=author&query=Firoozi%2C+R">Roya Firoozi</a>, <a href="/search/cs?searchtype=author&query=Seo%2C+S">Seung-Woo Seo</a>, <a href="/search/cs?searchtype=author&query=Schwager%2C+M">Mac Schwager</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+S">Seong-Woo Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2409.10027v2-abstract-short" style="display: inline;"> Large language models (LLMs) have shown significant potential in guiding embodied agents to execute language instructions across a range of tasks, including robotic manipulation and navigation. However, existing methods are primarily designed for static environments and do not leverage the agent's own experiences to refine its initial plans. Given that real-world environments are inherently stocha… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.10027v2-abstract-full').style.display = 'inline'; document.getElementById('2409.10027v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2409.10027v2-abstract-full" style="display: none;"> Large language models (LLMs) have shown significant potential in guiding embodied agents to execute language instructions across a range of tasks, including robotic manipulation and navigation. However, existing methods are primarily designed for static environments and do not leverage the agent's own experiences to refine its initial plans. Given that real-world environments are inherently stochastic, initial plans based solely on LLMs' general knowledge may fail to achieve their objectives, unlike in static scenarios. To address this limitation, this study introduces the Experience-and-Emotion Map (E2Map), which integrates not only LLM knowledge but also the agent's real-world experiences, drawing inspiration from human emotional responses. The proposed methodology enables one-shot behavior adjustments by updating the E2Map based on the agent's experiences. Our evaluation in stochastic navigation environments, including both simulations and real-world scenarios, demonstrates that the proposed method significantly enhances performance in stochastic environments compared to existing LLM-based approaches. Code and supplementary materials are available at https://e2map.github.io/. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.10027v2-abstract-full').style.display = 'none'; document.getElementById('2409.10027v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 1 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 16 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">19 pages, 28 figures. Project page: https://e2map.github.io</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2409.07913">arXiv:2409.07913</a> <span> [<a href="https://arxiv.org/pdf/2409.07913">pdf</a>, <a href="https://arxiv.org/format/2409.07913">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> <div class="is-inline-block" style="margin-left: 0.5rem"> <div class="tags has-addons"> <span class="tag is-dark is-size-7">doi</span> <span class="tag is-light is-size-7"><a class="" href="https://doi.org/10.1145/3627673.3680085">10.1145/3627673.3680085 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> UGAD: Universal Generative AI Detector utilizing Frequency Fingerprints </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Alam%2C+I">Inzamamul Alam</a>, <a href="/search/cs?searchtype=author&query=Muneer%2C+M+S">Muhammad Shahid Muneer</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2409.07913v1-abstract-short" style="display: inline;"> In the wake of a fabricated explosion image at the Pentagon, an ability to discern real images from fake counterparts has never been more critical. Our study introduces a novel multi-modal approach to detect AI-generated images amidst the proliferation of new generation methods such as Diffusion models. Our method, UGAD, encompasses three key detection steps: First, we transform the RGB images int… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.07913v1-abstract-full').style.display = 'inline'; document.getElementById('2409.07913v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2409.07913v1-abstract-full" style="display: none;"> In the wake of a fabricated explosion image at the Pentagon, an ability to discern real images from fake counterparts has never been more critical. Our study introduces a novel multi-modal approach to detect AI-generated images amidst the proliferation of new generation methods such as Diffusion models. Our method, UGAD, encompasses three key detection steps: First, we transform the RGB images into YCbCr channels and apply an Integral Radial Operation to emphasize salient radial features. Secondly, the Spatial Fourier Extraction operation is used for a spatial shift, utilizing a pre-trained deep learning network for optimal feature extraction. Finally, the deep neural network classification stage processes the data through dense layers using softmax for classification. Our approach significantly enhances the accuracy of differentiating between real and AI-generated images, as evidenced by a 12.64% increase in accuracy and 28.43% increase in AUC compared to existing state-of-the-art methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.07913v1-abstract-full').style.display = 'none'; document.getElementById('2409.07913v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 12 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2409.05346">arXiv:2409.05346</a> <span> [<a href="https://arxiv.org/pdf/2409.05346">pdf</a>, <a href="https://arxiv.org/format/2409.05346">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> GDFlow: Anomaly Detection with NCDE-based Normalizing Flow for Advanced Driver Assistance System </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lee%2C+K">Kangjun Lee</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+M">Minha Kim</a>, <a href="/search/cs?searchtype=author&query=Jun%2C+Y">Youngho Jun</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2409.05346v1-abstract-short" style="display: inline;"> For electric vehicles, the Adaptive Cruise Control (ACC) in Advanced Driver Assistance Systems (ADAS) is designed to assist braking based on driving conditions, road inclines, predefined deceleration strengths, and user braking patterns. However, the driving data collected during the development of ADAS are generally limited and lack diversity. This deficiency leads to late or aggressive braking f… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.05346v1-abstract-full').style.display = 'inline'; document.getElementById('2409.05346v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2409.05346v1-abstract-full" style="display: none;"> For electric vehicles, the Adaptive Cruise Control (ACC) in Advanced Driver Assistance Systems (ADAS) is designed to assist braking based on driving conditions, road inclines, predefined deceleration strengths, and user braking patterns. However, the driving data collected during the development of ADAS are generally limited and lack diversity. This deficiency leads to late or aggressive braking for different users. Crucially, it is necessary to effectively identify anomalies, such as unexpected or inconsistent braking patterns in ADAS, especially given the challenge of working with unlabelled, limited, and noisy datasets from real-world electric vehicles. In order to tackle the aforementioned challenges in ADAS, we propose Graph Neural Controlled Differential Equation Normalizing Flow (GDFlow), a model that leverages Normalizing Flow (NF) with Neural Controlled Differential Equations (NCDE) to learn the distribution of normal driving patterns continuously. Compared to the traditional clustering or anomaly detection algorithms, our approach effectively captures the spatio-temporal information from different sensor data and more accurately models continuous changes in driving patterns. Additionally, we introduce a quantile-based maximum likelihood objective to improve the likelihood estimate of the normal data near the boundary of the distribution, enhancing the model's ability to distinguish between normal and anomalous patterns. We validate GDFlow using real-world electric vehicle driving data that we collected from Hyundai IONIQ5 and GV80EV, achieving state-of-the-art performance compared to six baselines across four dataset configurations of different vehicle types and drivers. Furthermore, our model outperforms the latest anomaly detection methods across four time series benchmark datasets. Our approach demonstrates superior efficiency in inference time compared to existing methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.05346v1-abstract-full').style.display = 'none'; document.getElementById('2409.05346v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 9 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2409.01201">arXiv:2409.01201</a> <span> [<a href="https://arxiv.org/pdf/2409.01201">pdf</a>, <a href="https://arxiv.org/format/2409.01201">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Audio and Speech Processing">eess.AS</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Sound">cs.SD</span> </div> </div> <p class="title is-5 mathjax"> EnCLAP++: Analyzing the EnCLAP Framework for Optimizing Automated Audio Captioning Performance </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jaeyeon Kim</a>, <a href="/search/cs?searchtype=author&query=Jeon%2C+M">Minjeon Jeon</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+J">Jaeyoon Jung</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+H">Sang Hoon Woo</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jinjoo Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2409.01201v1-abstract-short" style="display: inline;"> In this work, we aim to analyze and optimize the EnCLAP framework, a state-of-the-art model in automated audio captioning. We investigate the impact of modifying the acoustic encoder components, explore pretraining with different dataset scales, and study the effectiveness of a reranking scheme. Through extensive experimentation and quantitative analysis of generated captions, we develop EnCLAP++,… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.01201v1-abstract-full').style.display = 'inline'; document.getElementById('2409.01201v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2409.01201v1-abstract-full" style="display: none;"> In this work, we aim to analyze and optimize the EnCLAP framework, a state-of-the-art model in automated audio captioning. We investigate the impact of modifying the acoustic encoder components, explore pretraining with different dataset scales, and study the effectiveness of a reranking scheme. Through extensive experimentation and quantitative analysis of generated captions, we develop EnCLAP++, an enhanced version that significantly surpasses the original. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.01201v1-abstract-full').style.display = 'none'; document.getElementById('2409.01201v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 2 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to DCASE2024 Workshop</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2409.01160">arXiv:2409.01160</a> <span> [<a href="https://arxiv.org/pdf/2409.01160">pdf</a>, <a href="https://arxiv.org/ps/2409.01160">ps</a>, <a href="https://arxiv.org/format/2409.01160">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Audio and Speech Processing">eess.AS</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Sound">cs.SD</span> </div> </div> <p class="title is-5 mathjax"> Expanding on EnCLAP with Auxiliary Retrieval Model for Automated Audio Captioning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jaeyeon Kim</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+J">Jaeyoon Jung</a>, <a href="/search/cs?searchtype=author&query=Jeon%2C+M">Minjeong Jeon</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+H">Sang Hoon Woo</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jinjoo Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2409.01160v1-abstract-short" style="display: inline;"> In this technical report, we describe our submission to DCASE2024 Challenge Task6 (Automated Audio Captioning) and Task8 (Language-based Audio Retrieval). We develop our approach building upon the EnCLAP audio captioning framework and optimizing it for Task6 of the challenge. Notably, we outline the changes in the underlying components and the incorporation of the reranking process. Additionally,… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.01160v1-abstract-full').style.display = 'inline'; document.getElementById('2409.01160v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2409.01160v1-abstract-full" style="display: none;"> In this technical report, we describe our submission to DCASE2024 Challenge Task6 (Automated Audio Captioning) and Task8 (Language-based Audio Retrieval). We develop our approach building upon the EnCLAP audio captioning framework and optimizing it for Task6 of the challenge. Notably, we outline the changes in the underlying components and the incorporation of the reranking process. Additionally, we submit a supplementary retriever model, a byproduct of our modified framework, to Task8. Our proposed systems achieve FENSE score of 0.542 on Task6 and mAP@10 score of 0.386 on Task8, significantly outperforming the baseline models. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2409.01160v1-abstract-full').style.display = 'none'; document.getElementById('2409.01160v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 2 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">DCASE2024 Challenge Technical Report. Ranked 2nd in Task 6 Automated Audio Captioning</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2408.17066">arXiv:2408.17066</a> <span> [<a href="https://arxiv.org/pdf/2408.17066">pdf</a>, <a href="https://arxiv.org/format/2408.17066">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> Non-verbal Interaction and Interface with a Quadruped Robot using Body and Hand Gestures: Design and User Experience Evaluation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Shin%2C+S">Soohyun Shin</a>, <a href="/search/cs?searchtype=author&query=Evetts%2C+T">Trevor Evetts</a>, <a href="/search/cs?searchtype=author&query=Saylor%2C+H">Hunter Saylor</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hyunji Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Soojin Woo</a>, <a href="/search/cs?searchtype=author&query=Rhee%2C+W">Wonhwha Rhee</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+S">Seong-Woo Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2408.17066v1-abstract-short" style="display: inline;"> In recent years, quadruped robots have attracted significant attention due to their practical advantages in maneuverability, particularly when navigating rough terrain and climbing stairs. As these robots become more integrated into various industries, including construction and healthcare, researchers have increasingly focused on developing intuitive interaction methods such as speech and gesture… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.17066v1-abstract-full').style.display = 'inline'; document.getElementById('2408.17066v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2408.17066v1-abstract-full" style="display: none;"> In recent years, quadruped robots have attracted significant attention due to their practical advantages in maneuverability, particularly when navigating rough terrain and climbing stairs. As these robots become more integrated into various industries, including construction and healthcare, researchers have increasingly focused on developing intuitive interaction methods such as speech and gestures that do not require separate devices such as keyboards or joysticks. This paper aims at investigating a comfortable and efficient interaction method with quadruped robots that possess a familiar form factor. To this end, we conducted two preliminary studies to observe how individuals naturally interact with a quadruped robot in natural and controlled settings, followed by a prototype experiment to examine human preferences for body-based and hand-based gesture controls using a Unitree Go1 Pro quadruped robot. We assessed the user experience of 13 participants using the User Experience Questionnaire and measured the time taken to complete specific tasks. The findings of our preliminary results indicate that humans have a natural preference for communicating with robots through hand and body gestures rather than speech. In addition, participants reported higher satisfaction and completed tasks more quickly when using body gestures to interact with the robot. This contradicts the fact that most gesture-based control technologies for quadruped robots are hand-based. The video is available at https://youtu.be/rysv1p1zvp4. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.17066v1-abstract-full').style.display = 'none'; document.getElementById('2408.17066v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 30 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">16 pages</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2408.06167">arXiv:2408.06167</a> <span> [<a href="https://arxiv.org/pdf/2408.06167">pdf</a>, <a href="https://arxiv.org/format/2408.06167">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> </div> <div class="is-inline-block" style="margin-left: 0.5rem"> <div class="tags has-addons"> <span class="tag is-dark is-size-7">doi</span> <span class="tag is-light is-size-7"><a class="" href="https://doi.org/10.1145/3627673.3680017">10.1145/3627673.3680017 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Blind-Match: Efficient Homomorphic Encryption-Based 1:N Matching for Privacy-Preserving Biometric Identification </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Choi%2C+H">Hyunmin Choi</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jiwon Kim</a>, <a href="/search/cs?searchtype=author&query=Song%2C+C">Chiyoung Song</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hyoungshick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2408.06167v2-abstract-short" style="display: inline;"> We present Blind-Match, a novel biometric identification system that leverages homomorphic encryption (HE) for efficient and privacy-preserving 1:N matching. Blind-Match introduces a HE-optimized cosine similarity computation method, where the key idea is to divide the feature vector into smaller parts for processing rather than computing the entire vector at once. By optimizing the number of thes… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.06167v2-abstract-full').style.display = 'inline'; document.getElementById('2408.06167v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2408.06167v2-abstract-full" style="display: none;"> We present Blind-Match, a novel biometric identification system that leverages homomorphic encryption (HE) for efficient and privacy-preserving 1:N matching. Blind-Match introduces a HE-optimized cosine similarity computation method, where the key idea is to divide the feature vector into smaller parts for processing rather than computing the entire vector at once. By optimizing the number of these parts, Blind-Match minimizes execution time while ensuring data privacy through HE. Blind-Match achieves superior performance compared to state-of-the-art methods across various biometric datasets. On the LFW face dataset, Blind-Match attains a 99.63% Rank-1 accuracy with a 128-dimensional feature vector, demonstrating its robustness in face recognition tasks. For fingerprint identification, Blind-Match achieves a remarkable 99.55% Rank-1 accuracy on the PolyU dataset, even with a compact 16-dimensional feature vector, significantly outperforming the state-of-the-art method, Blind-Touch, which achieves only 59.17%. Furthermore, Blind-Match showcases practical efficiency in large-scale biometric identification scenarios, such as Naver Cloud's FaceSign, by processing 6,144 biometric samples in 0.74 seconds using a 128-dimensional feature vector. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.06167v2-abstract-full').style.display = 'none'; document.getElementById('2408.06167v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 13 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 12 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to CIKM 2024 (Applied Research Track)</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.19102">arXiv:2407.19102</a> <span> [<a href="https://arxiv.org/pdf/2407.19102">pdf</a>, <a href="https://arxiv.org/format/2407.19102">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computational Complexity">cs.CC</span> </div> </div> <p class="title is-5 mathjax"> The Computational Complexity of Factored Graphs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Gupta%2C+S">Shreya Gupta</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+B">Boyang Huang</a>, <a href="/search/cs?searchtype=author&query=Impagliazzo%2C+R">Russell Impagliazzo</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Stanley Woo</a>, <a href="/search/cs?searchtype=author&query=Ye%2C+C">Christopher Ye</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.19102v2-abstract-short" style="display: inline;"> While graphs and abstract data structures can be large and complex, practical instances are often regular or highly structured. If the instance has sufficient structure, we might hope to compress the object into a more succinct representation. An efficient algorithm (with respect to the compressed input size) could then lead to more efficient computations than algorithms taking the explicit, uncom… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.19102v2-abstract-full').style.display = 'inline'; document.getElementById('2407.19102v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.19102v2-abstract-full" style="display: none;"> While graphs and abstract data structures can be large and complex, practical instances are often regular or highly structured. If the instance has sufficient structure, we might hope to compress the object into a more succinct representation. An efficient algorithm (with respect to the compressed input size) could then lead to more efficient computations than algorithms taking the explicit, uncompressed object as input. This leads to a natural question: when does knowing the input instance has a more succinct representation make computation easier? We initiate the study of the computational complexity of problems on factored graphs: graphs that are given as a formula of products and unions on smaller graphs. For any graph problem, we define a parameterized version that takes factored graphs as input, parameterized by the number of smaller graphs used to construct the factored graph. In this setting, we characterize the parameterized complexity of several natural graph problems, exhibiting a variety of complexities. We show that a decision version of lexicographically first maximal independent set is unconditionally XP-complete, and therefore not fixed parameter tractable (FPT). On the other hand, we show that clique counting is FPT. Finally, we show that reachability is FPT if and only if $\mathbf{NL}$ is in some fixed polynomial time. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.19102v2-abstract-full').style.display = 'none'; document.getElementById('2407.19102v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 5 September, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 26 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.15554">arXiv:2407.15554</a> <span> [<a href="https://arxiv.org/pdf/2407.15554">pdf</a>, <a href="https://arxiv.org/format/2407.15554">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Decomposition of Neural Discrete Representations for Large-Scale 3D Mapping </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Park%2C+M">Minseong Park</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Suhan Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+E">Euntai Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.15554v1-abstract-short" style="display: inline;"> Learning efficient representations of local features is a key challenge in feature volume-based 3D neural mapping, especially in large-scale environments. In this paper, we introduce Decomposition-based Neural Mapping (DNMap), a storage-efficient large-scale 3D mapping method that employs a discrete representation based on a decomposition strategy. This decomposition strategy aims to efficiently c… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.15554v1-abstract-full').style.display = 'inline'; document.getElementById('2407.15554v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.15554v1-abstract-full" style="display: none;"> Learning efficient representations of local features is a key challenge in feature volume-based 3D neural mapping, especially in large-scale environments. In this paper, we introduce Decomposition-based Neural Mapping (DNMap), a storage-efficient large-scale 3D mapping method that employs a discrete representation based on a decomposition strategy. This decomposition strategy aims to efficiently capture repetitive and representative patterns of shapes by decomposing each discrete embedding into component vectors that are shared across the embedding space. Our DNMap optimizes a set of component vectors, rather than entire discrete embeddings, and learns composition rather than indexing the discrete embeddings. Furthermore, to complement the mapping quality, we additionally learn low-resolution continuous embeddings that require tiny storage space. By combining these representations with a shallow neural network and an efficient octree-based feature volume, our DNMap successfully approximates signed distance functions and compresses the feature volume while preserving mapping quality. Our source code is available at https://github.com/minseong-p/dnmap. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.15554v1-abstract-full').style.display = 'none'; document.getElementById('2407.15554v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 22 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ECCV 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.11714">arXiv:2407.11714</a> <span> [<a href="https://arxiv.org/pdf/2407.11714">pdf</a>, <a href="https://arxiv.org/format/2407.11714">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Improving Unsupervised Video Object Segmentation via Fake Flow Generation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Cho%2C+S">Suhwan Cho</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+M">Minhyeok Lee</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jungho Lee</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+D">Donghyeong Kim</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Seunghoon Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sungmin Woo</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sangyoun Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.11714v1-abstract-short" style="display: inline;"> Unsupervised video object segmentation (VOS), also known as video salient object detection, aims to detect the most prominent object in a video at the pixel level. Recently, two-stream approaches that leverage both RGB images and optical flow maps have gained significant attention. However, the limited amount of training data remains a substantial challenge. In this study, we propose a novel data… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.11714v1-abstract-full').style.display = 'inline'; document.getElementById('2407.11714v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.11714v1-abstract-full" style="display: none;"> Unsupervised video object segmentation (VOS), also known as video salient object detection, aims to detect the most prominent object in a video at the pixel level. Recently, two-stream approaches that leverage both RGB images and optical flow maps have gained significant attention. However, the limited amount of training data remains a substantial challenge. In this study, we propose a novel data generation method that simulates fake optical flows from single images, thereby creating large-scale training data for stable network learning. Inspired by the observation that optical flow maps are highly dependent on depth maps, we generate fake optical flows by refining and augmenting the estimated depth maps of each image. By incorporating our simulated image-flow pairs, we achieve new state-of-the-art performance on all public benchmark datasets without relying on complex modules. We believe that our data generation method represents a potential breakthrough for future VOS research. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.11714v1-abstract-full').style.display = 'none'; document.getElementById('2407.11714v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 16 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.10784">arXiv:2407.10784</a> <span> [<a href="https://arxiv.org/pdf/2407.10784">pdf</a>, <a href="https://arxiv.org/format/2407.10784">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">stat.ML</span> </div> </div> <p class="title is-5 mathjax"> AdapTable: Test-Time Adaptation for Tabular Data via Shift-Aware Uncertainty Calibrator and Label Distribution Handler </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changhun Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+T">Taewon Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Seungyeon Woo</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+J+Y">June Yong Yang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+E">Eunho Yang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.10784v3-abstract-short" style="display: inline;"> In real-world scenarios, tabular data often suffer from distribution shifts that threaten the performance of machine learning models. Despite its prevalence and importance, handling distribution shifts in the tabular domain remains underexplored due to the inherent challenges within the tabular data itself. In this sense, test-time adaptation (TTA) offers a promising solution by adapting models to… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10784v3-abstract-full').style.display = 'inline'; document.getElementById('2407.10784v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.10784v3-abstract-full" style="display: none;"> In real-world scenarios, tabular data often suffer from distribution shifts that threaten the performance of machine learning models. Despite its prevalence and importance, handling distribution shifts in the tabular domain remains underexplored due to the inherent challenges within the tabular data itself. In this sense, test-time adaptation (TTA) offers a promising solution by adapting models to target data without accessing source data, crucial for privacy-sensitive tabular domains. However, existing TTA methods either 1) overlook the nature of tabular distribution shifts, often involving label distribution shifts, or 2) impose architectural constraints on the model, leading to a lack of applicability. To this end, we propose AdapTable, a novel TTA framework for tabular data. AdapTable operates in two stages: 1) calibrating model predictions using a shift-aware uncertainty calibrator, and 2) adjusting these predictions to match the target label distribution with a label distribution handler. We validate the effectiveness of AdapTable through theoretical analysis and extensive experiments on various distribution shift scenarios. Our results demonstrate AdapTable's ability to handle various real-world distribution shifts, achieving up to a 16% improvement on the HELOC dataset. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10784v3-abstract-full').style.display = 'none'; document.getElementById('2407.10784v3-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 15 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Under Review at AAAI 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.10399">arXiv:2407.10399</a> <span> [<a href="https://arxiv.org/pdf/2407.10399">pdf</a>, <a href="https://arxiv.org/format/2407.10399">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Exploring the Impact of Moire Pattern on Deepfake Detectors </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Tariq%2C+R">Razaib Tariq</a>, <a href="/search/cs?searchtype=author&query=Tariq%2C+S">Shahroz Tariq</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.10399v1-abstract-short" style="display: inline;"> Deepfake detection is critical in mitigating the societal threats posed by manipulated videos. While various algorithms have been developed for this purpose, challenges arise when detectors operate externally, such as on smartphones, when users take a photo of deepfake images and upload on the Internet. One significant challenge in such scenarios is the presence of Moir茅 patterns, which degrade im… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10399v1-abstract-full').style.display = 'inline'; document.getElementById('2407.10399v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.10399v1-abstract-full" style="display: none;"> Deepfake detection is critical in mitigating the societal threats posed by manipulated videos. While various algorithms have been developed for this purpose, challenges arise when detectors operate externally, such as on smartphones, when users take a photo of deepfake images and upload on the Internet. One significant challenge in such scenarios is the presence of Moir茅 patterns, which degrade image quality and confound conventional classification algorithms, including deep neural networks (DNNs). The impact of Moir茅 patterns remains largely unexplored for deepfake detectors. In this study, we investigate how camera-captured deepfake videos from digital screens affect detector performance. We conducted experiments using two prominent datasets, CelebDF and FF++, comparing the performance of four state-of-the-art detectors on camera-captured deepfake videos with introduced Moir茅 patterns. Our findings reveal a significant decline in detector accuracy, with none achieving above 68% on average. This underscores the critical need to address Moir茅 pattern challenges in real-world deepfake detection scenarios. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10399v1-abstract-full').style.display = 'none'; document.getElementById('2407.10399v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 14 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">7 page, 4 figures, 1 table, Accepted for publication in IEEE International Conference on Image Processing (ICIP 2024)</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.10277">arXiv:2407.10277</a> <span> [<a href="https://arxiv.org/pdf/2407.10277">pdf</a>, <a href="https://arxiv.org/format/2407.10277">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Disrupting Diffusion-based Inpainters with Semantic Digression </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Son%2C+G">Geonho Son</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Juhun Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.10277v1-abstract-short" style="display: inline;"> The fabrication of visual misinformation on the web and social media has increased exponentially with the advent of foundational text-to-image diffusion models. Namely, Stable Diffusion inpainters allow the synthesis of maliciously inpainted images of personal and private figures, and copyrighted contents, also known as deepfakes. To combat such generations, a disruption framework, namely Photogua… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10277v1-abstract-full').style.display = 'inline'; document.getElementById('2407.10277v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.10277v1-abstract-full" style="display: none;"> The fabrication of visual misinformation on the web and social media has increased exponentially with the advent of foundational text-to-image diffusion models. Namely, Stable Diffusion inpainters allow the synthesis of maliciously inpainted images of personal and private figures, and copyrighted contents, also known as deepfakes. To combat such generations, a disruption framework, namely Photoguard, has been proposed, where it adds adversarial noise to the context image to disrupt their inpainting synthesis. While their framework suggested a diffusion-friendly approach, the disruption is not sufficiently strong and it requires a significant amount of GPU and time to immunize the context image. In our work, we re-examine both the minimal and favorable conditions for a successful inpainting disruption, proposing DDD, a "Digression guided Diffusion Disruption" framework. First, we identify the most adversarially vulnerable diffusion timestep range with respect to the hidden space. Within this scope of noised manifold, we pose the problem as a semantic digression optimization. We maximize the distance between the inpainting instance's hidden states and a semantic-aware hidden state centroid, calibrated both by Monte Carlo sampling of hidden states and a discretely projected optimization in the token space. Effectively, our approach achieves stronger disruption and a higher success rate than Photoguard while lowering the GPU memory requirement, and speeding the optimization up to three times faster. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.10277v1-abstract-full').style.display = 'none'; document.getElementById('2407.10277v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 14 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">16 pages, 13 figures, IJCAI 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.09303">arXiv:2407.09303</a> <span> [<a href="https://arxiv.org/pdf/2407.09303">pdf</a>, <a href="https://arxiv.org/format/2407.09303">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> ProDepth: Boosting Self-Supervised Multi-Frame Monocular Depth with Probabilistic Fusion </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sungmin Woo</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+W">Wonjoon Lee</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+W+J">Woo Jin Kim</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+D">Dogyoon Lee</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sangyoun Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.09303v1-abstract-short" style="display: inline;"> Self-supervised multi-frame monocular depth estimation relies on the geometric consistency between successive frames under the assumption of a static scene. However, the presence of moving objects in dynamic scenes introduces inevitable inconsistencies, causing misaligned multi-frame feature matching and misleading self-supervision during training. In this paper, we propose a novel framework calle… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.09303v1-abstract-full').style.display = 'inline'; document.getElementById('2407.09303v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.09303v1-abstract-full" style="display: none;"> Self-supervised multi-frame monocular depth estimation relies on the geometric consistency between successive frames under the assumption of a static scene. However, the presence of moving objects in dynamic scenes introduces inevitable inconsistencies, causing misaligned multi-frame feature matching and misleading self-supervision during training. In this paper, we propose a novel framework called ProDepth, which effectively addresses the mismatch problem caused by dynamic objects using a probabilistic approach. We initially deduce the uncertainty associated with static scene assumption by adopting an auxiliary decoder. This decoder analyzes inconsistencies embedded in the cost volume, inferring the probability of areas being dynamic. We then directly rectify the erroneous cost volume for dynamic areas through a Probabilistic Cost Volume Modulation (PCVM) module. Specifically, we derive probability distributions of depth candidates from both single-frame and multi-frame cues, modulating the cost volume by adaptively fusing those distributions based on the inferred uncertainty. Additionally, we present a self-supervision loss reweighting strategy that not only masks out incorrect supervision with high uncertainty but also mitigates the risks in remaining possible dynamic areas in accordance with the probability. Our proposed method excels over state-of-the-art approaches in all metrics on both Cityscapes and KITTI datasets, and demonstrates superior generalization ability on the Waymo Open dataset. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.09303v1-abstract-full').style.display = 'none'; document.getElementById('2407.09303v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 12 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by ECCV 2024. Project Page: https://sungmin-woo.github.io/prodepth/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.01073">arXiv:2407.01073</a> <span> [<a href="https://arxiv.org/pdf/2407.01073">pdf</a>, <a href="https://arxiv.org/format/2407.01073">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> No More Potentially Dynamic Objects: Static Point Cloud Map Generation based on 3D Object Detection and Ground Projection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Soojin Woo</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+D">Donghwi Jung</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+S">Seong-Woo Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.01073v1-abstract-short" style="display: inline;"> In this paper, we propose an algorithm to generate a static point cloud map based on LiDAR point cloud data. Our proposed pipeline detects dynamic objects using 3D object detectors and projects points of dynamic objects onto the ground. Typically, point cloud data acquired in real-time serves as a snapshot of the surrounding areas containing both static objects and dynamic objects. The static obje… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.01073v1-abstract-full').style.display = 'inline'; document.getElementById('2407.01073v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.01073v1-abstract-full" style="display: none;"> In this paper, we propose an algorithm to generate a static point cloud map based on LiDAR point cloud data. Our proposed pipeline detects dynamic objects using 3D object detectors and projects points of dynamic objects onto the ground. Typically, point cloud data acquired in real-time serves as a snapshot of the surrounding areas containing both static objects and dynamic objects. The static objects include buildings and trees, otherwise, the dynamic objects contain objects such as parked cars that change their position over time. Removing dynamic objects from the point cloud map is crucial as they can degrade the quality and localization accuracy of the map. To address this issue, in this paper, we propose an algorithm that creates a map only consisting of static objects. We apply a 3D object detection algorithm to the point cloud data which are obtained from LiDAR to implement our pipeline. We then stack the points to create the map after performing ground segmentation and projection. As a result, not only we can eliminate currently dynamic objects at the time of map generation but also potentially dynamic objects such as parked vehicles. We validate the performance of our method using two kinds of datasets collected on real roads: KITTI and our dataset. The result demonstrates the capability of our proposal to create an accurate static map excluding dynamic objects from input point clouds. Also, we verified the improved performance of localization using a generated map based on our method. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.01073v1-abstract-full').style.display = 'none'; document.getElementById('2407.01073v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 1 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.16860">arXiv:2406.16860</a> <span> [<a href="https://arxiv.org/pdf/2406.16860">pdf</a>, <a href="https://arxiv.org/format/2406.16860">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Tong%2C+S">Shengbang Tong</a>, <a href="/search/cs?searchtype=author&query=Brown%2C+E">Ellis Brown</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+P">Penghao Wu</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sanghyun Woo</a>, <a href="/search/cs?searchtype=author&query=Middepogu%2C+M">Manoj Middepogu</a>, <a href="/search/cs?searchtype=author&query=Akula%2C+S+C">Sai Charitha Akula</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+J">Jihan Yang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+S">Shusheng Yang</a>, <a href="/search/cs?searchtype=author&query=Iyer%2C+A">Adithya Iyer</a>, <a href="/search/cs?searchtype=author&query=Pan%2C+X">Xichen Pan</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+A">Austin Wang</a>, <a href="/search/cs?searchtype=author&query=Fergus%2C+R">Rob Fergus</a>, <a href="/search/cs?searchtype=author&query=LeCun%2C+Y">Yann LeCun</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+S">Saining Xie</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2406.16860v1-abstract-short" style="display: inline;"> We introduce Cambrian-1, a family of multimodal LLMs (MLLMs) designed with a vision-centric approach. While stronger language models can enhance multimodal capabilities, the design choices for vision components are often insufficiently explored and disconnected from visual representation learning research. This gap hinders accurate sensory grounding in real-world scenarios. Our study uses LLMs and… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16860v1-abstract-full').style.display = 'inline'; document.getElementById('2406.16860v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.16860v1-abstract-full" style="display: none;"> We introduce Cambrian-1, a family of multimodal LLMs (MLLMs) designed with a vision-centric approach. While stronger language models can enhance multimodal capabilities, the design choices for vision components are often insufficiently explored and disconnected from visual representation learning research. This gap hinders accurate sensory grounding in real-world scenarios. Our study uses LLMs and visual instruction tuning as an interface to evaluate various visual representations, offering new insights into different models and architectures -- self-supervised, strongly supervised, or combinations thereof -- based on experiments with over 20 vision encoders. We critically examine existing MLLM benchmarks, addressing the difficulties involved in consolidating and interpreting results from various tasks, and introduce a new vision-centric benchmark, CV-Bench. To further improve visual grounding, we propose the Spatial Vision Aggregator (SVA), a dynamic and spatially-aware connector that integrates high-resolution vision features with LLMs while reducing the number of tokens. Additionally, we discuss the curation of high-quality visual instruction-tuning data from publicly available sources, emphasizing the importance of data source balancing and distribution ratio. Collectively, Cambrian-1 not only achieves state-of-the-art performance but also serves as a comprehensive, open cookbook for instruction-tuned MLLMs. We provide model weights, code, supporting tools, datasets, and detailed instruction-tuning and evaluation recipes. We hope our release will inspire and accelerate advancements in multimodal systems and visual representation learning. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16860v1-abstract-full').style.display = 'none'; document.getElementById('2406.16860v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 24 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Website at https://cambrian-mllm.github.io</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.18012">arXiv:2405.18012</a> <span> [<a href="https://arxiv.org/pdf/2405.18012">pdf</a>, <a href="https://arxiv.org/format/2405.18012">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Image and Video Processing">eess.IV</span> </div> </div> <p class="title is-5 mathjax"> Flow-Assisted Motion Learning Network for Weakly-Supervised Group Activity Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Nugroho%2C+M+A">Muhammad Adi Nugroho</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sumin Lee</a>, <a href="/search/cs?searchtype=author&query=Park%2C+J">Jinyoung Park</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yooseung Wang</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+D">Donguk Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.18012v1-abstract-short" style="display: inline;"> Weakly-Supervised Group Activity Recognition (WSGAR) aims to understand the activity performed together by a group of individuals with the video-level label and without actor-level labels. We propose Flow-Assisted Motion Learning Network (Flaming-Net) for WSGAR, which consists of the motion-aware actor encoder to extract actor features and the two-pathways relation module to infer the interaction… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.18012v1-abstract-full').style.display = 'inline'; document.getElementById('2405.18012v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.18012v1-abstract-full" style="display: none;"> Weakly-Supervised Group Activity Recognition (WSGAR) aims to understand the activity performed together by a group of individuals with the video-level label and without actor-level labels. We propose Flow-Assisted Motion Learning Network (Flaming-Net) for WSGAR, which consists of the motion-aware actor encoder to extract actor features and the two-pathways relation module to infer the interaction among actors and their activity. Flaming-Net leverages an additional optical flow modality in the training stage to enhance its motion awareness when finding locally active actors. The first pathway of the relation module, the actor-centric path, initially captures the temporal dynamics of individual actors and then constructs inter-actor relationships. In parallel, the group-centric path starts by building spatial connections between actors within the same timeframe and then captures simultaneous spatio-temporal dynamics among them. We demonstrate that Flaming-Net achieves new state-of-the-art WSGAR results on two benchmarks, including a 2.8%p higher MPCA score on the NBA dataset. Importantly, we use the optical flow modality only for training and not for inference. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.18012v1-abstract-full').style.display = 'none'; document.getElementById('2405.18012v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 28 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.17928">arXiv:2405.17928</a> <span> [<a href="https://arxiv.org/pdf/2405.17928">pdf</a>, <a href="https://arxiv.org/format/2405.17928">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Relational Self-supervised Distillation with Compact Descriptors for Image Copy Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+J">Juntae Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sungwon Woo</a>, <a href="/search/cs?searchtype=author&query=Nang%2C+J">Jongho Nang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.17928v5-abstract-short" style="display: inline;"> Image copy detection is the task of detecting edited copies of any image within a reference database. While previous approaches have shown remarkable progress, the large size of their networks and descriptors remains a disadvantage, complicating their practical application. In this paper, we propose a novel method that achieves competitive performance by using a lightweight network and compact des… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17928v5-abstract-full').style.display = 'inline'; document.getElementById('2405.17928v5-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.17928v5-abstract-full" style="display: none;"> Image copy detection is the task of detecting edited copies of any image within a reference database. While previous approaches have shown remarkable progress, the large size of their networks and descriptors remains a disadvantage, complicating their practical application. In this paper, we propose a novel method that achieves competitive performance by using a lightweight network and compact descriptors. By utilizing relational self-supervised distillation to transfer knowledge from a large network to a small network, we enable the training of lightweight networks with smaller descriptor sizes. We introduce relational self-supervised distillation for flexible representation in a smaller feature space and apply contrastive learning with a hard negative loss to prevent dimensional collapse. For the DISC2021 benchmark, ResNet-50 and EfficientNet-B0 are used as the teacher and student models, respectively, with micro average precision improving by 5.0\%/4.9\%/5.9\% for 64/128/256 descriptor sizes compared to the baseline method. The code is available at \href{https://github.com/juntae9926/RDCD}{https://github.com/juntae9926/RDCD}. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17928v5-abstract-full').style.display = 'none'; document.getElementById('2405.17928v5-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 9 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 28 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">WACV 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.17825">arXiv:2405.17825</a> <span> [<a href="https://arxiv.org/pdf/2405.17825">pdf</a>, <a href="https://arxiv.org/format/2405.17825">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Diffusion Model Patching via Mixture-of-Prompts </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Ham%2C+S">Seokil Ham</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jin-Young Kim</a>, <a href="/search/cs?searchtype=author&query=Go%2C+H">Hyojun Go</a>, <a href="/search/cs?searchtype=author&query=Park%2C+B">Byeongjun Park</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.17825v2-abstract-short" style="display: inline;"> We present Diffusion Model Patching (DMP), a simple method to boost the performance of pre-trained diffusion models that have already reached convergence, with a negligible increase in parameters. DMP inserts a small, learnable set of prompts into the model's input space while keeping the original model frozen. The effectiveness of DMP is not merely due to the addition of parameters but stems from… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17825v2-abstract-full').style.display = 'inline'; document.getElementById('2405.17825v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.17825v2-abstract-full" style="display: none;"> We present Diffusion Model Patching (DMP), a simple method to boost the performance of pre-trained diffusion models that have already reached convergence, with a negligible increase in parameters. DMP inserts a small, learnable set of prompts into the model's input space while keeping the original model frozen. The effectiveness of DMP is not merely due to the addition of parameters but stems from its dynamic gating mechanism, which selects and combines a subset of learnable prompts at every step of the generative process (e.g., reverse denoising steps). This strategy, which we term "mixture-of-prompts", enables the model to draw on the distinct expertise of each prompt, essentially "patching" the model's functionality at every step with minimal yet specialized parameters. Uniquely, DMP enhances the model by further training on the same dataset on which it was originally trained, even in a scenario where significant improvements are typically not expected due to model convergence. Experiments show that DMP significantly enhances the converged FID of DiT-L/2 on FFHQ 256x256 by 10.38%, achieved with only a 1.43% parameter increase and 50K additional training iterations. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17825v2-abstract-full').style.display = 'none'; document.getElementById('2405.17825v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 30 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 28 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://sangminwoo.github.io/DMP/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.17821">arXiv:2405.17821</a> <span> [<a href="https://arxiv.org/pdf/2405.17821">pdf</a>, <a href="https://arxiv.org/format/2405.17821">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> RITUAL: Random Image Transformations as a Universal Anti-hallucination Lever in LVLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Jang%2C+J">Jaehyuk Jang</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+D">Donguk Kim</a>, <a href="/search/cs?searchtype=author&query=Choi%2C+Y">Yubin Choi</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.17821v1-abstract-short" style="display: inline;"> Recent advancements in Large Vision Language Models (LVLMs) have revolutionized how machines understand and generate textual responses based on visual inputs. Despite their impressive capabilities, they often produce "hallucinatory" outputs that do not accurately reflect the visual information, posing challenges in reliability and trustworthiness. Current methods such as contrastive decoding have… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17821v1-abstract-full').style.display = 'inline'; document.getElementById('2405.17821v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.17821v1-abstract-full" style="display: none;"> Recent advancements in Large Vision Language Models (LVLMs) have revolutionized how machines understand and generate textual responses based on visual inputs. Despite their impressive capabilities, they often produce "hallucinatory" outputs that do not accurately reflect the visual information, posing challenges in reliability and trustworthiness. Current methods such as contrastive decoding have made strides in addressing these issues by contrasting the original probability distribution of generated tokens with distorted counterparts; yet, generating visually-faithful outputs remains a challenge. In this work, we shift our focus to the opposite: What could serve as a complementary enhancement to the original probability distribution? We propose a simple, training-free method termed RITUAL to enhance robustness against hallucinations in LVLMs. Our approach employs random image transformations as complements to the original probability distribution, aiming to mitigate the likelihood of hallucinatory visual explanations by enriching the model's exposure to varied visual scenarios. Our empirical results show that while the isolated use of transformed images initially degrades performance, strategic implementation of these transformations can indeed serve as effective complements. Notably, our method is compatible with current contrastive decoding methods and does not require external models or costly self-feedback mechanisms, making it a practical addition. In experiments, RITUAL significantly outperforms existing contrastive decoding methods across several object hallucination benchmarks, including POPE, CHAIR, and MME. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17821v1-abstract-full').style.display = 'none'; document.getElementById('2405.17821v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 28 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://sangminwoo.github.io/RITUAL/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.17820">arXiv:2405.17820</a> <span> [<a href="https://arxiv.org/pdf/2405.17820">pdf</a>, <a href="https://arxiv.org/format/2405.17820">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+D">Donguk Kim</a>, <a href="/search/cs?searchtype=author&query=Jang%2C+J">Jaehyuk Jang</a>, <a href="/search/cs?searchtype=author&query=Choi%2C+Y">Yubin Choi</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.17820v1-abstract-short" style="display: inline;"> This study addresses the issue observed in Large Vision Language Models (LVLMs), where excessive attention on a few image tokens, referred to as blind tokens, leads to hallucinatory responses in tasks requiring fine-grained understanding of visual objects. We found that tokens receiving lower attention weights often hold essential information for identifying nuanced object details -- ranging from… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17820v1-abstract-full').style.display = 'inline'; document.getElementById('2405.17820v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.17820v1-abstract-full" style="display: none;"> This study addresses the issue observed in Large Vision Language Models (LVLMs), where excessive attention on a few image tokens, referred to as blind tokens, leads to hallucinatory responses in tasks requiring fine-grained understanding of visual objects. We found that tokens receiving lower attention weights often hold essential information for identifying nuanced object details -- ranging from merely recognizing object existence to identifying their attributes (color, position, etc.) and understanding their relationships. To counteract the over-emphasis on blind tokens and to accurately respond to user queries, we introduce a technique called Attentional Vision Calibration (AVC). During the decoding phase, AVC identifies blind tokens by analyzing the image-related attention distribution. It then dynamically adjusts the logits for the next token prediction by contrasting the logits conditioned on the original visual tokens with those conditioned on the blind tokens. This effectively lowers the dependency on blind tokens and promotes a more balanced consideration of all tokens. We validate AVC on benchmarks such as POPE, MME, and AMBER, where it consistently outperforms existing decoding techniques in mitigating object hallucinations in LVLMs. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.17820v1-abstract-full').style.display = 'none'; document.getElementById('2405.17820v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 28 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://sangminwoo.github.io/AvisC/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.01934">arXiv:2405.01934</a> <span> [<a href="https://arxiv.org/pdf/2405.01934">pdf</a>, <a href="https://arxiv.org/format/2405.01934">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Impact of Architectural Modifications on Deep Learning Adversarial Robustness </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Juraev%2C+F">Firuz Juraev</a>, <a href="/search/cs?searchtype=author&query=Abuhamad%2C+M">Mohammed Abuhamad</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a>, <a href="/search/cs?searchtype=author&query=Thiruvathukal%2C+G+K">George K Thiruvathukal</a>, <a href="/search/cs?searchtype=author&query=Abuhmed%2C+T">Tamer Abuhmed</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.01934v1-abstract-short" style="display: inline;"> Rapid advancements of deep learning are accelerating adoption in a wide variety of applications, including safety-critical applications such as self-driving vehicles, drones, robots, and surveillance systems. These advancements include applying variations of sophisticated techniques that improve the performance of models. However, such models are not immune to adversarial manipulations, which can… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.01934v1-abstract-full').style.display = 'inline'; document.getElementById('2405.01934v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.01934v1-abstract-full" style="display: none;"> Rapid advancements of deep learning are accelerating adoption in a wide variety of applications, including safety-critical applications such as self-driving vehicles, drones, robots, and surveillance systems. These advancements include applying variations of sophisticated techniques that improve the performance of models. However, such models are not immune to adversarial manipulations, which can cause the system to misbehave and remain unnoticed by experts. The frequency of modifications to existing deep learning models necessitates thorough analysis to determine the impact on models' robustness. In this work, we present an experimental evaluation of the effects of model modifications on deep learning model robustness using adversarial attacks. Our methodology involves examining the robustness of variations of models against various adversarial attacks. By conducting our experiments, we aim to shed light on the critical issue of maintaining the reliability and safety of deep learning models in safety- and security-critical applications. Our results indicate the pressing demand for an in-depth assessment of the effects of model changes on the robustness of models. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.01934v1-abstract-full').style.display = 'none'; document.getElementById('2405.01934v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 3 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2404.14617">arXiv:2404.14617</a> <span> [<a href="https://arxiv.org/pdf/2404.14617">pdf</a>, <a href="https://arxiv.org/format/2404.14617">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Hardware Architecture">cs.AR</span> </div> </div> <p class="title is-5 mathjax"> TDRAM: Tag-enhanced DRAM for Efficient Caching </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Babaie%2C+M">Maryam Babaie</a>, <a href="/search/cs?searchtype=author&query=Akram%2C+A">Ayaz Akram</a>, <a href="/search/cs?searchtype=author&query=Elsasser%2C+W">Wendy Elsasser</a>, <a href="/search/cs?searchtype=author&query=Haukness%2C+B">Brent Haukness</a>, <a href="/search/cs?searchtype=author&query=Miller%2C+M">Michael Miller</a>, <a href="/search/cs?searchtype=author&query=Song%2C+T">Taeksang Song</a>, <a href="/search/cs?searchtype=author&query=Vogelsang%2C+T">Thomas Vogelsang</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Steven Woo</a>, <a href="/search/cs?searchtype=author&query=Lowe-Power%2C+J">Jason Lowe-Power</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2404.14617v1-abstract-short" style="display: inline;"> As SRAM-based caches are hitting a scaling wall, manufacturers are integrating DRAM-based caches into system designs to continue increasing cache sizes. While DRAM caches can improve the performance of memory systems, existing DRAM cache designs suffer from high miss penalties, wasted data movement, and interference between misses and demand requests. In this paper, we propose TDRAM, a novel DRAM… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.14617v1-abstract-full').style.display = 'inline'; document.getElementById('2404.14617v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2404.14617v1-abstract-full" style="display: none;"> As SRAM-based caches are hitting a scaling wall, manufacturers are integrating DRAM-based caches into system designs to continue increasing cache sizes. While DRAM caches can improve the performance of memory systems, existing DRAM cache designs suffer from high miss penalties, wasted data movement, and interference between misses and demand requests. In this paper, we propose TDRAM, a novel DRAM microarchitecture tailored for caching. TDRAM enhances HBM3 by adding a set of small low-latency mats to store tags and metadata on the same die as the data mats. These mats enable fast parallel tag and data access, on-DRAM-die tag comparison, and conditional data response based on comparison result (reducing wasted data transfers) akin to SRAM caches mechanism. TDRAM further optimizes the hit and miss latencies by performing opportunistic early tag probing. Moreover, TDRAM introduces a flush buffer to store conflicting dirty data on write misses, eliminating turnaround delays on data bus. We evaluate TDRAM using a full-system simulator and a set of HPC workloads with large memory footprints showing TDRAM provides at least 2.6$\times$ faster tag check, 1.2$\times$ speedup, and 21% less energy consumption, compared to the state-of-the-art commercial and research designs. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.14617v1-abstract-full').style.display = 'none'; document.getElementById('2404.14617v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 22 April, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> April 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.20225">arXiv:2403.20225</a> <span> [<a href="https://arxiv.org/pdf/2403.20225">pdf</a>, <a href="https://arxiv.org/format/2403.20225">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> MTMMC: A Large-Scale Real-World Multi-Modal Camera Tracking Benchmark </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sanghyun Woo</a>, <a href="/search/cs?searchtype=author&query=Park%2C+K">Kwanyong Park</a>, <a href="/search/cs?searchtype=author&query=Shin%2C+I">Inkyu Shin</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+M">Myungchul Kim</a>, <a href="/search/cs?searchtype=author&query=Kweon%2C+I+S">In So Kweon</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.20225v1-abstract-short" style="display: inline;"> Multi-target multi-camera tracking is a crucial task that involves identifying and tracking individuals over time using video streams from multiple cameras. This task has practical applications in various fields, such as visual surveillance, crowd behavior analysis, and anomaly detection. However, due to the difficulty and cost of collecting and labeling data, existing datasets for this task are e… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.20225v1-abstract-full').style.display = 'inline'; document.getElementById('2403.20225v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.20225v1-abstract-full" style="display: none;"> Multi-target multi-camera tracking is a crucial task that involves identifying and tracking individuals over time using video streams from multiple cameras. This task has practical applications in various fields, such as visual surveillance, crowd behavior analysis, and anomaly detection. However, due to the difficulty and cost of collecting and labeling data, existing datasets for this task are either synthetically generated or artificially constructed within a controlled camera network setting, which limits their ability to model real-world dynamics and generalize to diverse camera configurations. To address this issue, we present MTMMC, a real-world, large-scale dataset that includes long video sequences captured by 16 multi-modal cameras in two different environments - campus and factory - across various time, weather, and season conditions. This dataset provides a challenging test-bed for studying multi-camera tracking under diverse real-world complexities and includes an additional input modality of spatially aligned and temporally synchronized RGB and thermal cameras, which enhances the accuracy of multi-camera tracking. MTMMC is a super-set of existing datasets, benefiting independent fields such as person detection, re-identification, and multiple object tracking. We provide baselines and new learning setups on this dataset and set the reference scores for future studies. The datasets, models, and test server will be made publicly available. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.20225v1-abstract-full').style.display = 'none'; document.getElementById('2403.20225v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 29 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted on CVPR 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.14113">arXiv:2403.14113</a> <span> [<a href="https://arxiv.org/pdf/2403.14113">pdf</a>, <a href="https://arxiv.org/format/2403.14113">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Spatio-Temporal Proximity-Aware Dual-Path Model for Panoramic Activity Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sumin Lee</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yooseung Wang</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.14113v1-abstract-short" style="display: inline;"> Panoramic Activity Recognition (PAR) seeks to identify diverse human activities across different scales, from individual actions to social group and global activities in crowded panoramic scenes. PAR presents two major challenges: 1) recognizing the nuanced interactions among numerous individuals and 2) understanding multi-granular human activities. To address these, we propose Social Proximity-aw… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.14113v1-abstract-full').style.display = 'inline'; document.getElementById('2403.14113v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.14113v1-abstract-full" style="display: none;"> Panoramic Activity Recognition (PAR) seeks to identify diverse human activities across different scales, from individual actions to social group and global activities in crowded panoramic scenes. PAR presents two major challenges: 1) recognizing the nuanced interactions among numerous individuals and 2) understanding multi-granular human activities. To address these, we propose Social Proximity-aware Dual-Path Network (SPDP-Net) based on two key design principles. First, while previous works often focus on spatial distance among individuals within an image, we argue to consider the spatio-temporal proximity. It is crucial for individual relation encoding to correctly understand social dynamics. Secondly, deviating from existing hierarchical approaches (individual-to-social-to-global activity), we introduce a dual-path architecture for multi-granular activity recognition. This architecture comprises individual-to-global and individual-to-social paths, mutually reinforcing each other's task with global-local context through multiple layers. Through extensive experiments, we validate the effectiveness of the spatio-temporal proximity among individuals and the dual-path architecture in PAR. Furthermore, SPDP-Net achieves new state-of-the-art performance with 46.5\% of overall F1 score on JRDB-PAR dataset. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.14113v1-abstract-full').style.display = 'none'; document.getElementById('2403.14113v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.11582">arXiv:2403.11582</a> <span> [<a href="https://arxiv.org/pdf/2403.11582">pdf</a>, <a href="https://arxiv.org/format/2403.11582">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> OurDB: Ouroboric Domain Bridging for Multi-Target Domain Adaptive Semantic Segmentation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Seungbeom Woo</a>, <a href="/search/cs?searchtype=author&query=Baek%2C+G">Geonwoo Baek</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+T">Taehoon Kim</a>, <a href="/search/cs?searchtype=author&query=Na%2C+J">Jaemin Na</a>, <a href="/search/cs?searchtype=author&query=Hwang%2C+J">Joong-won Hwang</a>, <a href="/search/cs?searchtype=author&query=Hwang%2C+W">Wonjun Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.11582v1-abstract-short" style="display: inline;"> Multi-target domain adaptation (MTDA) for semantic segmentation poses a significant challenge, as it involves multiple target domains with varying distributions. The goal of MTDA is to minimize the domain discrepancies among a single source and multi-target domains, aiming to train a single model that excels across all target domains. Previous MTDA approaches typically employ multiple teacher arch… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.11582v1-abstract-full').style.display = 'inline'; document.getElementById('2403.11582v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.11582v1-abstract-full" style="display: none;"> Multi-target domain adaptation (MTDA) for semantic segmentation poses a significant challenge, as it involves multiple target domains with varying distributions. The goal of MTDA is to minimize the domain discrepancies among a single source and multi-target domains, aiming to train a single model that excels across all target domains. Previous MTDA approaches typically employ multiple teacher architectures, where each teacher specializes in one target domain to simplify the task. However, these architectures hinder the student model from fully assimilating comprehensive knowledge from all target-specific teachers and escalate training costs with increasing target domains. In this paper, we propose an ouroboric domain bridging (OurDB) framework, offering an efficient solution to the MTDA problem using a single teacher architecture. This framework dynamically cycles through multiple target domains, aligning each domain individually to restrain the biased alignment problem, and utilizes Fisher information to minimize the forgetting of knowledge from previous target domains. We also propose a context-guided class-wise mixup (CGMix) that leverages contextual information tailored to diverse target contexts in MTDA. Experimental evaluations conducted on four urban driving datasets (i.e., GTA5, Cityscapes, IDD, and Mapillary) demonstrate the superiority of our method over existing state-of-the-art approaches. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.11582v1-abstract-full').style.display = 'none'; document.getElementById('2403.11582v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 18 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.09176">arXiv:2403.09176</a> <span> [<a href="https://arxiv.org/pdf/2403.09176">pdf</a>, <a href="https://arxiv.org/format/2403.09176">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Switch Diffusion Transformer: Synergizing Denoising Tasks with Sparse Mixture-of-Experts </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Park%2C+B">Byeongjun Park</a>, <a href="/search/cs?searchtype=author&query=Go%2C+H">Hyojun Go</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jin-Young Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Ham%2C+S">Seokil Ham</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.09176v2-abstract-short" style="display: inline;"> Diffusion models have achieved remarkable success across a range of generative tasks. Recent efforts to enhance diffusion model architectures have reimagined them as a form of multi-task learning, where each task corresponds to a denoising task at a specific noise level. While these efforts have focused on parameter isolation and task routing, they fall short of capturing detailed inter-task relat… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.09176v2-abstract-full').style.display = 'inline'; document.getElementById('2403.09176v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.09176v2-abstract-full" style="display: none;"> Diffusion models have achieved remarkable success across a range of generative tasks. Recent efforts to enhance diffusion model architectures have reimagined them as a form of multi-task learning, where each task corresponds to a denoising task at a specific noise level. While these efforts have focused on parameter isolation and task routing, they fall short of capturing detailed inter-task relationships and risk losing semantic information, respectively. In response, we introduce Switch Diffusion Transformer (Switch-DiT), which establishes inter-task relationships between conflicting tasks without compromising semantic information. To achieve this, we employ a sparse mixture-of-experts within each transformer block to utilize semantic information and facilitate handling conflicts in tasks through parameter isolation. Additionally, we propose a diffusion prior loss, encouraging similar tasks to share their denoising paths while isolating conflicting ones. Through these, each transformer block contains a shared expert across all tasks, where the common and task-specific denoising paths enable the diffusion model to construct its beneficial way of synergizing denoising tasks. Extensive experiments validate the effectiveness of our approach in improving both image quality and convergence rate, and further analysis demonstrates that Switch-DiT constructs tailored denoising paths across various generation scenarios. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.09176v2-abstract-full').style.display = 'none'; document.getElementById('2403.09176v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 10 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 14 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project Page: https://byeongjun-park.github.io/Switch-DiT/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.04981">arXiv:2403.04981</a> <span> [<a href="https://arxiv.org/pdf/2403.04981">pdf</a>, <a href="https://arxiv.org/format/2403.04981">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Emerging Technologies">cs.ET</span> </div> </div> <p class="title is-5 mathjax"> Paving the Way for Pass Disturb Free Vertical NAND Storage via A Dedicated and String-Compatible Pass Gate </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhao%2C+Z">Zijian Zhao</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sola Woo</a>, <a href="/search/cs?searchtype=author&query=Aabrar%2C+K+A">Khandker Akif Aabrar</a>, <a href="/search/cs?searchtype=author&query=Kirtania%2C+S+G">Sharadindu Gopal Kirtania</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Z">Zhouhang Jiang</a>, <a href="/search/cs?searchtype=author&query=Deng%2C+S">Shan Deng</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+Y">Yi Xiao</a>, <a href="/search/cs?searchtype=author&query=Mulaosmanovic%2C+H">Halid Mulaosmanovic</a>, <a href="/search/cs?searchtype=author&query=Duenkel%2C+S">Stefan Duenkel</a>, <a href="/search/cs?searchtype=author&query=Kleimaier%2C+D">Dominik Kleimaier</a>, <a href="/search/cs?searchtype=author&query=Soss%2C+S">Steven Soss</a>, <a href="/search/cs?searchtype=author&query=Beyer%2C+S">Sven Beyer</a>, <a href="/search/cs?searchtype=author&query=Joshi%2C+R">Rajiv Joshi</a>, <a href="/search/cs?searchtype=author&query=Meninger%2C+S">Scott Meninger</a>, <a href="/search/cs?searchtype=author&query=Mohamed%2C+M">Mohamed Mohamed</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+K">Kijoon Kim</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+J">Jongho Woo</a>, <a href="/search/cs?searchtype=author&query=Lim%2C+S">Suhwan Lim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+K">Kwangsoo Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+W">Wanki Kim</a>, <a href="/search/cs?searchtype=author&query=Ha%2C+D">Daewon Ha</a>, <a href="/search/cs?searchtype=author&query=Narayanan%2C+V">Vijaykrishnan Narayanan</a>, <a href="/search/cs?searchtype=author&query=Datta%2C+S">Suman Datta</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+S">Shimeng Yu</a>, <a href="/search/cs?searchtype=author&query=Ni%2C+K">Kai Ni</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.04981v1-abstract-short" style="display: inline;"> In this work, we propose a dual-port cell design to address the pass disturb in vertical NAND storage, which can pass signals through a dedicated and string-compatible pass gate. We demonstrate that: i) the pass disturb-free feature originates from weakening of the depolarization field by the pass bias at the high-${V}_{TH}$ (HVT) state and the screening of the applied field by channel at the low-… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.04981v1-abstract-full').style.display = 'inline'; document.getElementById('2403.04981v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.04981v1-abstract-full" style="display: none;"> In this work, we propose a dual-port cell design to address the pass disturb in vertical NAND storage, which can pass signals through a dedicated and string-compatible pass gate. We demonstrate that: i) the pass disturb-free feature originates from weakening of the depolarization field by the pass bias at the high-${V}_{TH}$ (HVT) state and the screening of the applied field by channel at the low-${V}_{TH}$ (LVT) state; ii) combined simulations and experimental demonstrations of dual-port design verify the disturb-free operation in a NAND string, overcoming a key challenge in single-port designs; iii) the proposed design can be incorporated in a highly scaled vertical NAND FeFET string and the pass gate can be incorporated into the existing 3D NAND with the negligible overhead of the pass gate interconnection through a global bottom pass gate contact in the substrate. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.04981v1-abstract-full').style.display = 'none'; document.getElementById('2403.04981v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 7 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">29 pages, 7 figures</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2402.18848">arXiv:2402.18848</a> <span> [<a href="https://arxiv.org/pdf/2402.18848">pdf</a>, <a href="https://arxiv.org/format/2402.18848">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> SwitchLight: Co-design of Physics-driven Architecture and Pre-training Framework for Human Portrait Relighting </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hoon Kim</a>, <a href="/search/cs?searchtype=author&query=Jang%2C+M">Minje Jang</a>, <a href="/search/cs?searchtype=author&query=Yoon%2C+W">Wonjun Yoon</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jisoo Lee</a>, <a href="/search/cs?searchtype=author&query=Na%2C+D">Donghyun Na</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sanghyun Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.18848v1-abstract-short" style="display: inline;"> We introduce a co-designed approach for human portrait relighting that combines a physics-guided architecture with a pre-training framework. Drawing on the Cook-Torrance reflectance model, we have meticulously configured the architecture design to precisely simulate light-surface interactions. Furthermore, to overcome the limitation of scarce high-quality lightstage data, we have developed a self-… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18848v1-abstract-full').style.display = 'inline'; document.getElementById('2402.18848v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.18848v1-abstract-full" style="display: none;"> We introduce a co-designed approach for human portrait relighting that combines a physics-guided architecture with a pre-training framework. Drawing on the Cook-Torrance reflectance model, we have meticulously configured the architecture design to precisely simulate light-surface interactions. Furthermore, to overcome the limitation of scarce high-quality lightstage data, we have developed a self-supervised pre-training strategy. This novel combination of accurate physical modeling and expanded training dataset establishes a new benchmark in relighting realism. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18848v1-abstract-full').style.display = 'none'; document.getElementById('2402.18848v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 28 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">CVPR2024. Live demos available at https://www.beeble.ai/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2402.18817">arXiv:2402.18817</a> <span> [<a href="https://arxiv.org/pdf/2402.18817">pdf</a>, <a href="https://arxiv.org/format/2402.18817">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Gradient Alignment for Cross-Domain Face Anti-Spoofing </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Le%2C+B+M">Binh M. Le</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.18817v2-abstract-short" style="display: inline;"> Recent advancements in domain generalization (DG) for face anti-spoofing (FAS) have garnered considerable attention. Traditional methods have focused on designing learning objectives and additional modules to isolate domain-specific features while retaining domain-invariant characteristics in their representations. However, such approaches often lack guarantees of consistent maintenance of domain-… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18817v2-abstract-full').style.display = 'inline'; document.getElementById('2402.18817v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.18817v2-abstract-full" style="display: none;"> Recent advancements in domain generalization (DG) for face anti-spoofing (FAS) have garnered considerable attention. Traditional methods have focused on designing learning objectives and additional modules to isolate domain-specific features while retaining domain-invariant characteristics in their representations. However, such approaches often lack guarantees of consistent maintenance of domain-invariant features or the complete removal of domain-specific features. Furthermore, most prior works of DG for FAS do not ensure convergence to a local flat minimum, which has been shown to be advantageous for DG. In this paper, we introduce GAC-FAS, a novel learning objective that encourages the model to converge towards an optimal flat minimum without necessitating additional learning modules. Unlike conventional sharpness-aware minimizers, GAC-FAS identifies ascending points for each domain and regulates the generalization gradient updates at these points to align coherently with empirical risk minimization (ERM) gradient updates. This unique approach specifically guides the model to be robust against domain shifts. We demonstrate the efficacy of GAC-FAS through rigorous testing on challenging cross-domain FAS datasets, where it establishes state-of-the-art performance. The code is available at https://github.com/leminhbinh0209/CVPR24-FAS. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18817v2-abstract-full').style.display = 'none'; document.getElementById('2402.18817v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 11 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 28 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2024 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2402.18293">arXiv:2402.18293</a> <span> [<a href="https://arxiv.org/pdf/2402.18293">pdf</a>, <a href="https://arxiv.org/format/2402.18293">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Continuous Memory Representation for Anomaly Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lee%2C+J+C">Joo Chan Lee</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+T">Taejune Kim</a>, <a href="/search/cs?searchtype=author&query=Park%2C+E">Eunbyung Park</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a>, <a href="/search/cs?searchtype=author&query=Ko%2C+J+H">Jong Hwan Ko</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.18293v3-abstract-short" style="display: inline;"> There have been significant advancements in anomaly detection in an unsupervised manner, where only normal images are available for training. Several recent methods aim to detect anomalies based on a memory, comparing or reconstructing the input with directly stored normal features (or trained features with normal images). However, such memory-based approaches operate on a discrete feature space i… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18293v3-abstract-full').style.display = 'inline'; document.getElementById('2402.18293v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.18293v3-abstract-full" style="display: none;"> There have been significant advancements in anomaly detection in an unsupervised manner, where only normal images are available for training. Several recent methods aim to detect anomalies based on a memory, comparing or reconstructing the input with directly stored normal features (or trained features with normal images). However, such memory-based approaches operate on a discrete feature space implemented by the nearest neighbor or attention mechanism, suffering from poor generalization or an identity shortcut issue outputting the same as input, respectively. Furthermore, the majority of existing methods are designed to detect single-class anomalies, resulting in unsatisfactory performance when presented with multiple classes of objects. To tackle all of the above challenges, we propose CRAD, a novel anomaly detection method for representing normal features within a "continuous" memory, enabled by transforming spatial features into coordinates and mapping them to continuous grids. Furthermore, we carefully design the grids tailored for anomaly detection, representing both local and global normal features and fusing them effectively. Our extensive experiments demonstrate that CRAD successfully generalizes the normal features and mitigates the identity shortcut, furthermore, CRAD effectively handles diverse classes in a single model thanks to the high-granularity continuous representation. In an evaluation using the MVTec AD dataset, CRAD significantly outperforms the previous state-of-the-art method by reducing 65.0% of the error for multi-class unified anomaly detection. The project page is available at https://tae-mo.github.io/crad/. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.18293v3-abstract-full').style.display = 'none'; document.getElementById('2402.18293v3-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 24 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 28 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://tae-mo.github.io/crad/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2402.17812">arXiv:2402.17812</a> <span> [<a href="https://arxiv.org/pdf/2402.17812">pdf</a>, <a href="https://arxiv.org/format/2402.17812">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> DropBP: Accelerating Fine-Tuning of Large Language Models by Dropping Backward Propagation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sunghyeon Woo</a>, <a href="/search/cs?searchtype=author&query=Park%2C+B">Baeseong Park</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+B">Byeongwook Kim</a>, <a href="/search/cs?searchtype=author&query=Jo%2C+M">Minjung Jo</a>, <a href="/search/cs?searchtype=author&query=Kwon%2C+S+J">Se Jung Kwon</a>, <a href="/search/cs?searchtype=author&query=Jeon%2C+D">Dongsuk Jeon</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+D">Dongsoo Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.17812v2-abstract-short" style="display: inline;"> Large language models (LLMs) have achieved significant success across various domains. However, training these LLMs typically involves substantial memory and computational costs during both forward and backward propagation. While parameter-efficient fine-tuning (PEFT) considerably reduces the training memory associated with parameters, it does not address the significant computational costs and ac… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.17812v2-abstract-full').style.display = 'inline'; document.getElementById('2402.17812v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.17812v2-abstract-full" style="display: none;"> Large language models (LLMs) have achieved significant success across various domains. However, training these LLMs typically involves substantial memory and computational costs during both forward and backward propagation. While parameter-efficient fine-tuning (PEFT) considerably reduces the training memory associated with parameters, it does not address the significant computational costs and activation memory. In this paper, we propose Dropping Backward Propagation (DropBP), a novel approach designed to reduce computational costs and activation memory while maintaining accuracy. DropBP randomly drops layers during backward propagation, which is essentially equivalent to training shallow submodules generated by undropped layers and residual connections. Additionally, DropBP calculates the sensitivity of each layer to assign an appropriate drop rate, thereby stabilizing the training process. DropBP is not only applicable to full fine-tuning but can also be orthogonally integrated with all types of PEFT by dropping layers during backward propagation. Specifically, DropBP can reduce training time by 44% with comparable accuracy to the baseline, accelerate convergence to the same perplexity by 1.5x, and enable training with a sequence length 6.2x larger on a single NVIDIA-A100 GPU. Furthermore, our DropBP enabled a throughput increase of 79% on a NVIDIA A100 GPU and 117% on an Intel Gaudi2 HPU. The code is available at https://github.com/WooSunghyeon/dropbp. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.17812v2-abstract-full').style.display = 'none'; document.getElementById('2402.17812v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 6 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 27 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.17690">arXiv:2401.17690</a> <span> [<a href="https://arxiv.org/pdf/2401.17690">pdf</a>, <a href="https://arxiv.org/format/2401.17690">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Audio and Speech Processing">eess.AS</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Sound">cs.SD</span> </div> </div> <p class="title is-5 mathjax"> EnCLAP: Combining Neural Audio Codec and Audio-Text Joint Embedding for Automated Audio Captioning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jaeyeon Kim</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+J">Jaeyoon Jung</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Jinjoo Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+H">Sang Hoon Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.17690v1-abstract-short" style="display: inline;"> We propose EnCLAP, a novel framework for automated audio captioning. EnCLAP employs two acoustic representation models, EnCodec and CLAP, along with a pretrained language model, BART. We also introduce a new training objective called masked codec modeling that improves acoustic awareness of the pretrained language model. Experimental results on AudioCaps and Clotho demonstrate that our model surpa… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.17690v1-abstract-full').style.display = 'inline'; document.getElementById('2401.17690v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.17690v1-abstract-full" style="display: none;"> We propose EnCLAP, a novel framework for automated audio captioning. EnCLAP employs two acoustic representation models, EnCodec and CLAP, along with a pretrained language model, BART. We also introduce a new training objective called masked codec modeling that improves acoustic awareness of the pretrained language model. Experimental results on AudioCaps and Clotho demonstrate that our model surpasses the performance of baseline models. Source code will be available at https://github.com/jaeyeonkim99/EnCLAP . An online demo is available at https://huggingface.co/spaces/enclap-team/enclap . <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.17690v1-abstract-full').style.display = 'none'; document.getElementById('2401.17690v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 31 January, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to ICASSP 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.16189">arXiv:2401.16189</a> <span> [<a href="https://arxiv.org/pdf/2401.16189">pdf</a>, <a href="https://arxiv.org/format/2401.16189">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> FIMP: Future Interaction Modeling for Multi-Agent Motion Prediction </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sungmin Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+M">Minjung Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+D">Donghyeong Kim</a>, <a href="/search/cs?searchtype=author&query=Jang%2C+S">Sungjun Jang</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sangyoun Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.16189v1-abstract-short" style="display: inline;"> Multi-agent motion prediction is a crucial concern in autonomous driving, yet it remains a challenge owing to the ambiguous intentions of dynamic agents and their intricate interactions. Existing studies have attempted to capture interactions between road entities by using the definite data in history timesteps, as future information is not available and involves high uncertainty. However, without… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.16189v1-abstract-full').style.display = 'inline'; document.getElementById('2401.16189v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.16189v1-abstract-full" style="display: none;"> Multi-agent motion prediction is a crucial concern in autonomous driving, yet it remains a challenge owing to the ambiguous intentions of dynamic agents and their intricate interactions. Existing studies have attempted to capture interactions between road entities by using the definite data in history timesteps, as future information is not available and involves high uncertainty. However, without sufficient guidance for capturing future states of interacting agents, they frequently produce unrealistic trajectory overlaps. In this work, we propose Future Interaction modeling for Motion Prediction (FIMP), which captures potential future interactions in an end-to-end manner. FIMP adopts a future decoder that implicitly extracts the potential future information in an intermediate feature-level, and identifies the interacting entity pairs through future affinity learning and top-k filtering strategy. Experiments show that our future interaction modeling improves the performance remarkably, leading to superior performance on the Argoverse motion forecasting benchmark. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.16189v1-abstract-full').style.display = 'none'; document.getElementById('2401.16189v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 29 January, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by ICRA 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.04364">arXiv:2401.04364</a> <span> [<a href="https://arxiv.org/pdf/2401.04364">pdf</a>, <a href="https://arxiv.org/format/2401.04364">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> SoK: Facial Deepfake Detectors </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Le%2C+B+M">Binh M. Le</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jiwon Kim</a>, <a href="/search/cs?searchtype=author&query=Tariq%2C+S">Shahroz Tariq</a>, <a href="/search/cs?searchtype=author&query=Moore%2C+K">Kristen Moore</a>, <a href="/search/cs?searchtype=author&query=Abuadbba%2C+A">Alsharif Abuadbba</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.04364v2-abstract-short" style="display: inline;"> Deepfakes have rapidly emerged as a profound and serious threat to society, primarily due to their ease of creation and dissemination. This situation has triggered an accelerated development of deepfake detection technologies. However, many existing detectors rely heavily on lab-generated datasets for validation, which may not effectively prepare them for novel, emerging, and real-world deepfake t… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.04364v2-abstract-full').style.display = 'inline'; document.getElementById('2401.04364v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.04364v2-abstract-full" style="display: none;"> Deepfakes have rapidly emerged as a profound and serious threat to society, primarily due to their ease of creation and dissemination. This situation has triggered an accelerated development of deepfake detection technologies. However, many existing detectors rely heavily on lab-generated datasets for validation, which may not effectively prepare them for novel, emerging, and real-world deepfake techniques. In this paper, we conduct an extensive and comprehensive review and analysis of the latest state-of-the-art deepfake detectors, evaluating them against several critical criteria. These criteria facilitate the categorization of these detectors into 4 high-level groups and 13 fine-grained sub-groups, all aligned with a unified standard conceptual framework. This classification and framework offer deep and practical insights into the factors that affect detector efficacy. We assess the generalizability of 16 leading detectors across various standard attack scenarios, including black-box, white-box, and gray-box settings. Our systematized analysis and experimentation lay the groundwork for a deeper understanding of deepfake detectors and their generalizability, paving the way for future research focused on creating detectors adept at countering various attack scenarios. Additionally, this work offers insights for developing more proactive defenses against deepfakes. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.04364v2-abstract-full').style.display = 'none'; document.getElementById('2401.04364v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 25 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 9 January, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">18 pages, 6 figures, 5 table, under peer-review</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.02113">arXiv:2401.02113</a> <span> [<a href="https://arxiv.org/pdf/2401.02113">pdf</a>, <a href="https://arxiv.org/format/2401.02113">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Source-Free Online Domain Adaptive Semantic Segmentation of Satellite Images under Image Degradation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Niloy%2C+F+F">Fahim Faisal Niloy</a>, <a href="/search/cs?searchtype=author&query=Bhaumik%2C+K+K">Kishor Kumar Bhaumik</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.02113v1-abstract-short" style="display: inline;"> Online adaptation to distribution shifts in satellite image segmentation stands as a crucial yet underexplored problem. In this paper, we address source-free and online domain adaptation, i.e., test-time adaptation (TTA), for satellite images, with the focus on mitigating distribution shifts caused by various forms of image degradation. Towards achieving this goal, we propose a novel TTA approach… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.02113v1-abstract-full').style.display = 'inline'; document.getElementById('2401.02113v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.02113v1-abstract-full" style="display: none;"> Online adaptation to distribution shifts in satellite image segmentation stands as a crucial yet underexplored problem. In this paper, we address source-free and online domain adaptation, i.e., test-time adaptation (TTA), for satellite images, with the focus on mitigating distribution shifts caused by various forms of image degradation. Towards achieving this goal, we propose a novel TTA approach involving two effective strategies. First, we progressively estimate the global Batch Normalization (BN) statistics of the target distribution with incoming data stream. Leveraging these statistics during inference has the ability to effectively reduce domain gap. Furthermore, we enhance prediction quality by refining the predicted masks using global class centers. Both strategies employ dynamic momentum for fast and stable convergence. Notably, our method is backpropagation-free and hence fast and lightweight, making it highly suitable for on-the-fly adaptation to new domain. Through comprehensive experiments across various domain adaptation scenarios, we demonstrate the robust performance of our method. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.02113v1-abstract-full').style.display = 'none'; document.getElementById('2401.02113v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 4 January, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ICASSP 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2312.16823">arXiv:2312.16823</a> <span> [<a href="https://arxiv.org/pdf/2312.16823">pdf</a>, <a href="https://arxiv.org/format/2312.16823">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> </div> </div> <p class="title is-5 mathjax"> Layer Attack Unlearning: Fast and Accurate Machine Unlearning via Layer Level Attack and Knowledge Distillation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hyunjune Kim</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sangyong Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.16823v1-abstract-short" style="display: inline;"> Recently, serious concerns have been raised about the privacy issues related to training datasets in machine learning algorithms when including personal data. Various regulations in different countries, including the GDPR grant individuals to have personal data erased, known as 'the right to be forgotten' or 'the right to erasure'. However, there has been less research on effectively and practical… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.16823v1-abstract-full').style.display = 'inline'; document.getElementById('2312.16823v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.16823v1-abstract-full" style="display: none;"> Recently, serious concerns have been raised about the privacy issues related to training datasets in machine learning algorithms when including personal data. Various regulations in different countries, including the GDPR grant individuals to have personal data erased, known as 'the right to be forgotten' or 'the right to erasure'. However, there has been less research on effectively and practically deleting the requested personal data from the training set while not jeopardizing the overall machine learning performance. In this work, we propose a fast and novel machine unlearning paradigm at the layer level called layer attack unlearning, which is highly accurate and fast compared to existing machine unlearning algorithms. We introduce the Partial-PGD algorithm to locate the samples to forget efficiently. In addition, we only use the last layer of the model inspired by the Forward-Forward algorithm for unlearning process. Lastly, we use Knowledge Distillation (KD) to reliably learn the decision boundaries from the teacher using soft label information to improve accuracy performance. We conducted extensive experiments with SOTA machine unlearning models and demonstrated the effectiveness of our approach for accuracy and end-to-end unlearning performance. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.16823v1-abstract-full').style.display = 'none'; document.getElementById('2312.16823v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 27 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2312.15980">arXiv:2312.15980</a> <span> [<a href="https://arxiv.org/pdf/2312.15980">pdf</a>, <a href="https://arxiv.org/format/2312.15980">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> HarmonyView: Harmonizing Consistency and Diversity in One-Image-to-3D </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Park%2C+B">Byeongjun Park</a>, <a href="/search/cs?searchtype=author&query=Go%2C+H">Hyojun Go</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jin-Young Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.15980v1-abstract-short" style="display: inline;"> Recent progress in single-image 3D generation highlights the importance of multi-view coherency, leveraging 3D priors from large-scale diffusion models pretrained on Internet-scale images. However, the aspect of novel-view diversity remains underexplored within the research landscape due to the ambiguity in converting a 2D image into 3D content, where numerous potential shapes can emerge. Here, we… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.15980v1-abstract-full').style.display = 'inline'; document.getElementById('2312.15980v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.15980v1-abstract-full" style="display: none;"> Recent progress in single-image 3D generation highlights the importance of multi-view coherency, leveraging 3D priors from large-scale diffusion models pretrained on Internet-scale images. However, the aspect of novel-view diversity remains underexplored within the research landscape due to the ambiguity in converting a 2D image into 3D content, where numerous potential shapes can emerge. Here, we aim to address this research gap by simultaneously addressing both consistency and diversity. Yet, striking a balance between these two aspects poses a considerable challenge due to their inherent trade-offs. This work introduces HarmonyView, a simple yet effective diffusion sampling technique adept at decomposing two intricate aspects in single-image 3D generation: consistency and diversity. This approach paves the way for a more nuanced exploration of the two critical dimensions within the sampling process. Moreover, we propose a new evaluation metric based on CLIP image and text encoders to comprehensively assess the diversity of the generated views, which closely aligns with human evaluators' judgments. In experiments, HarmonyView achieves a harmonious balance, demonstrating a win-win scenario in both consistency and diversity. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.15980v1-abstract-full').style.display = 'none'; document.getElementById('2312.15980v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://byeongjun-park.github.io/HarmonyView/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2312.12807">arXiv:2312.12807</a> <span> [<a href="https://arxiv.org/pdf/2312.12807">pdf</a>, <a href="https://arxiv.org/format/2312.12807">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> All but One: Surgical Concept Erasing with Model Preservation in Text-to-Image Diffusion Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Hong%2C+S">Seunghoo Hong</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+J">Juhun Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S+S">Simon S. Woo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.12807v1-abstract-short" style="display: inline;"> Text-to-Image models such as Stable Diffusion have shown impressive image generation synthesis, thanks to the utilization of large-scale datasets. However, these datasets may contain sexually explicit, copyrighted, or undesirable content, which allows the model to directly generate them. Given that retraining these large models on individual concept deletion requests is infeasible, fine-tuning alg… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.12807v1-abstract-full').style.display = 'inline'; document.getElementById('2312.12807v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.12807v1-abstract-full" style="display: none;"> Text-to-Image models such as Stable Diffusion have shown impressive image generation synthesis, thanks to the utilization of large-scale datasets. However, these datasets may contain sexually explicit, copyrighted, or undesirable content, which allows the model to directly generate them. Given that retraining these large models on individual concept deletion requests is infeasible, fine-tuning algorithms have been developed to tackle concept erasing in diffusion models. While these algorithms yield good concept erasure, they all present one of the following issues: 1) the corrupted feature space yields synthesis of disintegrated objects, 2) the initially synthesized content undergoes a divergence in both spatial structure and semantics in the generated images, and 3) sub-optimal training updates heighten the model's susceptibility to utility harm. These issues severely degrade the original utility of generative models. In this work, we present a new approach that solves all of these challenges. We take inspiration from the concept of classifier guidance and propose a surgical update on the classifier guidance term while constraining the drift of the unconditional score term. Furthermore, our algorithm empowers the user to select an alternative to the erasing concept, allowing for more controllability. Our experimental results show that our algorithm not only erases the target concept effectively but also preserves the model's generation capability. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.12807v1-abstract-full').style.display = 'none'; document.getElementById('2312.12807v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Main paper with supplementary materials</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2312.11575">arXiv:2312.11575</a> <span> [<a href="https://arxiv.org/pdf/2312.11575">pdf</a>, <a href="https://arxiv.org/format/2312.11575">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> </div> <div class="is-inline-block" style="margin-left: 0.5rem"> <div class="tags has-addons"> <span class="tag is-dark is-size-7">doi</span> <span class="tag is-light is-size-7"><a class="" href="https://doi.org/10.1609/aaai.v38i20.30200">10.1609/aaai.v38i20.30200 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Blind-Touch: Homomorphic Encryption-Based Distributed Neural Network Inference for Privacy-Preserving Fingerprint Authentication </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Choi%2C+H">Hyunmin Choi</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Simon Woo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+H">Hyoungshick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.11575v2-abstract-short" style="display: inline;"> Fingerprint authentication is a popular security mechanism for smartphones and laptops. However, its adoption in web and cloud environments has been limited due to privacy concerns over storing and processing biometric data on servers. This paper introduces Blind-Touch, a novel machine learning-based fingerprint authentication system leveraging homomorphic encryption to address these privacy conce… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.11575v2-abstract-full').style.display = 'inline'; document.getElementById('2312.11575v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.11575v2-abstract-full" style="display: none;"> Fingerprint authentication is a popular security mechanism for smartphones and laptops. However, its adoption in web and cloud environments has been limited due to privacy concerns over storing and processing biometric data on servers. This paper introduces Blind-Touch, a novel machine learning-based fingerprint authentication system leveraging homomorphic encryption to address these privacy concerns. Homomorphic encryption allows computations on encrypted data without decrypting. Thus, Blind-Touch can keep fingerprint data encrypted on the server while performing machine learning operations. Blind-Touch combines three strategies to efficiently utilize homomorphic encryption in machine learning: (1) It optimizes the feature vector for a distributed architecture, processing the first fully connected layer (FC-16) in plaintext on the client side and the subsequent layer (FC-1) post-encryption on the server, thereby minimizing encrypted computations; (2) It employs a homomorphic encryption compatible data compression technique capable of handling 8,192 authentication results concurrently; and (3) It utilizes a clustered server architecture to simultaneously process authentication results, thereby enhancing scalability with increasing user numbers. Blind-Touch achieves high accuracy on two benchmark fingerprint datasets, with a 93.6% F1- score for the PolyU dataset and a 98.2% F1-score for the SOKOTO dataset. Moreover, Blind-Touch can match a fingerprint among 5,000 in about 0.65 seconds. With its privacy focused design, high accuracy, and efficiency, Blind-Touch is a promising alternative to conventional fingerprint authentication for web and cloud applications. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.11575v2-abstract-full').style.display = 'none'; document.getElementById('2312.11575v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 1 April, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 18 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">The 38th Annual AAAI Conference on Artificial Intelligence (AAAI) 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2311.12344">arXiv:2311.12344</a> <span> [<a href="https://arxiv.org/pdf/2311.12344">pdf</a>, <a href="https://arxiv.org/format/2311.12344">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Modality Mixer Exploiting Complementary Information for Multi-modal Action Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lee%2C+S">Sumin Lee</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Nugroho%2C+M+A">Muhammad Adi Nugroho</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2311.12344v1-abstract-short" style="display: inline;"> Due to the distinctive characteristics of sensors, each modality exhibits unique physical properties. For this reason, in the context of multi-modal action recognition, it is important to consider not only the overall action content but also the complementary nature of different modalities. In this paper, we propose a novel network, named Modality Mixer (M-Mixer) network, which effectively leverag… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.12344v1-abstract-full').style.display = 'inline'; document.getElementById('2311.12344v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2311.12344v1-abstract-full" style="display: none;"> Due to the distinctive characteristics of sensors, each modality exhibits unique physical properties. For this reason, in the context of multi-modal action recognition, it is important to consider not only the overall action content but also the complementary nature of different modalities. In this paper, we propose a novel network, named Modality Mixer (M-Mixer) network, which effectively leverages and incorporates the complementary information across modalities with the temporal context of actions for action recognition. A key component of our proposed M-Mixer is the Multi-modal Contextualization Unit (MCU), a simple yet effective recurrent unit. Our MCU is responsible for temporally encoding a sequence of one modality (e.g., RGB) with action content features of other modalities (e.g., depth and infrared modalities). This process encourages M-Mixer network to exploit global action content and also to supplement complementary information of other modalities. Furthermore, to extract appropriate complementary information regarding to the given modality settings, we introduce a new module, named Complementary Feature Extraction Module (CFEM). CFEM incorporates sepearte learnable query embeddings for each modality, which guide CFEM to extract complementary information and global action content from the other modalities. As a result, our proposed method outperforms state-of-the-art methods on NTU RGB+D 60, NTU RGB+D 120, and NW-UCLA datasets. Moreover, through comprehensive ablation studies, we further validate the effectiveness of our proposed method. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.12344v1-abstract-full').style.display = 'none'; document.getElementById('2311.12344v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 November, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">arXiv admin note: substantial text overlap with arXiv:2208.11314</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.16354">arXiv:2310.16354</a> <span> [<a href="https://arxiv.org/pdf/2310.16354">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Hardware Architecture">cs.AR</span> </div> </div> <p class="title is-5 mathjax"> RAMPART: RowHammer Mitigation and Repair for Server Memory Systems </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Woo%2C+S+C">Steven C. Woo</a>, <a href="/search/cs?searchtype=author&query=Elsasser%2C+W">Wendy Elsasser</a>, <a href="/search/cs?searchtype=author&query=Hamburg%2C+M">Mike Hamburg</a>, <a href="/search/cs?searchtype=author&query=Linstadt%2C+E">Eric Linstadt</a>, <a href="/search/cs?searchtype=author&query=Miller%2C+M+R">Michael R. Miller</a>, <a href="/search/cs?searchtype=author&query=Song%2C+T">Taeksang Song</a>, <a href="/search/cs?searchtype=author&query=Tringali%2C+J">James Tringali</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.16354v1-abstract-short" style="display: inline;"> RowHammer attacks are a growing security and reliability concern for DRAMs and computer systems as they can induce many bit errors that overwhelm error detection and correction capabilities. System-level solutions are needed as process technology and circuit improvements alone are unlikely to provide complete protection against RowHammer attacks in the future. This paper introduces RAMPART, a nove… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.16354v1-abstract-full').style.display = 'inline'; document.getElementById('2310.16354v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.16354v1-abstract-full" style="display: none;"> RowHammer attacks are a growing security and reliability concern for DRAMs and computer systems as they can induce many bit errors that overwhelm error detection and correction capabilities. System-level solutions are needed as process technology and circuit improvements alone are unlikely to provide complete protection against RowHammer attacks in the future. This paper introduces RAMPART, a novel approach to mitigating RowHammer attacks and improving server memory system reliability by remapping addresses in each DRAM in a way that confines RowHammer bit flips to a single device for any victim row address. When RAMPART is paired with Single Device Data Correction (SDDC) and patrol scrub, error detection and correction methods in use today, the system can detect and correct bit flips from a successful attack, allowing the memory system to heal itself. RAMPART is compatible with DDR5 RowHammer mitigation features, as well as a wide variety of algorithmic and probabilistic tracking methods. We also introduce BRC-VL, a variation of DDR5 Bounded Refresh Configuration (BRC) that improves system performance by reducing mitigation overhead and show that it works well with probabilistic sampling methods to combat traditional and victim-focused mitigation attacks like Half-Double. The combination of RAMPART, SDDC, and scrubbing enables stronger RowHammer resistance by correcting bit flips from one successful attack. Uncorrectable errors are much less likely, requiring two successful attacks before the memory system is scrubbed. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.16354v1-abstract-full').style.display = 'none'; document.getElementById('2310.16354v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 25 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">16 pages, 13 figures. A version of this paper will appear in the Proceedings of MEMSYS23</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">ACM Class:</span> B.3.1; B.3.4 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.07138">arXiv:2310.07138</a> <span> [<a href="https://arxiv.org/pdf/2310.07138">pdf</a>, <a href="https://arxiv.org/format/2310.07138">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Denoising Task Routing for Diffusion Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Park%2C+B">Byeongjun Park</a>, <a href="/search/cs?searchtype=author&query=Woo%2C+S">Sangmin Woo</a>, <a href="/search/cs?searchtype=author&query=Go%2C+H">Hyojun Go</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+J">Jin-Young Kim</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+C">Changick Kim</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.07138v3-abstract-short" style="display: inline;"> Diffusion models generate highly realistic images by learning a multi-step denoising process, naturally embodying the principles of multi-task learning (MTL). Despite the inherent connection between diffusion models and MTL, there remains an unexplored area in designing neural architectures that explicitly incorporate MTL into the framework of diffusion models. In this paper, we present Denoising… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.07138v3-abstract-full').style.display = 'inline'; document.getElementById('2310.07138v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.07138v3-abstract-full" style="display: none;"> Diffusion models generate highly realistic images by learning a multi-step denoising process, naturally embodying the principles of multi-task learning (MTL). Despite the inherent connection between diffusion models and MTL, there remains an unexplored area in designing neural architectures that explicitly incorporate MTL into the framework of diffusion models. In this paper, we present Denoising Task Routing (DTR), a simple add-on strategy for existing diffusion model architectures to establish distinct information pathways for individual tasks within a single architecture by selectively activating subsets of channels in the model. What makes DTR particularly compelling is its seamless integration of prior knowledge of denoising tasks into the framework: (1) Task Affinity: DTR activates similar channels for tasks at adjacent timesteps and shifts activated channels as sliding windows through timesteps, capitalizing on the inherent strong affinity between tasks at adjacent timesteps. (2) Task Weights: During the early stages (higher timesteps) of the denoising process, DTR assigns a greater number of task-specific channels, leveraging the insight that diffusion models prioritize reconstructing global structure and perceptually rich contents in earlier stages, and focus on simple noise removal in later stages. Our experiments reveal that DTR not only consistently boosts diffusion models' performance across different evaluation protocols without adding extra parameters but also accelerates training convergence. Finally, we show the complementarity between our architectural approach and existing MTL optimization techniques, providing a more complete view of MTL in the context of diffusion training. Significantly, by leveraging this complementarity, we attain matched performance of DiT-XL using the smaller DiT-L with a reduction in training iterations from 7M to 2M. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.07138v3-abstract-full').style.display = 'none'; document.getElementById('2310.07138v3-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 10 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ICLR 2024</span> </p> </li> </ol> <nav class="pagination is-small is-centered breathe-horizontal" role="navigation" aria-label="pagination"> <a href="" class="pagination-previous is-invisible">Previous </a> <a href="/search/?searchtype=author&query=Woo%2C+S&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Woo%2C+S&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> </ul> </nav> <div class="is-hidden-tablet">  <span class="help" style="display: inline-block;"><a href="https://github.com/arXiv/arxiv-search/releases">Search v0.5.6 released 2020-02-24</a>  </span> </div> </div> </main> <footer> <div class="columns is-desktop" role="navigation" aria-label="Secondary">  <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/about">About</a></li> <li><a href="https://info.arxiv.org/help">Help</a></li> </ul> </div> <div class="column"> <ul class="nav-spaced"> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>contact arXiv</title><desc>Click here to contact arXiv</desc><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg> <a href="https://info.arxiv.org/help/contact.html"> Contact</a> </li> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>subscribe to arXiv mailings</title><desc>Click here to subscribe</desc><path d="M476 3.2L12.5 270.6c-18.1 10.4-15.8 35.6 2.2 43.2L121 358.4l287.3-253.2c5.5-4.9 13.3 2.6 8.6 8.3L176 407v80.5c0 23.6 28.5 32.9 42.5 15.8L282 426l124.6 52.2c14.2 6 30.4-2.9 33-18.2l72-432C515 7.8 493.3-6.8 476 3.2z"/></svg> <a href="https://info.arxiv.org/help/subscribe"> Subscribe</a> </li> </ul> </div> </div> </div>   <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/license/index.html">Copyright</a></li> <li><a href="https://info.arxiv.org/help/policies/privacy_policy.html">Privacy Policy</a></li> </ul> </div> <div class="column sorry-app-links"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/web_accessibility.html">Web Accessibility Assistance</a></li> <li> <p class="help"> <a class="a11y-main-link" href="https://status.arxiv.org" target="_blank">arXiv Operational Status <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 256 512" class="icon filter-dark_grey" role="presentation"><path d="M224.3 273l-136 136c-9.4 9.4-24.6 9.4-33.9 0l-22.6-22.6c-9.4-9.4-9.4-24.6 0-33.9l96.4-96.4-96.4-96.4c-9.4-9.4-9.4-24.6 0-33.9L54.3 103c9.4-9.4 24.6-9.4 33.9 0l136 136c9.5 9.4 9.5 24.6.1 34z"/></svg></a><br> Get status notifications via <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/email/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg>email</a> or <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/slack/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon filter-black" role="presentation"><path d="M94.12 315.1c0 25.9-21.16 47.06-47.06 47.06S0 341 0 315.1c0-25.9 21.16-47.06 47.06-47.06h47.06v47.06zm23.72 0c0-25.9 21.16-47.06 47.06-47.06s47.06 21.16 47.06 47.06v117.84c0 25.9-21.16 47.06-47.06 47.06s-47.06-21.16-47.06-47.06V315.1zm47.06-188.98c-25.9 0-47.06-21.16-47.06-47.06S139 32 164.9 32s47.06 21.16 47.06 47.06v47.06H164.9zm0 23.72c25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06H47.06C21.16 243.96 0 222.8 0 196.9s21.16-47.06 47.06-47.06H164.9zm188.98 47.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06h-47.06V196.9zm-23.72 0c0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06V79.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06V196.9zM283.1 385.88c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06v-47.06h47.06zm0-23.72c-25.9 0-47.06-21.16-47.06-47.06 0-25.9 21.16-47.06 47.06-47.06h117.84c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06H283.1z"/></svg>slack</a> </p> </li> </ul> </div> </div> </div>  </div> </footer> <script src="https://static.arxiv.org/static/base/1.0.0a5/js/member_acknowledgement.js"></script> </body> </html>

CINXE.COM

Search | arXiv e-print repository