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 376 results for author: <span class="mathjax">Luo, C</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=Luo%2C+C">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="Luo, C"> </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=Luo%2C+C&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="Luo, C"> <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=Luo%2C+C&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=200" class="pagination-link " aria-label="Page 5" aria-current="page">5 </a> </li> <li><span class="pagination-ellipsis">…</span></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/2503.18434">arXiv:2503.18434</a> <span> [<a href="https://arxiv.org/pdf/2503.18434">pdf</a>, <a href="https://arxiv.org/format/2503.18434">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"> A Simple yet Effective Layout Token in Large Language Models for Document Understanding </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhu%2C+Z">Zhaoqing Zhu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuwei Luo</a>, <a href="/search/cs?searchtype=author&query=Shao%2C+Z">Zirui Shao</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+F">Feiyu Gao</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+H">Hangdi Xing</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+Q">Qi Zheng</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Ji Zhang</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="2503.18434v1-abstract-short" style="display: inline;"> Recent methods that integrate spatial layouts with text for document understanding in large language models (LLMs) have shown promising results. A commonly used method is to represent layout information as text tokens and interleave them with text content as inputs to the LLMs. However, such a method still demonstrates limitations, as it requires additional position IDs for tokens that are used to… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.18434v1-abstract-full').style.display = 'inline'; document.getElementById('2503.18434v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.18434v1-abstract-full" style="display: none;"> Recent methods that integrate spatial layouts with text for document understanding in large language models (LLMs) have shown promising results. A commonly used method is to represent layout information as text tokens and interleave them with text content as inputs to the LLMs. However, such a method still demonstrates limitations, as it requires additional position IDs for tokens that are used to represent layout information. Due to the constraint on max position IDs, assigning them to layout information reduces those available for text content, reducing the capacity for the model to learn from the text during training, while also introducing a large number of potentially untrained position IDs during long-context inference, which can hinder performance on document understanding tasks. To address these issues, we propose LayTokenLLM, a simple yet effective method for document understanding. LayTokenLLM represents layout information as a single token per text segment and uses a specialized positional encoding scheme. It shares position IDs between text and layout tokens, eliminating the need for additional position IDs. This design maintains the model's capacity to learn from text while mitigating long-context issues during inference. Furthermore, a novel pre-training objective called Next Interleaved Text and Layout Token Prediction (NTLP) is devised to enhance cross-modality learning between text and layout tokens. Extensive experiments show that LayTokenLLM outperforms existing layout-integrated LLMs and MLLMs of similar scales on multi-page document understanding tasks, as well as most single-page tasks. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.18434v1-abstract-full').style.display = 'none'; document.getElementById('2503.18434v1-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </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">CVPR 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/2503.16463">arXiv:2503.16463</a> <span> [<a href="https://arxiv.org/pdf/2503.16463">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Human-Computer Interaction">cs.HC</span> </div> </div> <p class="title is-5 mathjax"> Improving Interactive Diagnostic Ability of a Large Language Model Agent Through Clinical Experience Learning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sun%2C+Z">Zhoujian Sun</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Z">Ziyi Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Chu%2C+J">Jiebin Chu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+Z">Zhengxing Huang</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="2503.16463v1-abstract-short" style="display: inline;"> Recent advances in large language models (LLMs) have shown promising results in medical diagnosis, with some studies indicating superior performance compared to human physicians in specific scenarios. However, the diagnostic capabilities of LLMs are often overestimated, as their performance significantly deteriorates in interactive diagnostic settings that require active information gathering. Thi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.16463v1-abstract-full').style.display = 'inline'; document.getElementById('2503.16463v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.16463v1-abstract-full" style="display: none;"> Recent advances in large language models (LLMs) have shown promising results in medical diagnosis, with some studies indicating superior performance compared to human physicians in specific scenarios. However, the diagnostic capabilities of LLMs are often overestimated, as their performance significantly deteriorates in interactive diagnostic settings that require active information gathering. This study investigates the underlying mechanisms behind the performance degradation phenomenon and proposes a solution. We identified that the primary deficiency of LLMs lies in the initial diagnosis phase, particularly in information-gathering efficiency and initial diagnosis formation, rather than in the subsequent differential diagnosis phase. To address this limitation, we developed a plug-and-play method enhanced (PPME) LLM agent, leveraging over 3.5 million electronic medical records from Chinese and American healthcare facilities. Our approach integrates specialized models for initial disease diagnosis and inquiry into the history of the present illness, trained through supervised and reinforcement learning techniques. The experimental results indicate that the PPME LLM achieved over 30% improvement compared to baselines. The final diagnostic accuracy of the PPME LLM in interactive diagnostic scenarios approached levels comparable to those achieved using complete clinical data. These findings suggest a promising potential for developing autonomous diagnostic systems, although further validation studies are needed. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.16463v1-abstract-full').style.display = 'none'; document.getElementById('2503.16463v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </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">30 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/2503.13518">arXiv:2503.13518</a> <span> [<a href="https://arxiv.org/pdf/2503.13518">pdf</a>, <a href="https://arxiv.org/format/2503.13518">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Examples as the Prompt: A Scalable Approach for Efficient LLM Adaptation in E-Commerce </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zeng%2C+J">Jingying Zeng</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Z">Zhenwei Dai</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Varshney%2C+S">Samarth Varshney</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Z">Zhiji Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zhen Li</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</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="2503.13518v1-abstract-short" style="display: inline;"> Prompting LLMs offers an efficient way to guide output generation without explicit model training. In the e-commerce domain, prompting-based applications are widely used for tasks such as query understanding, recommender systems, and customer support. However, adapting LLMs to different tasks often requires extensive prompt engineering by domain experts, along with frequent updates to align with e… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.13518v1-abstract-full').style.display = 'inline'; document.getElementById('2503.13518v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.13518v1-abstract-full" style="display: none;"> Prompting LLMs offers an efficient way to guide output generation without explicit model training. In the e-commerce domain, prompting-based applications are widely used for tasks such as query understanding, recommender systems, and customer support. However, adapting LLMs to different tasks often requires extensive prompt engineering by domain experts, along with frequent updates to align with evolving business needs. Additionally, crafting fully unbiased natural language prompts remains a challenge for humans. To address these challenges, we propose a novel framework, Examples as the Prompt (EaP) which leverages labeled data to enhance prompts. Specifically, EaP automatically selects the most representative examples to maximize the few-shot capability of LLMs. It is efficient due to its unsupervised example selection and adaptive to potential data distribution shifts. We validate EaP on four real-world production use cases, demonstrating that it achieves comparable or even superior performance comparing to hand-crafted prompts designed by domain experts. Additionally, we introduce EaP_lite, which entirely replaces the natural language components of prompts with labeled examples. EaP_lite improves LLM inference speed by up to 70% without compromising performance. Latest online A/B test shows that using EaP and EaP_lite for data labeling can bring significant composite revenue gain by 0.06%. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.13518v1-abstract-full').style.display = 'none'; document.getElementById('2503.13518v1-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2503.11513">arXiv:2503.11513</a> <span> [<a href="https://arxiv.org/pdf/2503.11513">pdf</a>, <a href="https://arxiv.org/format/2503.11513">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"> HiTVideo: Hierarchical Tokenizers for Enhancing Text-to-Video Generation with Autoregressive Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhou%2C+Z">Ziqin Zhou</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yifan Yang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yuqing Yang</a>, <a href="/search/cs?searchtype=author&query=He%2C+T">Tianyu He</a>, <a href="/search/cs?searchtype=author&query=Peng%2C+H">Houwen Peng</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+K">Kai Qiu</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+L">Lili Qiu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+L">Lingqiao 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="2503.11513v1-abstract-short" style="display: inline;"> Text-to-video generation poses significant challenges due to the inherent complexity of video data, which spans both temporal and spatial dimensions. It introduces additional redundancy, abrupt variations, and a domain gap between language and vision tokens while generation. Addressing these challenges requires an effective video tokenizer that can efficiently encode video data while preserving es… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.11513v1-abstract-full').style.display = 'inline'; document.getElementById('2503.11513v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.11513v1-abstract-full" style="display: none;"> Text-to-video generation poses significant challenges due to the inherent complexity of video data, which spans both temporal and spatial dimensions. It introduces additional redundancy, abrupt variations, and a domain gap between language and vision tokens while generation. Addressing these challenges requires an effective video tokenizer that can efficiently encode video data while preserving essential semantic and spatiotemporal information, serving as a critical bridge between text and vision. Inspired by the observation in VQ-VAE-2 and workflows of traditional animation, we propose HiTVideo for text-to-video generation with hierarchical tokenizers. It utilizes a 3D causal VAE with a multi-layer discrete token framework, encoding video content into hierarchically structured codebooks. Higher layers capture semantic information with higher compression, while lower layers focus on fine-grained spatiotemporal details, striking a balance between compression efficiency and reconstruction quality. Our approach efficiently encodes longer video sequences (e.g., 8 seconds, 64 frames), reducing bits per pixel (bpp) by approximately 70\% compared to baseline tokenizers, while maintaining competitive reconstruction quality. We explore the trade-offs between compression and reconstruction, while emphasizing the advantages of high-compressed semantic tokens in text-to-video tasks. HiTVideo aims to address the potential limitations of existing video tokenizers in text-to-video generation tasks, striving for higher compression ratios and simplify LLMs modeling under language guidance, offering a scalable and promising framework for advancing text to video generation. Demo page: https://ziqinzhou66.github.io/project/HiTVideo. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.11513v1-abstract-full').style.display = 'none'; document.getElementById('2503.11513v1-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2503.04830">arXiv:2503.04830</a> <span> [<a href="https://arxiv.org/pdf/2503.04830">pdf</a>, <a href="https://arxiv.org/format/2503.04830">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Cite Before You Speak: Enhancing Context-Response Grounding in E-commerce Conversational LLM-Agents </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zeng%2C+J">Jingying Zeng</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Z">Zhenwei Dai</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Varshney%2C+S">Samarth Varshney</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zhen Li</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</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="2503.04830v2-abstract-short" style="display: inline;"> With the advancement of conversational large language models (LLMs), several LLM-based Conversational Shopping Agents (CSA) have been developed to help customers answer questions and smooth their shopping journey in e-commerce domain. The primary objective in building a trustworthy CSA is to ensure the agent's responses are accurate and factually grounded, which is essential for building customer… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.04830v2-abstract-full').style.display = 'inline'; document.getElementById('2503.04830v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.04830v2-abstract-full" style="display: none;"> With the advancement of conversational large language models (LLMs), several LLM-based Conversational Shopping Agents (CSA) have been developed to help customers answer questions and smooth their shopping journey in e-commerce domain. The primary objective in building a trustworthy CSA is to ensure the agent's responses are accurate and factually grounded, which is essential for building customer trust and encouraging continuous engagement. However, two challenges remain. First, LLMs produce hallucinated or unsupported claims. Such inaccuracies risk spreading misinformation and diminishing customer trust. Second, without providing knowledge source attribution in CSA response, customers struggle to verify LLM-generated information. To address these challenges, we present an easily productionized solution that enables a "citation experience" utilizing In-context Learning (ICL) and Multi-UX-Inference (MUI) to generate responses with citations to attribute its original sources without interfering other existing UX features. With proper UX design, these citation marks can be linked to the related product information and display the source to our customers. In this work, we also build auto-metrics and scalable benchmarks to holistically evaluate LLM's grounding and attribution capabilities. Our experiments demonstrate that incorporating this citation generation paradigm can substantially enhance the grounding of LLM responses by 13.83% on the real-world data. As such, our solution not only addresses the immediate challenges of LLM grounding issues but also adds transparency to conversational AI. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.04830v2-abstract-full').style.display = 'none'; document.getElementById('2503.04830v2-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 5 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2503.04782">arXiv:2503.04782</a> <span> [<a href="https://arxiv.org/pdf/2503.04782">pdf</a>, <a href="https://arxiv.org/format/2503.04782">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Logic in Computer Science">cs.LO</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"> SMT(LIA) Sampling with High Diversity </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lai%2C+Y">Yong Lai</a>, <a href="/search/cs?searchtype=author&query=Li%2C+J">Junjie Li</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuan Luo</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="2503.04782v1-abstract-short" style="display: inline;"> Satisfiability Modulo Linear Integer Arithmetic, SMT(LIA) for short, is pivotal across various critical domains. Previous research has primarily focused on SMT solving techniques. However, in practical applications such as software and hardware testing, there is a need to generate a diverse set of solutions for use as test inputs. We have developed the first sampling framework that integrates loca… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.04782v1-abstract-full').style.display = 'inline'; document.getElementById('2503.04782v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.04782v1-abstract-full" style="display: none;"> Satisfiability Modulo Linear Integer Arithmetic, SMT(LIA) for short, is pivotal across various critical domains. Previous research has primarily focused on SMT solving techniques. However, in practical applications such as software and hardware testing, there is a need to generate a diverse set of solutions for use as test inputs. We have developed the first sampling framework that integrates local search with CDCL(T) techniques, named HighDiv, capable of generating a highly diverse set of solutions for constraints under linear integer theory. Initially, in the local search phase, we introduced a novel operator called boundary-aware movement. This operator performs random moves by considering the current state's constraints on variables, thereby enhancing the diversity of variables during the search process. Furthermore, we have conducted an in-depth study of the preprocessing and variable initialization mechanisms within the framework, which significantly enhances the efficiency of subsequent local searches. Lastly, we use the solutions obtained from local search sampling as additional constraints to further explore the solution space using the stochastic CDCL(T) method. Experimental results demonstrate that \HighDiv generates solutions with greater diversity compared to the state-of-the-art SMT(LIA) sampling tool, MeGASampler. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.04782v1-abstract-full').style.display = 'none'; document.getElementById('2503.04782v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2503.01743">arXiv:2503.01743</a> <span> [<a href="https://arxiv.org/pdf/2503.01743">pdf</a>, <a href="https://arxiv.org/format/2503.01743">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Microsoft"> Microsoft</a>, <a href="/search/cs?searchtype=author&query=%3A"> :</a>, <a href="/search/cs?searchtype=author&query=Abouelenin%2C+A">Abdelrahman Abouelenin</a>, <a href="/search/cs?searchtype=author&query=Ashfaq%2C+A">Atabak Ashfaq</a>, <a href="/search/cs?searchtype=author&query=Atkinson%2C+A">Adam Atkinson</a>, <a href="/search/cs?searchtype=author&query=Awadalla%2C+H">Hany Awadalla</a>, <a href="/search/cs?searchtype=author&query=Bach%2C+N">Nguyen Bach</a>, <a href="/search/cs?searchtype=author&query=Bao%2C+J">Jianmin Bao</a>, <a href="/search/cs?searchtype=author&query=Benhaim%2C+A">Alon Benhaim</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+M">Martin Cai</a>, <a href="/search/cs?searchtype=author&query=Chaudhary%2C+V">Vishrav Chaudhary</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+C">Congcong Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+D">Dong Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+D">Dongdong Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+J">Junkun Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+W">Weizhu Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yen-Chun Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yi-ling Chen</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+X">Xiyang Dai</a>, <a href="/search/cs?searchtype=author&query=Fan%2C+R">Ruchao Fan</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+M">Mei Gao</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+M">Min Gao</a>, <a href="/search/cs?searchtype=author&query=Garg%2C+A">Amit Garg</a>, <a href="/search/cs?searchtype=author&query=Goswami%2C+A">Abhishek Goswami</a> , et al. (51 additional authors not shown) </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="2503.01743v2-abstract-short" style="display: inline;"> We introduce Phi-4-Mini and Phi-4-Multimodal, compact yet highly capable language and multimodal models. Phi-4-Mini is a 3.8-billion-parameter language model trained on high-quality web and synthetic data, significantly outperforming recent open-source models of similar size and matching the performance of models twice its size on math and coding tasks requiring complex reasoning. This achievement… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.01743v2-abstract-full').style.display = 'inline'; document.getElementById('2503.01743v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.01743v2-abstract-full" style="display: none;"> We introduce Phi-4-Mini and Phi-4-Multimodal, compact yet highly capable language and multimodal models. Phi-4-Mini is a 3.8-billion-parameter language model trained on high-quality web and synthetic data, significantly outperforming recent open-source models of similar size and matching the performance of models twice its size on math and coding tasks requiring complex reasoning. This achievement is driven by a carefully curated synthetic data recipe emphasizing high-quality math and coding datasets. Compared to its predecessor, Phi-3.5-Mini, Phi-4-Mini features an expanded vocabulary size of 200K tokens to better support multilingual applications, as well as group query attention for more efficient long-sequence generation. Phi-4-Multimodal is a multimodal model that integrates text, vision, and speech/audio input modalities into a single model. Its novel modality extension approach leverages LoRA adapters and modality-specific routers to allow multiple inference modes combining various modalities without interference. For example, it now ranks first in the OpenASR leaderboard to date, although the LoRA component of the speech/audio modality has just 460 million parameters. Phi-4-Multimodal supports scenarios involving (vision + language), (vision + speech), and (speech/audio) inputs, outperforming larger vision-language and speech-language models on a wide range of tasks. Additionally, we experiment to further train Phi-4-Mini to enhance its reasoning capabilities. Despite its compact 3.8-billion-parameter size, this experimental version achieves reasoning performance on par with or surpassing significantly larger models, including DeepSeek-R1-Distill-Qwen-7B and DeepSeek-R1-Distill-Llama-8B. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.01743v2-abstract-full').style.display = 'none'; document.getElementById('2503.01743v2-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, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 3 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </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">39 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/2503.00334">arXiv:2503.00334</a> <span> [<a href="https://arxiv.org/pdf/2503.00334">pdf</a>, <a href="https://arxiv.org/format/2503.00334">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 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/10.1145/3696410.3714802">10.1145/3696410.3714802 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> MCNet: Monotonic Calibration Networks for Expressive Uncertainty Calibration in Online Advertising </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Quanyu Dai</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+J">Jiaren Xiao</a>, <a href="/search/cs?searchtype=author&query=Du%2C+Z">Zhaocheng Du</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+J">Jieming Zhu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chengxiao Luo</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+X">Xiao-Ming Wu</a>, <a href="/search/cs?searchtype=author&query=Dong%2C+Z">Zhenhua Dong</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="2503.00334v1-abstract-short" style="display: inline;"> In online advertising, uncertainty calibration aims to adjust a ranking model's probability predictions to better approximate the true likelihood of an event, e.g., a click or a conversion. However, existing calibration approaches may lack the ability to effectively model complex nonlinear relations, consider context features, and achieve balanced performance across different data subsets. To tack… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.00334v1-abstract-full').style.display = 'inline'; document.getElementById('2503.00334v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2503.00334v1-abstract-full" style="display: none;"> In online advertising, uncertainty calibration aims to adjust a ranking model's probability predictions to better approximate the true likelihood of an event, e.g., a click or a conversion. However, existing calibration approaches may lack the ability to effectively model complex nonlinear relations, consider context features, and achieve balanced performance across different data subsets. To tackle these challenges, we introduce a novel model called Monotonic Calibration Networks, featuring three key designs: a monotonic calibration function (MCF), an order-preserving regularizer, and a field-balance regularizer. The nonlinear MCF is capable of naturally modeling and universally approximating the intricate relations between uncalibrated predictions and the posterior probabilities, thus being much more expressive than existing methods. MCF can also integrate context features using a flexible model architecture, thereby achieving context awareness. The order-preserving and field-balance regularizers promote the monotonic relationship between adjacent bins and the balanced calibration performance on data subsets, respectively. Experimental results on both public and industrial datasets demonstrate the superior performance of our method in generating well-calibrated probability predictions. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2503.00334v1-abstract-full').style.display = 'none'; document.getElementById('2503.00334v1-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, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2025. </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 WWW2025</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">ACM Class:</span> H.0 </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> THE ACM WEB CONFERENCE 2025 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.20061">arXiv:2502.20061</a> <span> [<a href="https://arxiv.org/pdf/2502.20061">pdf</a>, <a href="https://arxiv.org/format/2502.20061">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"> HiFAR: Multi-Stage Curriculum Learning for High-Dynamics Humanoid Fall Recovery </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Chen%2C+P">Penghui Chen</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yushi Wang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Changsheng Luo</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+W">Wenhan Cai</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+M">Mingguo Zhao</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="2502.20061v2-abstract-short" style="display: inline;"> Humanoid robots encounter considerable difficulties in autonomously recovering from falls, especially within dynamic and unstructured environments. Conventional control methodologies are often inadequate in addressing the complexities associated with high-dimensional dynamics and the contact-rich nature of fall recovery. Meanwhile, reinforcement learning techniques are hindered by issues related t… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.20061v2-abstract-full').style.display = 'inline'; document.getElementById('2502.20061v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.20061v2-abstract-full" style="display: none;"> Humanoid robots encounter considerable difficulties in autonomously recovering from falls, especially within dynamic and unstructured environments. Conventional control methodologies are often inadequate in addressing the complexities associated with high-dimensional dynamics and the contact-rich nature of fall recovery. Meanwhile, reinforcement learning techniques are hindered by issues related to sparse rewards, intricate collision scenarios, and discrepancies between simulation and real-world applications. In this study, we introduce a multi-stage curriculum learning framework, termed HiFAR. This framework employs a staged learning approach that progressively incorporates increasingly complex and high-dimensional recovery tasks, thereby facilitating the robot's acquisition of efficient and stable fall recovery strategies. Furthermore, it enables the robot to adapt its policy to effectively manage real-world fall incidents. We assess the efficacy of the proposed method using a real humanoid robot, showcasing its capability to autonomously recover from a diverse range of falls with high success rates, rapid recovery times, robustness, and generalization. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.20061v2-abstract-full').style.display = 'none'; document.getElementById('2502.20061v2-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, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 27 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.18387">arXiv:2502.18387</a> <span> [<a href="https://arxiv.org/pdf/2502.18387">pdf</a>, <a href="https://arxiv.org/format/2502.18387">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> How Far are LLMs from Real Search? A Comprehensive Study on Efficiency, Completeness, and Inherent Capabilities </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lin%2C+M">Minhua Lin</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Zeng%2C+J">Jingying Zeng</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Z">Zhenwei Dai</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zheng Li</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+X">Xiang Zhang</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+S">Suhang Wang</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="2502.18387v2-abstract-short" style="display: inline;"> Search plays a fundamental role in problem-solving across various domains, with most real-world decision-making problems being solvable through systematic search. Drawing inspiration from recent discussions on search and learning, we systematically explore the complementary relationship between search and Large Language Models (LLMs) from three perspectives. First, we analyze how learning can enha… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.18387v2-abstract-full').style.display = 'inline'; document.getElementById('2502.18387v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.18387v2-abstract-full" style="display: none;"> Search plays a fundamental role in problem-solving across various domains, with most real-world decision-making problems being solvable through systematic search. Drawing inspiration from recent discussions on search and learning, we systematically explore the complementary relationship between search and Large Language Models (LLMs) from three perspectives. First, we analyze how learning can enhance search efficiency and propose Search via Learning (SeaL), a framework that leverages LLMs for effective and efficient search. Second, we further extend SeaL to SeaL-C to ensure rigorous completeness during search. Our evaluation across three real-world planning tasks demonstrates that SeaL achieves near-perfect accuracy while reducing search spaces by up to 99.1% compared to traditional approaches. Finally, we explore how far LLMs are from real search by investigating whether they can develop search capabilities independently. Our analysis reveals that while current LLMs struggle with efficient search in complex problems, incorporating systematic search strategies significantly enhances their problem-solving capabilities. These findings not only validate the effectiveness of our approach but also highlight the need for improving LLMs' search abilities for real-world applications. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.18387v2-abstract-full').style.display = 'none'; document.getElementById('2502.18387v2-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 25 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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">31 pages, 9 figures, 18 tables</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.17943">arXiv:2502.17943</a> <span> [<a href="https://arxiv.org/pdf/2502.17943">pdf</a>, <a href="https://arxiv.org/format/2502.17943">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> CaseGen: A Benchmark for Multi-Stage Legal Case Documents Generation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Li%2C+H">Haitao Li</a>, <a href="/search/cs?searchtype=author&query=Ye%2C+J">Jiaying Ye</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+Y">Yiran Hu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+J">Jia Chen</a>, <a href="/search/cs?searchtype=author&query=Ai%2C+Q">Qingyao Ai</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Y">Yueyue Wu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+J">Junjie Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yifan Chen</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Q">Quan Zhou</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Y">Yiqun 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="2502.17943v1-abstract-short" style="display: inline;"> Legal case documents play a critical role in judicial proceedings. As the number of cases continues to rise, the reliance on manual drafting of legal case documents is facing increasing pressure and challenges. The development of large language models (LLMs) offers a promising solution for automating document generation. However, existing benchmarks fail to fully capture the complexities involved… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.17943v1-abstract-full').style.display = 'inline'; document.getElementById('2502.17943v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.17943v1-abstract-full" style="display: none;"> Legal case documents play a critical role in judicial proceedings. As the number of cases continues to rise, the reliance on manual drafting of legal case documents is facing increasing pressure and challenges. The development of large language models (LLMs) offers a promising solution for automating document generation. However, existing benchmarks fail to fully capture the complexities involved in drafting legal case documents in real-world scenarios. To address this gap, we introduce CaseGen, the benchmark for multi-stage legal case documents generation in the Chinese legal domain. CaseGen is based on 500 real case samples annotated by legal experts and covers seven essential case sections. It supports four key tasks: drafting defense statements, writing trial facts, composing legal reasoning, and generating judgment results. To the best of our knowledge, CaseGen is the first benchmark designed to evaluate LLMs in the context of legal case document generation. To ensure an accurate and comprehensive evaluation, we design the LLM-as-a-judge evaluation framework and validate its effectiveness through human annotations. We evaluate several widely used general-domain LLMs and legal-specific LLMs, highlighting their limitations in case document generation and pinpointing areas for potential improvement. This work marks a step toward a more effective framework for automating legal case documents drafting, paving the way for the reliable application of AI in the legal field. The dataset and code are publicly available at https://github.com/CSHaitao/CaseGen. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.17943v1-abstract-full').style.display = 'none'; document.getElementById('2502.17943v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.16868">arXiv:2502.16868</a> <span> [<a href="https://arxiv.org/pdf/2502.16868">pdf</a>, <a href="https://arxiv.org/format/2502.16868">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Databases">cs.DB</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="Human-Computer Interaction">cs.HC</span> </div> </div> <p class="title is-5 mathjax"> Graphy'our Data: Towards End-to-End Modeling, Exploring and Generating Report from Raw Data </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lai%2C+L">Longbin Lai</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Changwei Luo</a>, <a href="/search/cs?searchtype=author&query=Lou%2C+Y">Yunkai Lou</a>, <a href="/search/cs?searchtype=author&query=Ju%2C+M">Mingchen Ju</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Z">Zhengyi 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="2502.16868v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) have recently demonstrated remarkable performance in tasks such as Retrieval-Augmented Generation (RAG) and autonomous AI agent workflows. Yet, when faced with large sets of unstructured documents requiring progressive exploration, analysis, and synthesis, such as conducting literature survey, existing approaches often fall short. We address this challenge -- termed Pr… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16868v1-abstract-full').style.display = 'inline'; document.getElementById('2502.16868v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.16868v1-abstract-full" style="display: none;"> Large Language Models (LLMs) have recently demonstrated remarkable performance in tasks such as Retrieval-Augmented Generation (RAG) and autonomous AI agent workflows. Yet, when faced with large sets of unstructured documents requiring progressive exploration, analysis, and synthesis, such as conducting literature survey, existing approaches often fall short. We address this challenge -- termed Progressive Document Investigation -- by introducing Graphy, an end-to-end platform that automates data modeling, exploration and high-quality report generation in a user-friendly manner. Graphy comprises an offline Scrapper that transforms raw documents into a structured graph of Fact and Dimension nodes, and an online Surveyor that enables iterative exploration and LLM-driven report generation. We showcase a pre-scrapped graph of over 50,000 papers -- complete with their references -- demonstrating how Graphy facilitates the literature-survey scenario. The demonstration video can be found at https://youtu.be/uM4nzkAdGlM. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16868v1-abstract-full').style.display = 'none'; document.getElementById('2502.16868v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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">4 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/2502.14768">arXiv:2502.14768</a> <span> [<a href="https://arxiv.org/pdf/2502.14768">pdf</a>, <a href="https://arxiv.org/format/2502.14768">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xie%2C+T">Tian Xie</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+Z">Zitian Gao</a>, <a href="/search/cs?searchtype=author&query=Ren%2C+Q">Qingnan Ren</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+H">Haoming Luo</a>, <a href="/search/cs?searchtype=author&query=Hong%2C+Y">Yuqian Hong</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+B">Bryan Dai</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+J">Joey Zhou</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+K">Kai Qiu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zhirong Wu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</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="2502.14768v1-abstract-short" style="display: inline;"> Inspired by the success of DeepSeek-R1, we explore the potential of rule-based reinforcement learning (RL) in large reasoning models. To analyze reasoning dynamics, we use synthetic logic puzzles as training data due to their controllable complexity and straightforward answer verification. We make some key technical contributions that lead to effective and stable RL training: a system prompt that… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.14768v1-abstract-full').style.display = 'inline'; document.getElementById('2502.14768v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.14768v1-abstract-full" style="display: none;"> Inspired by the success of DeepSeek-R1, we explore the potential of rule-based reinforcement learning (RL) in large reasoning models. To analyze reasoning dynamics, we use synthetic logic puzzles as training data due to their controllable complexity and straightforward answer verification. We make some key technical contributions that lead to effective and stable RL training: a system prompt that emphasizes the thinking and answering process, a stringent format reward function that penalizes outputs for taking shortcuts, and a straightforward training recipe that achieves stable convergence. Our 7B model develops advanced reasoning skills-such as reflection, verification, and summarization-that are absent from the logic corpus. Remarkably, after training on just 5K logic problems, it demonstrates generalization abilities to the challenging math benchmarks AIME and AMC. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.14768v1-abstract-full').style.display = 'none'; document.getElementById('2502.14768v1-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, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.13260">arXiv:2502.13260</a> <span> [<a href="https://arxiv.org/pdf/2502.13260">pdf</a>, <a href="https://arxiv.org/format/2502.13260">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Cui%2C+Y">Yingqian Cui</a>, <a href="/search/cs?searchtype=author&query=He%2C+P">Pengfei He</a>, <a href="/search/cs?searchtype=author&query=Zeng%2C+J">Jingying Zeng</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Z">Zhenwei Dai</a>, <a href="/search/cs?searchtype=author&query=Han%2C+Y">Yan Han</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+J">Jing Huang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zhen Li</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+S">Suhang Wang</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+Y">Yue Xing</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+J">Jiliang Tang</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</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="2502.13260v1-abstract-short" style="display: inline;"> Chain-of-Thought (CoT) reasoning, which breaks down complex tasks into intermediate reasoning steps, has significantly enhanced the performance of large language models (LLMs) on challenging tasks. However, the detailed reasoning process in CoT often incurs long generation times and high computational costs, partly due to the inclusion of unnecessary steps. To address this, we propose a method to… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.13260v1-abstract-full').style.display = 'inline'; document.getElementById('2502.13260v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.13260v1-abstract-full" style="display: none;"> Chain-of-Thought (CoT) reasoning, which breaks down complex tasks into intermediate reasoning steps, has significantly enhanced the performance of large language models (LLMs) on challenging tasks. However, the detailed reasoning process in CoT often incurs long generation times and high computational costs, partly due to the inclusion of unnecessary steps. To address this, we propose a method to identify critical reasoning steps using perplexity as a measure of their importance: a step is deemed critical if its removal causes a significant increase in perplexity. Our method enables models to focus solely on generating these critical steps. This can be achieved through two approaches: refining demonstration examples in few-shot CoT or fine-tuning the model using selected examples that include only critical steps. Comprehensive experiments validate the effectiveness of our method, which achieves a better balance between the reasoning accuracy and efficiency of CoT. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.13260v1-abstract-full').style.display = 'none'; document.getElementById('2502.13260v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.12928">arXiv:2502.12928</a> <span> [<a href="https://arxiv.org/pdf/2502.12928">pdf</a>, <a href="https://arxiv.org/format/2502.12928">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> Finedeep: Mitigating Sparse Activation in Dense LLMs via Multi-Layer Fine-Grained Experts </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Pan%2C+L">Leiyu Pan</a>, <a href="/search/cs?searchtype=author&query=Su%2C+Z">Zhenpeng Su</a>, <a href="/search/cs?searchtype=author&query=Lv%2C+M">Minxuan Lv</a>, <a href="/search/cs?searchtype=author&query=Xiong%2C+Y">Yizhe Xiong</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+X">Xiangwen Zhang</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+Z">Zijia Lin</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+H">Hui Chen</a>, <a href="/search/cs?searchtype=author&query=Han%2C+J">Jungong Han</a>, <a href="/search/cs?searchtype=author&query=Ding%2C+G">Guiguang Ding</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+D">Di Zhang</a>, <a href="/search/cs?searchtype=author&query=Gai%2C+K">Kun Gai</a>, <a href="/search/cs?searchtype=author&query=Xiong%2C+D">Deyi Xiong</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="2502.12928v1-abstract-short" style="display: inline;"> Large language models have demonstrated exceptional performance across a wide range of tasks. However, dense models usually suffer from sparse activation, where many activation values tend towards zero (i.e., being inactivated). We argue that this could restrict the efficient exploration of model representation space. To mitigate this issue, we propose Finedeep, a deep-layered fine-grained expert… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12928v1-abstract-full').style.display = 'inline'; document.getElementById('2502.12928v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.12928v1-abstract-full" style="display: none;"> Large language models have demonstrated exceptional performance across a wide range of tasks. However, dense models usually suffer from sparse activation, where many activation values tend towards zero (i.e., being inactivated). We argue that this could restrict the efficient exploration of model representation space. To mitigate this issue, we propose Finedeep, a deep-layered fine-grained expert architecture for dense models. Our framework partitions the feed-forward neural network layers of traditional dense models into small experts, arranges them across multiple sub-layers. A novel routing mechanism is proposed to determine each expert's contribution. We conduct extensive experiments across various model sizes, demonstrating that our approach significantly outperforms traditional dense architectures in terms of perplexity and benchmark performance while maintaining a comparable number of parameters and floating-point operations. Moreover, we find that Finedeep achieves optimal results when balancing depth and width, specifically by adjusting the number of expert sub-layers and the number of experts per sub-layer. Empirical results confirm that Finedeep effectively alleviates sparse activation and efficiently utilizes representation capacity in dense models. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12928v1-abstract-full').style.display = 'none'; document.getElementById('2502.12928v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.12574">arXiv:2502.12574</a> <span> [<a href="https://arxiv.org/pdf/2502.12574">pdf</a>, <a href="https://arxiv.org/format/2502.12574">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"> HeadInfer: Memory-Efficient LLM Inference by Head-wise Offloading </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+Z">Zefan Cai</a>, <a href="/search/cs?searchtype=author&query=Sun%2C+H">Hanshi Sun</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+J">Jinqi Xiao</a>, <a href="/search/cs?searchtype=author&query=Yuan%2C+B">Bo Yuan</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+W">Wen Xiao</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Junjie Hu</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+J">Jiawei Zhao</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+B">Beidi Chen</a>, <a href="/search/cs?searchtype=author&query=Anandkumar%2C+A">Anima Anandkumar</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="2502.12574v1-abstract-short" style="display: inline;"> Transformer-based large language models (LLMs) demonstrate impressive performance in long context generation. Extending the context length has disproportionately shifted the memory footprint of LLMs during inference to the key-value cache (KV cache). In this paper, we propose HEADINFER, which offloads the KV cache to CPU RAM while avoiding the need to fully store the KV cache for any transformer l… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12574v1-abstract-full').style.display = 'inline'; document.getElementById('2502.12574v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.12574v1-abstract-full" style="display: none;"> Transformer-based large language models (LLMs) demonstrate impressive performance in long context generation. Extending the context length has disproportionately shifted the memory footprint of LLMs during inference to the key-value cache (KV cache). In this paper, we propose HEADINFER, which offloads the KV cache to CPU RAM while avoiding the need to fully store the KV cache for any transformer layer on the GPU. HEADINFER employs a fine-grained, head-wise offloading strategy, maintaining only selective attention heads KV cache on the GPU while computing attention output dynamically. Through roofline analysis, we demonstrate that HEADINFER maintains computational efficiency while significantly reducing memory footprint. We evaluate HEADINFER on the Llama-3-8B model with a 1-million-token sequence, reducing the GPU memory footprint of the KV cache from 128 GB to 1 GB and the total GPU memory usage from 207 GB to 17 GB, achieving a 92% reduction compared to BF16 baseline inference. Notably, HEADINFER enables 4-million-token inference with an 8B model on a single consumer GPU with 24GB memory (e.g., NVIDIA RTX 4090) without approximation methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12574v1-abstract-full').style.display = 'none'; document.getElementById('2502.12574v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.12455">arXiv:2502.12455</a> <span> [<a href="https://arxiv.org/pdf/2502.12455">pdf</a>, <a href="https://arxiv.org/format/2502.12455">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> DSMoE: Matrix-Partitioned Experts with Dynamic Routing for Computation-Efficient Dense LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lv%2C+M">Minxuan Lv</a>, <a href="/search/cs?searchtype=author&query=Su%2C+Z">Zhenpeng Su</a>, <a href="/search/cs?searchtype=author&query=Pan%2C+L">Leiyu Pan</a>, <a href="/search/cs?searchtype=author&query=Xiong%2C+Y">Yizhe Xiong</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+Z">Zijia Lin</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+H">Hui Chen</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+W">Wei Zhou</a>, <a href="/search/cs?searchtype=author&query=Han%2C+J">Jungong Han</a>, <a href="/search/cs?searchtype=author&query=Ding%2C+G">Guiguang Ding</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+D">Di Zhang</a>, <a href="/search/cs?searchtype=author&query=Gai%2C+K">Kun Gai</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+S">Songlin Hu</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="2502.12455v2-abstract-short" style="display: inline;"> As large language models continue to scale, computational costs and resource consumption have emerged as significant challenges. While existing sparsification methods like pruning reduce computational overhead, they risk losing model knowledge through parameter removal. This paper proposes DSMoE (Dynamic Sparse Mixture-of-Experts), a novel approach that achieves sparsification by partitioning pre-… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12455v2-abstract-full').style.display = 'inline'; document.getElementById('2502.12455v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.12455v2-abstract-full" style="display: none;"> As large language models continue to scale, computational costs and resource consumption have emerged as significant challenges. While existing sparsification methods like pruning reduce computational overhead, they risk losing model knowledge through parameter removal. This paper proposes DSMoE (Dynamic Sparse Mixture-of-Experts), a novel approach that achieves sparsification by partitioning pre-trained FFN layers into computational blocks. We implement adaptive expert routing using sigmoid activation and straight-through estimators, enabling tokens to flexibly access different aspects of model knowledge based on input complexity. Additionally, we introduce a sparsity loss term to balance performance and computational efficiency. Extensive experiments on LLaMA models demonstrate that under equivalent computational constraints, DSMoE achieves superior performance compared to existing pruning and MoE approaches across language modeling and downstream tasks, particularly excelling in generation tasks. Analysis reveals that DSMoE learns distinctive layerwise activation patterns, providing new insights for future MoE architecture design. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12455v2-abstract-full').style.display = 'none'; document.getElementById('2502.12455v2-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 17 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.11168">arXiv:2502.11168</a> <span> [<a href="https://arxiv.org/pdf/2502.11168">pdf</a>, <a href="https://arxiv.org/format/2502.11168">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"> Knowing Your Target: Target-Aware Transformer Makes Better Spatio-Temporal Video Grounding </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Gu%2C+X">Xin Gu</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+Y">Yaojie Shen</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chenxi Luo</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+T">Tiejian Luo</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+Y">Yan Huang</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+Y">Yuewei Lin</a>, <a href="/search/cs?searchtype=author&query=Fan%2C+H">Heng Fan</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+L">Libo Zhang</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="2502.11168v1-abstract-short" style="display: inline;"> Transformer has attracted increasing interest in STVG, owing to its end-to-end pipeline and promising result. Existing Transformer-based STVG approaches often leverage a set of object queries, which are initialized simply using zeros and then gradually learn target position information via iterative interactions with multimodal features, for spatial and temporal localization. Despite simplicity, t… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11168v1-abstract-full').style.display = 'inline'; document.getElementById('2502.11168v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.11168v1-abstract-full" style="display: none;"> Transformer has attracted increasing interest in STVG, owing to its end-to-end pipeline and promising result. Existing Transformer-based STVG approaches often leverage a set of object queries, which are initialized simply using zeros and then gradually learn target position information via iterative interactions with multimodal features, for spatial and temporal localization. Despite simplicity, these zero object queries, due to lacking target-specific cues, are hard to learn discriminative target information from interactions with multimodal features in complicated scenarios (\e.g., with distractors or occlusion), resulting in degradation. Addressing this, we introduce a novel Target-Aware Transformer for STVG (TA-STVG), which seeks to adaptively generate object queries via exploring target-specific cues from the given video-text pair, for improving STVG. The key lies in two simple yet effective modules, comprising text-guided temporal sampling (TTS) and attribute-aware spatial activation (ASA), working in a cascade. The former focuses on selecting target-relevant temporal cues from a video utilizing holistic text information, while the latter aims at further exploiting the fine-grained visual attribute information of the object from previous target-aware temporal cues, which is applied for object query initialization. Compared to existing methods leveraging zero-initialized queries, object queries in our TA-STVG, directly generated from a given video-text pair, naturally carry target-specific cues, making them adaptive and better interact with multimodal features for learning more discriminative information to improve STVG. In our experiments on three benchmarks, TA-STVG achieves state-of-the-art performance and significantly outperforms the baseline, validating its efficacy. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11168v1-abstract-full').style.display = 'none'; document.getElementById('2502.11168v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.09888">arXiv:2502.09888</a> <span> [<a href="https://arxiv.org/pdf/2502.09888">pdf</a>, <a href="https://arxiv.org/format/2502.09888">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Information Retrieval">cs.IR</span> </div> </div> <p class="title is-5 mathjax"> An Efficient Large Recommendation Model: Towards a Resource-Optimal Scaling Law </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xu%2C+S">Songpei Xu</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+S">Shijia Wang</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+D">Da Guo</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+X">Xianwen Guo</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+Q">Qiang Xiao</a>, <a href="/search/cs?searchtype=author&query=Li%2C+F">Fangjian Li</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuanjiang Luo</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="2502.09888v1-abstract-short" style="display: inline;"> The pursuit of scaling up recommendation models confronts intrinsic tensions between expanding model capacity and preserving computational tractability. While prior studies have explored scaling laws for recommendation systems, their resource-intensive paradigms -- often requiring tens of thousands of A100 GPU hours -- remain impractical for most industrial applications. This work addresses a crit… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.09888v1-abstract-full').style.display = 'inline'; document.getElementById('2502.09888v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.09888v1-abstract-full" style="display: none;"> The pursuit of scaling up recommendation models confronts intrinsic tensions between expanding model capacity and preserving computational tractability. While prior studies have explored scaling laws for recommendation systems, their resource-intensive paradigms -- often requiring tens of thousands of A100 GPU hours -- remain impractical for most industrial applications. This work addresses a critical gap: achieving sustainable model scaling under strict computational budgets. We propose Climber, a resource-efficient recommendation framework comprising two synergistic components: the ASTRO model architecture for algorithmic innovation and the TURBO acceleration framework for engineering optimization. ASTRO (Adaptive Scalable Transformer for RecOmmendation) adopts two core innovations: (1) multi-scale sequence partitioning that reduces attention complexity from O(n^2d) to O(n^2d/Nb) via hierarchical blocks, enabling more efficient scaling with sequence length; (2) dynamic temperature modulation that adaptively adjusts attention scores for multimodal distributions arising from inherent multi-scenario and multi-behavior interactions. Complemented by TURBO (Two-stage Unified Ranking with Batched Output), a co-designed acceleration framework integrating gradient-aware feature compression and memory-efficient Key-Value caching, Climber achieves 5.15x throughput gains without performance degradation. Comprehensive offline experiments on multiple datasets validate that Climber exhibits a more ideal scaling curve. To our knowledge, this is the first publicly documented framework where controlled model scaling drives continuous online metric growth (12.19% overall lift) without prohibitive resource costs. Climber has been successfully deployed on Netease Cloud Music, one of China's largest music streaming platforms, serving tens of millions of users daily. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.09888v1-abstract-full').style.display = 'none'; document.getElementById('2502.09888v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.08244">arXiv:2502.08244</a> <span> [<a href="https://arxiv.org/pdf/2502.08244">pdf</a>, <a href="https://arxiv.org/format/2502.08244">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"> FloVD: Optical Flow Meets Video Diffusion Model for Enhanced Camera-Controlled Video Synthesis </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Jin%2C+W">Wonjoon Jin</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Baek%2C+S">Seung-Hwan Baek</a>, <a href="/search/cs?searchtype=author&query=Cho%2C+S">Sunghyun Cho</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="2502.08244v2-abstract-short" style="display: inline;"> We present FloVD, a novel video diffusion model for camera-controllable video generation. FloVD leverages optical flow to represent the motions of the camera and moving objects. This approach offers two key benefits. Since optical flow can be directly estimated from videos, our approach allows for the use of arbitrary training videos without ground-truth camera parameters. Moreover, as background… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.08244v2-abstract-full').style.display = 'inline'; document.getElementById('2502.08244v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.08244v2-abstract-full" style="display: none;"> We present FloVD, a novel video diffusion model for camera-controllable video generation. FloVD leverages optical flow to represent the motions of the camera and moving objects. This approach offers two key benefits. Since optical flow can be directly estimated from videos, our approach allows for the use of arbitrary training videos without ground-truth camera parameters. Moreover, as background optical flow encodes 3D correlation across different viewpoints, our method enables detailed camera control by leveraging the background motion. To synthesize natural object motion while supporting detailed camera control, our framework adopts a two-stage video synthesis pipeline consisting of optical flow generation and flow-conditioned video synthesis. Extensive experiments demonstrate the superiority of our method over previous approaches in terms of accurate camera control and natural object motion synthesis. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.08244v2-abstract-full').style.display = 'none'; document.getElementById('2502.08244v2-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 12 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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">Our paper has been accepted to CVPR 2025. Website: https://jinwonjoon.github.io/flovd_site/ Code: https://github.com/JinWonjoon/FloVD</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.04531">arXiv:2502.04531</a> <span> [<a href="https://arxiv.org/pdf/2502.04531">pdf</a>, <a href="https://arxiv.org/format/2502.04531">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> <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"> AnyPlace: Learning Generalized Object Placement for Robot Manipulation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhao%2C+Y">Yuchi Zhao</a>, <a href="/search/cs?searchtype=author&query=Bogdanovic%2C+M">Miroslav Bogdanovic</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chengyuan Luo</a>, <a href="/search/cs?searchtype=author&query=Tohme%2C+S">Steven Tohme</a>, <a href="/search/cs?searchtype=author&query=Darvish%2C+K">Kourosh Darvish</a>, <a href="/search/cs?searchtype=author&query=Aspuru-Guzik%2C+A">Al谩n Aspuru-Guzik</a>, <a href="/search/cs?searchtype=author&query=Shkurti%2C+F">Florian Shkurti</a>, <a href="/search/cs?searchtype=author&query=Garg%2C+A">Animesh Garg</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="2502.04531v1-abstract-short" style="display: inline;"> Object placement in robotic tasks is inherently challenging due to the diversity of object geometries and placement configurations. To address this, we propose AnyPlace, a two-stage method trained entirely on synthetic data, capable of predicting a wide range of feasible placement poses for real-world tasks. Our key insight is that by leveraging a Vision-Language Model (VLM) to identify rough plac… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.04531v1-abstract-full').style.display = 'inline'; document.getElementById('2502.04531v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.04531v1-abstract-full" style="display: none;"> Object placement in robotic tasks is inherently challenging due to the diversity of object geometries and placement configurations. To address this, we propose AnyPlace, a two-stage method trained entirely on synthetic data, capable of predicting a wide range of feasible placement poses for real-world tasks. Our key insight is that by leveraging a Vision-Language Model (VLM) to identify rough placement locations, we focus only on the relevant regions for local placement, which enables us to train the low-level placement-pose-prediction model to capture diverse placements efficiently. For training, we generate a fully synthetic dataset of randomly generated objects in different placement configurations (insertion, stacking, hanging) and train local placement-prediction models. We conduct extensive evaluations in simulation, demonstrating that our method outperforms baselines in terms of success rate, coverage of possible placement modes, and precision. In real-world experiments, we show how our approach directly transfers models trained purely on synthetic data to the real world, where it successfully performs placements in scenarios where other models struggle -- such as with varying object geometries, diverse placement modes, and achieving high precision for fine placement. More at: https://any-place.github.io. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.04531v1-abstract-full').style.display = 'none'; document.getElementById('2502.04531v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.03855">arXiv:2502.03855</a> <span> [<a href="https://arxiv.org/pdf/2502.03855">pdf</a>, <a href="https://arxiv.org/format/2502.03855">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"> Semi-rPPG: Semi-Supervised Remote Physiological Measurement with Curriculum Pseudo-Labeling </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wu%2C+B">Bingjie Wu</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+Z">Zitong Yu</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+Y">Yiping Xie</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+W">Wei Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chaoqi Luo</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Y">Yong Liu</a>, <a href="/search/cs?searchtype=author&query=Goh%2C+R+S+M">Rick Siow Mong Goh</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="2502.03855v1-abstract-short" style="display: inline;"> Remote Photoplethysmography (rPPG) is a promising technique to monitor physiological signals such as heart rate from facial videos. However, the labeled facial videos in this research are challenging to collect. Current rPPG research is mainly based on several small public datasets collected in simple environments, which limits the generalization and scale of the AI models. Semi-supervised methods… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.03855v1-abstract-full').style.display = 'inline'; document.getElementById('2502.03855v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.03855v1-abstract-full" style="display: none;"> Remote Photoplethysmography (rPPG) is a promising technique to monitor physiological signals such as heart rate from facial videos. However, the labeled facial videos in this research are challenging to collect. Current rPPG research is mainly based on several small public datasets collected in simple environments, which limits the generalization and scale of the AI models. Semi-supervised methods that leverage a small amount of labeled data and abundant unlabeled data can fill this gap for rPPG learning. In this study, a novel semi-supervised learning method named Semi-rPPG that combines curriculum pseudo-labeling and consistency regularization is proposed to extract intrinsic physiological features from unlabelled data without impairing the model from noises. Specifically, a curriculum pseudo-labeling strategy with signal-to-noise ratio (SNR) criteria is proposed to annotate the unlabelled data while adaptively filtering out the low-quality unlabelled data. Besides, a novel consistency regularization term for quasi-periodic signals is proposed through weak and strong augmented clips. To benefit the research on semi-supervised rPPG measurement, we establish a novel semi-supervised benchmark for rPPG learning through intra-dataset and cross-dataset evaluation on four public datasets. The proposed Semi-rPPG method achieves the best results compared with three classical semi-supervised methods under different protocols. Ablation studies are conducted to prove the effectiveness of the proposed methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.03855v1-abstract-full').style.display = 'none'; document.getElementById('2502.03855v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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 IEEE Transactions on Instrumentation and Measurement (TIM)</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.03447">arXiv:2502.03447</a> <span> [<a href="https://arxiv.org/pdf/2502.03447">pdf</a>, <a href="https://arxiv.org/format/2502.03447">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Human-Computer Interaction">cs.HC</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/10.1145/3708359.3712142.">10.1145/3708359.3712142. <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Designing LLM-simulated Immersive Spaces to Enhance Autistic Children's Social Affordances Understanding </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Cao%2C+Y">Yancheng Cao</a>, <a href="/search/cs?searchtype=author&query=HE%2C+Y">Yangyang HE</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yonglin Chen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+M">Menghan Chen</a>, <a href="/search/cs?searchtype=author&query=You%2C+S">Shanhe You</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+Y">Yulin Qiu</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+M">Min Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuan Luo</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+C">Chen Zheng</a>, <a href="/search/cs?searchtype=author&query=Tong%2C+X">Xin Tong</a>, <a href="/search/cs?searchtype=author&query=Liang%2C+J">Jing Liang</a>, <a href="/search/cs?searchtype=author&query=Gong%2C+J">Jiangtao Gong</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="2502.03447v1-abstract-short" style="display: inline;"> One of the key challenges faced by autistic children is understanding social affordances in complex environments, which further impacts their ability to respond appropriately to social signals. In traffic scenarios, this impairment can even lead to safety concerns. In this paper, we introduce an LLM-simulated immersive projection environment designed to improve this ability in autistic children wh… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.03447v1-abstract-full').style.display = 'inline'; document.getElementById('2502.03447v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.03447v1-abstract-full" style="display: none;"> One of the key challenges faced by autistic children is understanding social affordances in complex environments, which further impacts their ability to respond appropriately to social signals. In traffic scenarios, this impairment can even lead to safety concerns. In this paper, we introduce an LLM-simulated immersive projection environment designed to improve this ability in autistic children while ensuring their safety. We first propose 17 design considerations across four major categories, derived from a comprehensive review of previous research. Next, we developed a system called AIroad, which leverages LLMs to simulate drivers with varying social intents, expressed through explicit multimodal social signals. AIroad helps autistic children bridge the gap in recognizing the intentions behind behaviors and learning appropriate responses through various stimuli. A user study involving 14 participants demonstrated that this technology effectively engages autistic children and leads to significant improvements in their comprehension of social affordances in traffic scenarios. Additionally, parents reported high perceived usability of the system. These findings highlight the potential of combining LLM technology with immersive environments for the functional rehabilitation of autistic children in the future. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.03447v1-abstract-full').style.display = 'none'; document.getElementById('2502.03447v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </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">iui2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.11599">arXiv:2501.11599</a> <span> [<a href="https://arxiv.org/pdf/2501.11599">pdf</a>, <a href="https://arxiv.org/format/2501.11599">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</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"> SR-FoT: A Syllogistic-Reasoning Framework of Thought for Large Language Models Tackling Knowledge-based Reasoning Tasks </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wan%2C+W">Wentao Wan</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Z">Zhuojie Yang</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yongcan Chen</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chenglin Luo</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+R">Ruilin Wang</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+K">Kehao Cai</a>, <a href="/search/cs?searchtype=author&query=Kang%2C+N">Nan Kang</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+L">Liang Lin</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+K">Keze Wang</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="2501.11599v1-abstract-short" style="display: inline;"> Deductive reasoning is a crucial logical capability that assists us in solving complex problems based on existing knowledge. Although augmented by Chain-of-Thought prompts, Large Language Models (LLMs) might not follow the correct reasoning paths. Enhancing the deductive reasoning abilities of LLMs, and leveraging their extensive built-in knowledge for various reasoning tasks, remains an open ques… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.11599v1-abstract-full').style.display = 'inline'; document.getElementById('2501.11599v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.11599v1-abstract-full" style="display: none;"> Deductive reasoning is a crucial logical capability that assists us in solving complex problems based on existing knowledge. Although augmented by Chain-of-Thought prompts, Large Language Models (LLMs) might not follow the correct reasoning paths. Enhancing the deductive reasoning abilities of LLMs, and leveraging their extensive built-in knowledge for various reasoning tasks, remains an open question. Attempting to mimic the human deductive reasoning paradigm, we propose a multi-stage Syllogistic-Reasoning Framework of Thought (SR-FoT) that enables LLMs to perform syllogistic deductive reasoning to handle complex knowledge-based reasoning tasks. Our SR-FoT begins by interpreting the question and then uses the interpretation and the original question to propose a suitable major premise. It proceeds by generating and answering minor premise questions in two stages to match the minor premises. Finally, it guides LLMs to use the previously generated major and minor premises to perform syllogistic deductive reasoning to derive the answer to the original question. Extensive and thorough experiments on knowledge-based reasoning tasks have demonstrated the effectiveness and advantages of our SR-FoT. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.11599v1-abstract-full').style.display = 'none'; document.getElementById('2501.11599v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </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">This paper has been accepted by 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/2501.09341">arXiv:2501.09341</a> <span> [<a href="https://arxiv.org/pdf/2501.09341">pdf</a>, <a href="https://arxiv.org/format/2501.09341">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"> SE-BSFV: Online Subspace Learning based Shadow Enhancement and Background Suppression for ViSAR under Complex Background </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Yan%2C+S">Shangqu Yan</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chenyang Luo</a>, <a href="/search/cs?searchtype=author&query=Fu%2C+Y">Yaowen Fu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+W">Wenpeng Zhang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+W">Wei Yang</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+R">Ruofeng Yu</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="2501.09341v1-abstract-short" style="display: inline;"> Video synthetic aperture radar (ViSAR) has attracted substantial attention in the moving target detection (MTD) field due to its ability to continuously monitor changes in the target area. In ViSAR, the moving targets' shadows will not offset and defocus, which is widely used as a feature for MTD. However, the shadows are difficult to distinguish from the low scattering region in the background, w… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.09341v1-abstract-full').style.display = 'inline'; document.getElementById('2501.09341v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.09341v1-abstract-full" style="display: none;"> Video synthetic aperture radar (ViSAR) has attracted substantial attention in the moving target detection (MTD) field due to its ability to continuously monitor changes in the target area. In ViSAR, the moving targets' shadows will not offset and defocus, which is widely used as a feature for MTD. However, the shadows are difficult to distinguish from the low scattering region in the background, which will cause more missing and false alarms. Therefore, it is worth investigating how to enhance the distinction between the shadows and background. In this study, we proposed the Shadow Enhancement and Background Suppression for ViSAR (SE-BSFV) algorithm. The SE-BSFV algorithm is based on the low-rank representation (LRR) theory and adopts online subspace learning technique to enhance shadows and suppress background for ViSAR images. Firstly, we use a registration algorithm to register the ViSAR images and utilize Gaussian mixture distribution (GMD) to model the ViSAR data. Secondly, the knowledge learned from the previous frames is leveraged to estimate the GMD parameters of the current frame, and the Expectation-maximization (EM) algorithm is used to estimate the subspace parameters. Then, the foreground matrix of the current frame can be obtained. Finally, the alternating direction method of multipliers (ADMM) is used to eliminate strong scattering objects in the foreground matrix to obtain the final results. The experimental results indicate that the SE-BSFV algorithm significantly enhances the shadows' saliency and greatly improves the detection performance while ensuring efficiency compared with several other advanced pre-processing algorithms. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.09341v1-abstract-full').style.display = 'none'; document.getElementById('2501.09341v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.08305">arXiv:2501.08305</a> <span> [<a href="https://arxiv.org/pdf/2501.08305">pdf</a>, <a href="https://arxiv.org/format/2501.08305">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"> Benchmarking Graph Representations and Graph Neural Networks for Multivariate Time Series Classification </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Yang%2C+W">Wennuo Yang</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+S">Shiling Wu</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Y">Yuzhi Zhou</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=He%2C+X">Xilin He</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+W">Weicheng Xie</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+L">Linlin Shen</a>, <a href="/search/cs?searchtype=author&query=Song%2C+S">Siyang Song</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="2501.08305v2-abstract-short" style="display: inline;"> Multivariate Time Series Classification (MTSC) enables the analysis if complex temporal data, and thus serves as a cornerstone in various real-world applications, ranging from healthcare to finance. Since the relationship among variables in MTS usually contain crucial cues, a large number of graph-based MTSC approaches have been proposed, as the graph topology and edges can explicitly represent re… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.08305v2-abstract-full').style.display = 'inline'; document.getElementById('2501.08305v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.08305v2-abstract-full" style="display: none;"> Multivariate Time Series Classification (MTSC) enables the analysis if complex temporal data, and thus serves as a cornerstone in various real-world applications, ranging from healthcare to finance. Since the relationship among variables in MTS usually contain crucial cues, a large number of graph-based MTSC approaches have been proposed, as the graph topology and edges can explicitly represent relationships among variables (channels), where not only various MTS graph representation learning strategies but also different Graph Neural Networks (GNNs) have been explored. Despite such progresses, there is no comprehensive study that fairly benchmarks and investigates the performances of existing widely-used graph representation learning strategies/GNN classifiers in the application of different MTSC tasks. In this paper, we present the first benchmark which systematically investigates the effectiveness of the widely-used three node feature definition strategies, four edge feature learning strategies and five GNN architecture, resulting in 60 different variants for graph-based MTSC. These variants are developed and evaluated with a standardized data pipeline and training/validation/testing strategy on 26 widely-used suspensor MTSC datasets. Our experiments highlight that node features significantly influence MTSC performance, while the visualization of edge features illustrates why adaptive edge learning outperforms other edge feature learning methods. The code of the proposed benchmark is publicly available at \url{https://github.com/CVI-yangwn/Benchmark-GNN-for-Multivariate-Time-Series-Classification}. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.08305v2-abstract-full').style.display = 'none'; document.getElementById('2501.08305v2-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> 17 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 14 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">MSC Class:</span> 68T10 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.07845">arXiv:2501.07845</a> <span> [<a href="https://arxiv.org/pdf/2501.07845">pdf</a>, <a href="https://arxiv.org/format/2501.07845">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> Reasoning with Graphs: Structuring Implicit Knowledge to Enhance LLMs Reasoning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Han%2C+H">Haoyu Han</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+Y">Yaochen Xie</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Nag%2C+S">Sreyashi Nag</a>, <a href="/search/cs?searchtype=author&query=Headden%2C+W">William Headden</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yang Li</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Ji%2C+S">Shuiwang Ji</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+J">Jiliang Tang</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="2501.07845v1-abstract-short" style="display: inline;"> Large language models (LLMs) have demonstrated remarkable success across a wide range of tasks; however, they still encounter challenges in reasoning tasks that require understanding and inferring relationships between distinct pieces of information within text sequences. This challenge is particularly pronounced in tasks involving multi-step processes, such as logical reasoning and multi-hop ques… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.07845v1-abstract-full').style.display = 'inline'; document.getElementById('2501.07845v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.07845v1-abstract-full" style="display: none;"> Large language models (LLMs) have demonstrated remarkable success across a wide range of tasks; however, they still encounter challenges in reasoning tasks that require understanding and inferring relationships between distinct pieces of information within text sequences. This challenge is particularly pronounced in tasks involving multi-step processes, such as logical reasoning and multi-hop question answering, where understanding implicit relationships between entities and leveraging multi-hop connections in the given context are crucial. Graphs, as fundamental data structures, explicitly represent pairwise relationships between entities, thereby offering the potential to enhance LLMs' reasoning capabilities. External graphs have proven effective in supporting LLMs across multiple tasks. However, in many reasoning tasks, no pre-existing graph structure is provided. Can we structure implicit knowledge derived from context into graphs to assist LLMs in reasoning? In this paper, we propose Reasoning with Graphs (RwG) by first constructing explicit graphs from the context and then leveraging these graphs to enhance LLM reasoning performance on reasoning tasks. Extensive experiments demonstrate the effectiveness of the proposed method in improving both logical reasoning and multi-hop question answering tasks. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.07845v1-abstract-full').style.display = 'none'; document.getElementById('2501.07845v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.07837">arXiv:2501.07837</a> <span> [<a href="https://arxiv.org/pdf/2501.07837">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> A Driver Advisory System Based on Large Language Model for High-speed Train </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Luo%2C+Y+C">Y. C. Luo</a>, <a href="/search/cs?searchtype=author&query=Xun%2C+J">J. Xun</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+W">W. Wang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+R+Z">R. Z. Zhang</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+Z+C">Z. C. Zhao</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="2501.07837v1-abstract-short" style="display: inline;"> With the rapid development of China high-speed railway, drivers face increasingly significant technical challenges during operations, such as fault handling. Currently, drivers depend on the onboard mechanic when facing technical issues, for instance, traction loss or sensor faults. This dependency can hinder effective operation, even lead to accidents, while waiting for faults to be addressed. To… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.07837v1-abstract-full').style.display = 'inline'; document.getElementById('2501.07837v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.07837v1-abstract-full" style="display: none;"> With the rapid development of China high-speed railway, drivers face increasingly significant technical challenges during operations, such as fault handling. Currently, drivers depend on the onboard mechanic when facing technical issues, for instance, traction loss or sensor faults. This dependency can hinder effective operation, even lead to accidents, while waiting for faults to be addressed. To enhance the accuracy and explainability of actions during fault handling, an Intelligent Driver Advisory System (IDAS) framework based on a large language model (LLM) named IDAS-LLM, is introduced. Initially, domain-fine-tuning of the LLM is performed using a constructed railway knowledge question-and-answer dataset to improve answer accuracy in railway-related questions. Subsequently, integration of the Retrieval-augmented Generation (RAG) architecture is pursued for system design to enhance the explainability of generated responses. Comparative experiments are conducted using the constructed railway driving knowledge assessment dataset. Results indicate that domain-fine-tuned LLMs show an improvement in answer accuracy by an average of 10%, outperforming some current mainstream LLMs. Additionally, the inclusion of the RAG framework increases the average recall rate of question-and-answer sessions by about 4%. Finally, the fault handling capability of IDAS-LLM is demonstrated through simulations of real operational scenarios, proving that the proposed framework has practical application prospects. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.07837v1-abstract-full').style.display = 'none'; document.getElementById('2501.07837v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </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, 7 figures, presented at 104th TRB Annual Meeting</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.03257">arXiv:2501.03257</a> <span> [<a href="https://arxiv.org/pdf/2501.03257">pdf</a>, <a href="https://arxiv.org/format/2501.03257">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="Computation and Language">cs.CL</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"> Breaking Through the Spike: Spike Window Decoding for Accelerated and Precise Automatic Speech Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+W">Wei Zhang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+T">Tian-Hao Zhang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chao Luo</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+H">Hui Zhou</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+C">Chao Yang</a>, <a href="/search/cs?searchtype=author&query=Qian%2C+X">Xinyuan Qian</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+X">Xu-Cheng Yin</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="2501.03257v1-abstract-short" style="display: inline;"> Recently, end-to-end automatic speech recognition has become the mainstream approach in both industry and academia. To optimize system performance in specific scenarios, the Weighted Finite-State Transducer (WFST) is extensively used to integrate acoustic and language models, leveraging its capacity to implicitly fuse language models within static graphs, thereby ensuring robust recognition while… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.03257v1-abstract-full').style.display = 'inline'; document.getElementById('2501.03257v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.03257v1-abstract-full" style="display: none;"> Recently, end-to-end automatic speech recognition has become the mainstream approach in both industry and academia. To optimize system performance in specific scenarios, the Weighted Finite-State Transducer (WFST) is extensively used to integrate acoustic and language models, leveraging its capacity to implicitly fuse language models within static graphs, thereby ensuring robust recognition while also facilitating rapid error correction. However, WFST necessitates a frame-by-frame search of CTC posterior probabilities through autoregression, which significantly hampers inference speed. In this work, we thoroughly investigate the spike property of CTC outputs and further propose the conjecture that adjacent frames to non-blank spikes carry semantic information beneficial to the model. Building on this, we propose the Spike Window Decoding algorithm, which greatly improves the inference speed by making the number of frames decoded in WFST linearly related to the number of spiking frames in the CTC output, while guaranteeing the recognition performance. Our method achieves SOTA recognition accuracy with significantly accelerates decoding speed, proven across both AISHELL-1 and large-scale In-House datasets, establishing a pioneering approach for integrating CTC output with WFST. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.03257v1-abstract-full').style.display = 'none'; document.getElementById('2501.03257v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </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 ICASSP 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/2501.02379">arXiv:2501.02379</a> <span> [<a href="https://arxiv.org/pdf/2501.02379">pdf</a>, <a href="https://arxiv.org/format/2501.02379">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"> Tensor-GaLore: Memory-Efficient Training via Gradient Tensor Decomposition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=George%2C+R+J">Robert Joseph George</a>, <a href="/search/cs?searchtype=author&query=Pitt%2C+D">David Pitt</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+J">Jiawei Zhao</a>, <a href="/search/cs?searchtype=author&query=Kossaifi%2C+J">Jean Kossaifi</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Tian%2C+Y">Yuandong Tian</a>, <a href="/search/cs?searchtype=author&query=Anandkumar%2C+A">Anima Anandkumar</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="2501.02379v1-abstract-short" style="display: inline;"> We present Tensor-GaLore, a novel method for efficient training of neural networks with higher-order tensor weights. Many models, particularly those used in scientific computing, employ tensor-parameterized layers to capture complex, multidimensional relationships. When scaling these methods to high-resolution problems makes memory usage grow intractably, and matrix based optimization methods lead… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.02379v1-abstract-full').style.display = 'inline'; document.getElementById('2501.02379v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.02379v1-abstract-full" style="display: none;"> We present Tensor-GaLore, a novel method for efficient training of neural networks with higher-order tensor weights. Many models, particularly those used in scientific computing, employ tensor-parameterized layers to capture complex, multidimensional relationships. When scaling these methods to high-resolution problems makes memory usage grow intractably, and matrix based optimization methods lead to suboptimal performance and compression. We propose to work directly in the high-order space of the complex tensor parameter space using a tensor factorization of the gradients during optimization. We showcase its effectiveness on Fourier Neural Operators (FNOs), a class of models crucial for solving partial differential equations (PDE) and prove the theory of it. Across various PDE tasks like the Navier Stokes and Darcy Flow equations, Tensor-GaLore achieves substantial memory savings, reducing optimizer memory usage by up to 75%. These substantial memory savings across AI for science demonstrate Tensor-GaLore's potential. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.02379v1-abstract-full').style.display = 'none'; document.getElementById('2501.02379v1-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, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.16443">arXiv:2412.16443</a> <span> [<a href="https://arxiv.org/pdf/2412.16443">pdf</a>, <a href="https://arxiv.org/ps/2412.16443">ps</a>, <a href="https://arxiv.org/format/2412.16443">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"> Has LLM Reached the Scaling Ceiling Yet? Unified Insights into LLM Regularities and Constraints </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Luo%2C+C">Charles Luo</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="2412.16443v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their scalability raises a critical question: Have we reached the scaling ceiling? This paper addresses this pivotal question by developing a unified theoretical framework that integrates mathematical and statistical insights to explain the scaling dynamics of LLMs. We present: 1. Central Limit Theorem (CLT) for Hidden Rep… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16443v1-abstract-full').style.display = 'inline'; document.getElementById('2412.16443v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.16443v1-abstract-full" style="display: none;"> Large Language Models (LLMs) have demonstrated remarkable capabilities, yet their scalability raises a critical question: Have we reached the scaling ceiling? This paper addresses this pivotal question by developing a unified theoretical framework that integrates mathematical and statistical insights to explain the scaling dynamics of LLMs. We present: 1. Central Limit Theorem (CLT) for Hidden Representations: We show that noise in hidden representations scales inversely with context size, explaining stabilization effects and the limits of context length improvements. 2. Bias-Variance Decomposition: We decompose next-token prediction loss into irreducible entropy, capacity-driven bias, and finite sample variance, revealing trade-offs where scaling yields diminishing returns. 3. Emergent SNR Thresholds: By defining signal-to-noise ratio (SNR), we quantify how capabilities emerge abruptly once SNR surpasses a threshold, offering insights into when scaling becomes less effective. Through this framework, we conclude that while LLMs have not reached an absolute scaling ceiling, practical constraints are increasingly prominent: diminishing returns, resource inefficiencies, and data limitations. Future progress will require a shift from brute-force scaling to innovations in architecture, data quality, and training paradigms. This work provides a roadmap for guiding the efficient development of next-generation LLMs and advancing the field beyond traditional scaling strategies. Keywords: Large Language Models; Scaling Ceiling; Central Limit Theorem; Bias-Variance Trade-Off; Signal-to-Noise Ratio; Emergent Capabilities <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16443v1-abstract-full').style.display = 'none'; document.getElementById('2412.16443v1-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, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.12767">arXiv:2412.12767</a> <span> [<a href="https://arxiv.org/pdf/2412.12767">pdf</a>, <a href="https://arxiv.org/format/2412.12767">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</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"> A Survey of Calibration Process for Black-Box LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xie%2C+L">Liangru Xie</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Zeng%2C+J">Jingying Zeng</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Han%2C+Y">Yan Han</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+J">Jing Huang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zhen Li</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+S">Suhang Wang</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</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="2412.12767v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) demonstrate remarkable performance in semantic understanding and generation, yet accurately assessing their output reliability remains a significant challenge. While numerous studies have explored calibration techniques, they primarily focus on White-Box LLMs with accessible parameters. Black-Box LLMs, despite their superior performance, pose heightened requirements fo… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.12767v1-abstract-full').style.display = 'inline'; document.getElementById('2412.12767v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.12767v1-abstract-full" style="display: none;"> Large Language Models (LLMs) demonstrate remarkable performance in semantic understanding and generation, yet accurately assessing their output reliability remains a significant challenge. While numerous studies have explored calibration techniques, they primarily focus on White-Box LLMs with accessible parameters. Black-Box LLMs, despite their superior performance, pose heightened requirements for calibration techniques due to their API-only interaction constraints. Although recent researches have achieved breakthroughs in black-box LLMs calibration, a systematic survey of these methodologies is still lacking. To bridge this gap, we presents the first comprehensive survey on calibration techniques for black-box LLMs. We first define the Calibration Process of LLMs as comprising two interrelated key steps: Confidence Estimation and Calibration. Second, we conduct a systematic review of applicable methods within black-box settings, and provide insights on the unique challenges and connections in implementing these key steps. Furthermore, we explore typical applications of Calibration Process in black-box LLMs and outline promising future research directions, providing new perspectives for enhancing reliability and human-machine alignment. This is our GitHub link: https://github.com/LiangruXie/Calibration-Process-in-Black-Box-LLMs <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.12767v1-abstract-full').style.display = 'none'; document.getElementById('2412.12767v1-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> 17 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.11500">arXiv:2412.11500</a> <span> [<a href="https://arxiv.org/pdf/2412.11500">pdf</a>, <a href="https://arxiv.org/format/2412.11500">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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"> Intention Knowledge Graph Construction for User Intention Relation Modeling </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Bai%2C+J">Jiaxin Bai</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Z">Zhaobo Wang</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+J">Junfei Cheng</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+D">Dan Yu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+Z">Zerui Huang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+W">Weiqi Wang</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+X">Xin Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+Y">Yanming Zhu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+B">Bo Li</a>, <a href="/search/cs?searchtype=author&query=Song%2C+Y">Yangqiu Song</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="2412.11500v1-abstract-short" style="display: inline;"> Understanding user intentions is challenging for online platforms. Recent work on intention knowledge graphs addresses this but often lacks focus on connecting intentions, which is crucial for modeling user behavior and predicting future actions. This paper introduces a framework to automatically generate an intention knowledge graph, capturing connections between user intentions. Using the Amazon… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.11500v1-abstract-full').style.display = 'inline'; document.getElementById('2412.11500v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.11500v1-abstract-full" style="display: none;"> Understanding user intentions is challenging for online platforms. Recent work on intention knowledge graphs addresses this but often lacks focus on connecting intentions, which is crucial for modeling user behavior and predicting future actions. This paper introduces a framework to automatically generate an intention knowledge graph, capturing connections between user intentions. Using the Amazon m2 dataset, we construct an intention graph with 351 million edges, demonstrating high plausibility and acceptance. Our model effectively predicts new session intentions and enhances product recommendations, outperforming previous state-of-the-art methods and showcasing the approach's practical utility. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.11500v1-abstract-full').style.display = 'none'; document.getElementById('2412.11500v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.08109">arXiv:2412.08109</a> <span> [<a href="https://arxiv.org/pdf/2412.08109">pdf</a>, <a href="https://arxiv.org/format/2412.08109">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Software Engineering">cs.SE</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"> Unseen Horizons: Unveiling the Real Capability of LLM Code Generation Beyond the Familiar </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+Y">Yuanliang Zhang</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+Y">Yifan Xie</a>, <a href="/search/cs?searchtype=author&query=Li%2C+S">Shanshan Li</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+K">Ke Liu</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+C">Chong Wang</a>, <a href="/search/cs?searchtype=author&query=Jia%2C+Z">Zhouyang Jia</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xiangbing Huang</a>, <a href="/search/cs?searchtype=author&query=Song%2C+J">Jie Song</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chaopeng Luo</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+Z">Zhizheng Zheng</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+R">Rulin Xu</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Y">Yitong Liu</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+S">Si Zheng</a>, <a href="/search/cs?searchtype=author&query=Liao%2C+X">Xiangke Liao</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="2412.08109v2-abstract-short" style="display: inline;"> Recently, large language models (LLMs) have shown strong potential in code generation tasks. However, there are still gaps before they can be fully applied in actual software development processes. Accurately assessing the code generation capabilities of large language models has become an important basis for evaluating and improving the models. Some existing works have constructed datasets to eva… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.08109v2-abstract-full').style.display = 'inline'; document.getElementById('2412.08109v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.08109v2-abstract-full" style="display: none;"> Recently, large language models (LLMs) have shown strong potential in code generation tasks. However, there are still gaps before they can be fully applied in actual software development processes. Accurately assessing the code generation capabilities of large language models has become an important basis for evaluating and improving the models. Some existing works have constructed datasets to evaluate the capabilities of these models. However, the current evaluation process may encounter the illusion of "Specialist in Familiarity", primarily due to three gaps: the exposure of target code, case timeliness, and dependency availability. The fundamental reason for these gaps is that the code in current datasets may have been extensively exposed and exercised during the training phase, and due to the continuous training and development of LLM, their timeliness has been severely compromised. The key to solve the problem is to, as much as possible, evaluate the LLMs using code that they have not encountered before. Thus, the fundamental idea in this paper is to draw on the concept of code obfuscation, changing code at different levels while ensuring the functionality and output. To this end, we build a code-obfuscation based benchmark OBFUSEVAL. We first collect 1,354 raw cases from five real-world projects, including function description and code. Then we use three-level strategy (symbol, structure and semantic) to obfuscate descriptions, code and context dependencies. We evaluate four LLMs on OBFU- SEVAL and compared the effectiveness of different obfuscation strategy. We use official test suites of these projects to evaluate the generated code. The results show that after obfuscation, the average decrease ratio of test pass rate can up to 62.5%. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.08109v2-abstract-full').style.display = 'none'; document.getElementById('2412.08109v2-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> 15 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 11 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 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 47th International Conference on Software Engineering (ICSE 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/2412.04531">arXiv:2412.04531</a> <span> [<a href="https://arxiv.org/pdf/2412.04531">pdf</a>, <a href="https://arxiv.org/format/2412.04531">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"> MageBench: Bridging Large Multimodal Models to Agents </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+M">Miaosen Zhang</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yifan Yang</a>, <a href="/search/cs?searchtype=author&query=Bao%2C+J">Jianmin Bao</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+D">Dongdong Chen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+K">Kai Qiu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Geng%2C+X">Xin Geng</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+B">Baining Guo</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="2412.04531v1-abstract-short" style="display: inline;"> LMMs have shown impressive visual understanding capabilities, with the potential to be applied in agents, which demand strong reasoning and planning abilities. Nevertheless, existing benchmarks mostly assess their reasoning abilities in language part, where the chain-of-thought is entirely composed of text.We consider the scenario where visual signals are continuously updated and required along th… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.04531v1-abstract-full').style.display = 'inline'; document.getElementById('2412.04531v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.04531v1-abstract-full" style="display: none;"> LMMs have shown impressive visual understanding capabilities, with the potential to be applied in agents, which demand strong reasoning and planning abilities. Nevertheless, existing benchmarks mostly assess their reasoning abilities in language part, where the chain-of-thought is entirely composed of text.We consider the scenario where visual signals are continuously updated and required along the decision making process. Such vision-in-the-chain reasoning paradigm is more aligned with the needs of multimodal agents, while being rarely evaluated. In this paper, we introduce MageBench, a reasoning capability oriented multimodal agent benchmark that, while having light-weight environments, poses significant reasoning challenges and holds substantial practical value. This benchmark currently includes three types of environments: WebUI, Sokoban, and Football, comprising a total of 483 different scenarios. It thoroughly validates the agent's knowledge and engineering capabilities, visual intelligence, and interaction skills. The results show that only a few product-level models are better than random acting, and all of them are far inferior to human-level. More specifically, we found current models severely lack the ability to modify their planning based on visual feedback, as well as visual imagination, interleaved image-text long context handling, and other abilities. We hope that our work will provide optimization directions for LMM from the perspective of being an agent. We release our code and data at https://github.com/microsoft/MageBench. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.04531v1-abstract-full').style.display = 'none'; document.getElementById('2412.04531v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 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">37 pages, 32 figures, github link: https://github.com/microsoft/MageBench</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.02110">arXiv:2412.02110</a> <span> [<a href="https://arxiv.org/pdf/2412.02110">pdf</a>, <a href="https://arxiv.org/format/2412.02110">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> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Operating Systems">cs.OS</span> </div> </div> <p class="title is-5 mathjax"> Retrofitting XoM for Stripped Binaries without Embedded Data Relocation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chenke Luo</a>, <a href="/search/cs?searchtype=author&query=Ming%2C+J">Jiang Ming</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+M">Mengfei Xie</a>, <a href="/search/cs?searchtype=author&query=Peng%2C+G">Guojun Peng</a>, <a href="/search/cs?searchtype=author&query=Fu%2C+J">Jianming Fu</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="2412.02110v2-abstract-short" style="display: inline;"> In this paper, we present PXoM, a practical technique to seamlessly retrofit XoM into stripped binaries on the x86-64 platform. As handling the mixture of code and data is a well-known challenge for XoM, most existing methods require the strict separation of code and data areas via either compile-time transformation or binary patching, so that the unreadable permission can be safely enforced at th… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.02110v2-abstract-full').style.display = 'inline'; document.getElementById('2412.02110v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.02110v2-abstract-full" style="display: none;"> In this paper, we present PXoM, a practical technique to seamlessly retrofit XoM into stripped binaries on the x86-64 platform. As handling the mixture of code and data is a well-known challenge for XoM, most existing methods require the strict separation of code and data areas via either compile-time transformation or binary patching, so that the unreadable permission can be safely enforced at the granularity of memory pages. In contrast to previous approaches, we provide a fine-grained memory permission control mechanism to restrict the read permission of code while allowing legitimate data reads within code pages. This novelty enables PXoM to harden stripped binaries but without resorting to error-prone embedded data relocation. We leverage Intel's hardware feature, Memory Protection Keys, to offer an efficient fine-grained permission control. We measure PXoM's performance with both micro- and macro-benchmarks, and it only introduces negligible runtime overhead. Our security evaluation shows that PXoM leaves adversaries with little wiggle room to harvest all of the required gadgets, suggesting PXoM is practical for real-world deployment. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.02110v2-abstract-full').style.display = 'none'; document.getElementById('2412.02110v2-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 2 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.18660">arXiv:2411.18660</a> <span> [<a href="https://arxiv.org/pdf/2411.18660">pdf</a>, <a href="https://arxiv.org/format/2411.18660">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"> OOD-HOI: Text-Driven 3D Whole-Body Human-Object Interactions Generation Beyond Training Domains </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+Y">Yixuan Zhang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+H">Hui Yang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuanchen Luo</a>, <a href="/search/cs?searchtype=author&query=Peng%2C+J">Junran Peng</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yuxi Wang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Z">Zhaoxiang Zhang</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="2411.18660v1-abstract-short" style="display: inline;"> Generating realistic 3D human-object interactions (HOIs) from text descriptions is a active research topic with potential applications in virtual and augmented reality, robotics, and animation. However, creating high-quality 3D HOIs remains challenging due to the lack of large-scale interaction data and the difficulty of ensuring physical plausibility, especially in out-of-domain (OOD) scenarios.… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.18660v1-abstract-full').style.display = 'inline'; document.getElementById('2411.18660v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.18660v1-abstract-full" style="display: none;"> Generating realistic 3D human-object interactions (HOIs) from text descriptions is a active research topic with potential applications in virtual and augmented reality, robotics, and animation. However, creating high-quality 3D HOIs remains challenging due to the lack of large-scale interaction data and the difficulty of ensuring physical plausibility, especially in out-of-domain (OOD) scenarios. Current methods tend to focus either on the body or the hands, which limits their ability to produce cohesive and realistic interactions. In this paper, we propose OOD-HOI, a text-driven framework for generating whole-body human-object interactions that generalize well to new objects and actions. Our approach integrates a dual-branch reciprocal diffusion model to synthesize initial interaction poses, a contact-guided interaction refiner to improve physical accuracy based on predicted contact areas, and a dynamic adaptation mechanism which includes semantic adjustment and geometry deformation to improve robustness. Experimental results demonstrate that our OOD-HOI could generate more realistic and physically plausible 3D interaction pose in OOD scenarios compared to existing methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.18660v1-abstract-full').style.display = 'none'; document.getElementById('2411.18660v1-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 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/2411.17697">arXiv:2411.17697</a> <span> [<a href="https://arxiv.org/pdf/2411.17697">pdf</a>, <a href="https://arxiv.org/format/2411.17697">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"> StableAnimator: High-Quality Identity-Preserving Human Image Animation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Tu%2C+S">Shuyuan Tu</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+Z">Zhen Xing</a>, <a href="/search/cs?searchtype=author&query=Han%2C+X">Xintong Han</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+Z">Zhi-Qi Cheng</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</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="2411.17697v2-abstract-short" style="display: inline;"> Current diffusion models for human image animation struggle to ensure identity (ID) consistency. This paper presents StableAnimator, the first end-to-end ID-preserving video diffusion framework, which synthesizes high-quality videos without any post-processing, conditioned on a reference image and a sequence of poses. Building upon a video diffusion model, StableAnimator contains carefully designe… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.17697v2-abstract-full').style.display = 'inline'; document.getElementById('2411.17697v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.17697v2-abstract-full" style="display: none;"> Current diffusion models for human image animation struggle to ensure identity (ID) consistency. This paper presents StableAnimator, the first end-to-end ID-preserving video diffusion framework, which synthesizes high-quality videos without any post-processing, conditioned on a reference image and a sequence of poses. Building upon a video diffusion model, StableAnimator contains carefully designed modules for both training and inference striving for identity consistency. In particular, StableAnimator begins by computing image and face embeddings with off-the-shelf extractors, respectively and face embeddings are further refined by interacting with image embeddings using a global content-aware Face Encoder. Then, StableAnimator introduces a novel distribution-aware ID Adapter that prevents interference caused by temporal layers while preserving ID via alignment. During inference, we propose a novel Hamilton-Jacobi-Bellman (HJB) equation-based optimization to further enhance the face quality. We demonstrate that solving the HJB equation can be integrated into the diffusion denoising process, and the resulting solution constrains the denoising path and thus benefits ID preservation. Experiments on multiple benchmarks show the effectiveness of StableAnimator both qualitatively and quantitatively. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.17697v2-abstract-full').style.display = 'none'; document.getElementById('2411.17697v2-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 26 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/2411.13552">arXiv:2411.13552</a> <span> [<a href="https://arxiv.org/pdf/2411.13552">pdf</a>, <a href="https://arxiv.org/format/2411.13552">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"> REDUCIO! Generating 1024$\times$1024 Video within 16 Seconds using Extremely Compressed Motion Latents </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Tian%2C+R">Rui Tian</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Bao%2C+J">Jianmin Bao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+K">Kai Qiu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yifan Yang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</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="2411.13552v2-abstract-short" style="display: inline;"> Commercial video generation models have exhibited realistic, high-fidelity results but are still restricted to limited access. One crucial obstacle for large-scale applications is the expensive training and inference cost. In this paper, we argue that videos contain much more redundant information than images, thus can be encoded by very few motion latents based on a content image. Towards this go… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.13552v2-abstract-full').style.display = 'inline'; document.getElementById('2411.13552v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.13552v2-abstract-full" style="display: none;"> Commercial video generation models have exhibited realistic, high-fidelity results but are still restricted to limited access. One crucial obstacle for large-scale applications is the expensive training and inference cost. In this paper, we argue that videos contain much more redundant information than images, thus can be encoded by very few motion latents based on a content image. Towards this goal, we design an image-conditioned VAE to encode a video to an extremely compressed motion latent space. This magic Reducio charm enables 64x reduction of latents compared to a common 2D VAE, without sacrificing the quality. Training diffusion models on such a compact representation easily allows for generating 1K resolution videos. We then adopt a two-stage video generation paradigm, which performs text-to-image and text-image-to-video sequentially. Extensive experiments show that our Reducio-DiT achieves strong performance in evaluation, though trained with limited GPU resources. More importantly, our method significantly boost the efficiency of video LDMs both in training and inference. We train Reducio-DiT in around 3.2K training hours in total and generate a 16-frame 1024*1024 video clip within 15.5 seconds on a single A100 GPU. Code released at https://github.com/microsoft/Reducio-VAE . <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.13552v2-abstract-full').style.display = 'none'; document.getElementById('2411.13552v2-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 20 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 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">Code available at https://github.com/microsoft/Reducio-VAE</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.09118">arXiv:2411.09118</a> <span> [<a href="https://arxiv.org/pdf/2411.09118">pdf</a>, <a href="https://arxiv.org/format/2411.09118">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Optimization and Control">math.OC</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"> FxTS-Net: Fixed-Time Stable Learning Framework for Neural ODEs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chaoyang Luo</a>, <a href="/search/cs?searchtype=author&query=Zou%2C+Y">Yan Zou</a>, <a href="/search/cs?searchtype=author&query=Li%2C+W">Wanying Li</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+N">Nanjing Huang</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="2411.09118v1-abstract-short" style="display: inline;"> Neural Ordinary Differential Equations (Neural ODEs), as a novel category of modeling big data methods, cleverly link traditional neural networks and dynamical systems. However, it is challenging to ensure the dynamics system reaches a correctly predicted state within a user-defined fixed time. To address this problem, we propose a new method for training Neural ODEs using fixed-time stability (Fx… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.09118v1-abstract-full').style.display = 'inline'; document.getElementById('2411.09118v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.09118v1-abstract-full" style="display: none;"> Neural Ordinary Differential Equations (Neural ODEs), as a novel category of modeling big data methods, cleverly link traditional neural networks and dynamical systems. However, it is challenging to ensure the dynamics system reaches a correctly predicted state within a user-defined fixed time. To address this problem, we propose a new method for training Neural ODEs using fixed-time stability (FxTS) Lyapunov conditions. Our framework, called FxTS-Net, is based on the novel FxTS loss (FxTS-Loss) designed on Lyapunov functions, which aims to encourage convergence to accurate predictions in a user-defined fixed time. We also provide an innovative approach for constructing Lyapunov functions to meet various tasks and network architecture requirements, achieved by leveraging supervised information during training. By developing a more precise time upper bound estimation for bounded non-vanishingly perturbed systems, we demonstrate that minimizing FxTS-Loss not only guarantees FxTS behavior of the dynamics but also input perturbation robustness. For optimising FxTS-Loss, we also propose a learning algorithm, in which the simulated perturbation sampling method can capture sample points in critical regions to approximate FxTS-Loss. Experimentally, we find that FxTS-Net provides better prediction performance and better robustness under input perturbation. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.09118v1-abstract-full').style.display = 'none'; document.getElementById('2411.09118v1-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 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/2411.07722">arXiv:2411.07722</a> <span> [<a href="https://arxiv.org/pdf/2411.07722">pdf</a>, <a href="https://arxiv.org/format/2411.07722">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Is Cognition consistent with Perception? Assessing and Mitigating Multimodal Knowledge Conflicts in Document Understanding </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Shao%2C+Z">Zirui Shao</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuwei Luo</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+Z">Zhaoqing Zhu</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+H">Hangdi Xing</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+Z">Zhi Yu</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+Q">Qi Zheng</a>, <a href="/search/cs?searchtype=author&query=Bu%2C+J">Jiajun Bu</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="2411.07722v1-abstract-short" style="display: inline;"> Multimodal large language models (MLLMs) have shown impressive capabilities in document understanding, a rapidly growing research area with significant industrial demand in recent years. As a multimodal task, document understanding requires models to possess both perceptual and cognitive abilities. However, current MLLMs often face conflicts between perception and cognition. Taking a document VQA… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.07722v1-abstract-full').style.display = 'inline'; document.getElementById('2411.07722v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.07722v1-abstract-full" style="display: none;"> Multimodal large language models (MLLMs) have shown impressive capabilities in document understanding, a rapidly growing research area with significant industrial demand in recent years. As a multimodal task, document understanding requires models to possess both perceptual and cognitive abilities. However, current MLLMs often face conflicts between perception and cognition. Taking a document VQA task (cognition) as an example, an MLLM might generate answers that do not match the corresponding visual content identified by its OCR (perception). This conflict suggests that the MLLM might struggle to establish an intrinsic connection between the information it "sees" and what it "understands." Such conflicts challenge the intuitive notion that cognition is consistent with perception, hindering the performance and explainability of MLLMs. In this paper, we define the conflicts between cognition and perception as Cognition and Perception (C&P) knowledge conflicts, a form of multimodal knowledge conflicts, and systematically assess them with a focus on document understanding. Our analysis reveals that even GPT-4o, a leading MLLM, achieves only 68.6% C&P consistency. To mitigate the C&P knowledge conflicts, we propose a novel method called Multimodal Knowledge Consistency Fine-tuning. This method first ensures task-specific consistency and then connects the cognitive and perceptual knowledge. Our method significantly reduces C&P knowledge conflicts across all tested MLLMs and enhances their performance in both cognitive and perceptual tasks in most scenarios. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.07722v1-abstract-full').style.display = 'none'; document.getElementById('2411.07722v1-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 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">Preprint</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.04997">arXiv:2411.04997</a> <span> [<a href="https://arxiv.org/pdf/2411.04997">pdf</a>, <a href="https://arxiv.org/format/2411.04997">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="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> LLM2CLIP: Powerful Language Model Unlocks Richer Visual Representation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Huang%2C+W">Weiquan Huang</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+A">Aoqi Wu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yifan Yang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+X">Xufang Luo</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yuqing Yang</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+L">Liang Hu</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Q">Qi Dai</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+X">Xiyang Dai</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+D">Dongdong Chen</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+L">Lili Qiu</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="2411.04997v3-abstract-short" style="display: inline;"> CLIP is a foundational multimodal model that aligns image and text features into a shared space using contrastive learning on large-scale image-text pairs. Its strength lies in leveraging natural language as a rich supervisory signal. With the rapid progress of large language models (LLMs), we explore their potential to further enhance CLIP's multimodal representation learning. This work introduce… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.04997v3-abstract-full').style.display = 'inline'; document.getElementById('2411.04997v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.04997v3-abstract-full" style="display: none;"> CLIP is a foundational multimodal model that aligns image and text features into a shared space using contrastive learning on large-scale image-text pairs. Its strength lies in leveraging natural language as a rich supervisory signal. With the rapid progress of large language models (LLMs), we explore their potential to further enhance CLIP's multimodal representation learning. This work introduces a fine-tuning approach that integrates LLMs with the pretrained CLIP visual encoder, leveraging LLMs' advanced text understanding and open-world knowledge to improve CLIP's ability to process long and complex captions. To address the challenge of LLMs' autoregressive nature, we propose a caption-to-caption contrastive learning framework to enhance the discriminative power of their outputs. Our method achieves substantial performance gains on various downstream tasks, demonstrating the effectiveness of combining LLMs with CLIP for enhanced multimodal learning. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.04997v3-abstract-full').style.display = 'none'; document.getElementById('2411.04997v3-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 7 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/2411.01595">arXiv:2411.01595</a> <span> [<a href="https://arxiv.org/pdf/2411.01595">pdf</a>, <a href="https://arxiv.org/format/2411.01595">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"> RS-MoE: A Vision-Language Model with Mixture of Experts for Remote Sensing Image Captioning and Visual Question Answering </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Lin%2C+H">Hui Lin</a>, <a href="/search/cs?searchtype=author&query=Hong%2C+D">Danfeng Hong</a>, <a href="/search/cs?searchtype=author&query=Ge%2C+S">Shuhang Ge</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuyao Luo</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+K">Kai Jiang</a>, <a href="/search/cs?searchtype=author&query=Jin%2C+H">Hao Jin</a>, <a href="/search/cs?searchtype=author&query=Wen%2C+C">Congcong Wen</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="2411.01595v2-abstract-short" style="display: inline;"> Remote Sensing Image Captioning (RSIC) presents unique challenges and plays a critical role in applications. Traditional RSIC methods often struggle to produce rich and diverse descriptions. Recently, with advancements in VLMs, efforts have emerged to integrate these models into the remote sensing domain and to introduce descriptive datasets specifically designed to enhance VLM training. This pape… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.01595v2-abstract-full').style.display = 'inline'; document.getElementById('2411.01595v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.01595v2-abstract-full" style="display: none;"> Remote Sensing Image Captioning (RSIC) presents unique challenges and plays a critical role in applications. Traditional RSIC methods often struggle to produce rich and diverse descriptions. Recently, with advancements in VLMs, efforts have emerged to integrate these models into the remote sensing domain and to introduce descriptive datasets specifically designed to enhance VLM training. This paper proposes RS-MoE, a first Mixture of Expert based VLM specifically customized for remote sensing domain. Unlike traditional MoE models, the core of RS-MoE is the MoE Block, which incorporates a novel Instruction Router and multiple lightweight Large Language Models (LLMs) as expert models. The Instruction Router is designed to generate specific prompts tailored for each corresponding LLM, guiding them to focus on distinct aspects of the RSIC task. This design not only allows each expert LLM to concentrate on a specific subset of the task, thereby enhancing the specificity and accuracy of the generated captions, but also improves the scalability of the model by facilitating parallel processing of sub-tasks. Additionally, we present a two-stage training strategy for tuning our RS-MoE model to prevent performance degradation due to sparsity. We fine-tuned our model on the RSICap dataset using our proposed training strategy. Experimental results on the RSICap dataset, along with evaluations on other traditional datasets where no additional fine-tuning was applied, demonstrate that our model achieves state-of-the-art performance in generating precise and contextually relevant captions. Notably, our RS-MoE-1B variant achieves performance comparable to 13B VLMs, demonstrating the efficiency of our model design. Moreover, our model demonstrates promising generalization capabilities by consistently achieving state-of-the-art performance on the Remote Sensing Visual Question Answering (RSVQA) task. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.01595v2-abstract-full').style.display = 'none'; document.getElementById('2411.01595v2-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 3 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/2411.00771">arXiv:2411.00771</a> <span> [<a href="https://arxiv.org/pdf/2411.00771">pdf</a>, <a href="https://arxiv.org/format/2411.00771">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"> CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Liu%2C+Y">Yang Liu</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chuanchen Luo</a>, <a href="/search/cs?searchtype=author&query=Mao%2C+Z">Zhongkai Mao</a>, <a href="/search/cs?searchtype=author&query=Peng%2C+J">Junran Peng</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Z">Zhaoxiang Zhang</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="2411.00771v2-abstract-short" style="display: inline;"> Recently, 3D Gaussian Splatting (3DGS) has revolutionized radiance field reconstruction, manifesting efficient and high-fidelity novel view synthesis. However, accurately representing surfaces, especially in large and complex scenarios, remains a significant challenge due to the unstructured nature of 3DGS. In this paper, we present CityGaussianV2, a novel approach for large-scale scene reconstruc… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.00771v2-abstract-full').style.display = 'inline'; document.getElementById('2411.00771v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.00771v2-abstract-full" style="display: none;"> Recently, 3D Gaussian Splatting (3DGS) has revolutionized radiance field reconstruction, manifesting efficient and high-fidelity novel view synthesis. However, accurately representing surfaces, especially in large and complex scenarios, remains a significant challenge due to the unstructured nature of 3DGS. In this paper, we present CityGaussianV2, a novel approach for large-scale scene reconstruction that addresses critical challenges related to geometric accuracy and efficiency. Building on the favorable generalization capabilities of 2D Gaussian Splatting (2DGS), we address its convergence and scalability issues. Specifically, we implement a decomposed-gradient-based densification and depth regression technique to eliminate blurry artifacts and accelerate convergence. To scale up, we introduce an elongation filter that mitigates Gaussian count explosion caused by 2DGS degeneration. Furthermore, we optimize the CityGaussian pipeline for parallel training, achieving up to 10$\times$ compression, at least 25% savings in training time, and a 50% decrease in memory usage. We also established standard geometry benchmarks under large-scale scenes. Experimental results demonstrate that our method strikes a promising balance between visual quality, geometric accuracy, as well as storage and training costs. The project page is available at https://dekuliutesla.github.io/CityGaussianV2/. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.00771v2-abstract-full').style.display = 'none'; document.getElementById('2411.00771v2-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 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 1 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 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 ICLR2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.22128">arXiv:2410.22128</a> <span> [<a href="https://arxiv.org/pdf/2410.22128">pdf</a>, <a href="https://arxiv.org/format/2410.22128">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"> PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Hong%2C+S">Sunghwan Hong</a>, <a href="/search/cs?searchtype=author&query=Jung%2C+J">Jaewoo Jung</a>, <a href="/search/cs?searchtype=author&query=Shin%2C+H">Heeseong Shin</a>, <a href="/search/cs?searchtype=author&query=Han%2C+J">Jisang Han</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+J">Jiaolong Yang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chong Luo</a>, <a href="/search/cs?searchtype=author&query=Kim%2C+S">Seungryong 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="2410.22128v1-abstract-short" style="display: inline;"> We consider the problem of novel view synthesis from unposed images in a single feed-forward. Our framework capitalizes on fast speed, scalability, and high-quality 3D reconstruction and view synthesis capabilities of 3DGS, where we further extend it to offer a practical solution that relaxes common assumptions such as dense image views, accurate camera poses, and substantial image overlaps. We ac… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22128v1-abstract-full').style.display = 'inline'; document.getElementById('2410.22128v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.22128v1-abstract-full" style="display: none;"> We consider the problem of novel view synthesis from unposed images in a single feed-forward. Our framework capitalizes on fast speed, scalability, and high-quality 3D reconstruction and view synthesis capabilities of 3DGS, where we further extend it to offer a practical solution that relaxes common assumptions such as dense image views, accurate camera poses, and substantial image overlaps. We achieve this through identifying and addressing unique challenges arising from the use of pixel-aligned 3DGS: misaligned 3D Gaussians across different views induce noisy or sparse gradients that destabilize training and hinder convergence, especially when above assumptions are not met. To mitigate this, we employ pre-trained monocular depth estimation and visual correspondence models to achieve coarse alignments of 3D Gaussians. We then introduce lightweight, learnable modules to refine depth and pose estimates from the coarse alignments, improving the quality of 3D reconstruction and novel view synthesis. Furthermore, the refined estimates are leveraged to estimate geometry confidence scores, which assess the reliability of 3D Gaussian centers and condition the prediction of Gaussian parameters accordingly. Extensive evaluations on large-scale real-world datasets demonstrate that PF3plat sets a new state-of-the-art across all benchmarks, supported by comprehensive ablation studies validating our design choices. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22128v1-abstract-full').style.display = 'none'; document.getElementById('2410.22128v1-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 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">project page: https://cvlab-kaist.github.io/PF3plat/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.17952">arXiv:2410.17952</a> <span> [<a href="https://arxiv.org/pdf/2410.17952">pdf</a>, <a href="https://arxiv.org/format/2410.17952">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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="Information Retrieval">cs.IR</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"> SimRAG: Self-Improving Retrieval-Augmented Generation for Adapting Large Language Models to Specialized Domains </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xu%2C+R">Ran Xu</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+H">Hui Liu</a>, <a href="/search/cs?searchtype=author&query=Nag%2C+S">Sreyashi Nag</a>, <a href="/search/cs?searchtype=author&query=Dai%2C+Z">Zhenwei Dai</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+Y">Yaochen Xie</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yang Li</a>, <a href="/search/cs?searchtype=author&query=Ho%2C+J+C">Joyce C. Ho</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+C">Carl Yang</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</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="2410.17952v2-abstract-short" style="display: inline;"> Retrieval-augmented generation (RAG) enhances the question-answering (QA) abilities of large language models (LLMs) by integrating external knowledge. However, adapting general-purpose RAG systems to specialized fields such as science and medicine poses unique challenges due to distribution shifts and limited access to domain-specific data. To tackle this, we propose SimRAG, a self-training approa… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.17952v2-abstract-full').style.display = 'inline'; document.getElementById('2410.17952v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.17952v2-abstract-full" style="display: none;"> Retrieval-augmented generation (RAG) enhances the question-answering (QA) abilities of large language models (LLMs) by integrating external knowledge. However, adapting general-purpose RAG systems to specialized fields such as science and medicine poses unique challenges due to distribution shifts and limited access to domain-specific data. To tackle this, we propose SimRAG, a self-training approach that equips the LLM with joint capabilities of question answering and question generation for domain adaptation. Our method first fine-tunes the LLM on instruction-following, question-answering, and search-related data. Then, it prompts the same LLM to generate diverse domain-relevant questions from unlabeled corpora, with an additional filtering strategy to retain high-quality synthetic examples. By leveraging these self-generated synthetic examples, the LLM can improve their performance on domain-specific RAG tasks. Experiments on 11 datasets, spanning two backbone sizes and three domains, demonstrate that SimRAG outperforms baselines by 1.2\%--8.6\%. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.17952v2-abstract-full').style.display = 'none'; document.getElementById('2410.17952v2-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 23 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 to NAACL 2025 main conference</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> NAACL 2025 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.16540">arXiv:2410.16540</a> <span> [<a href="https://arxiv.org/pdf/2410.16540">pdf</a>, <a href="https://arxiv.org/format/2410.16540">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</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> <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"> A Theoretical Understanding of Chain-of-Thought: Coherent Reasoning and Error-Aware Demonstration </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Cui%2C+Y">Yingqian Cui</a>, <a href="/search/cs?searchtype=author&query=He%2C+P">Pengfei He</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+X">Xianfeng Tang</a>, <a href="/search/cs?searchtype=author&query=He%2C+Q">Qi He</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chen Luo</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+J">Jiliang Tang</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+Y">Yue Xing</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="2410.16540v1-abstract-short" style="display: inline;"> Few-shot Chain-of-Thought (CoT) prompting has demonstrated strong performance in improving the reasoning capabilities of large language models (LLMs). While theoretical investigations have been conducted to understand CoT, the underlying transformer used in these studies isolates the CoT reasoning process into separated in-context learning steps (Stepwise ICL). In this work, we theoretically show… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.16540v1-abstract-full').style.display = 'inline'; document.getElementById('2410.16540v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.16540v1-abstract-full" style="display: none;"> Few-shot Chain-of-Thought (CoT) prompting has demonstrated strong performance in improving the reasoning capabilities of large language models (LLMs). While theoretical investigations have been conducted to understand CoT, the underlying transformer used in these studies isolates the CoT reasoning process into separated in-context learning steps (Stepwise ICL). In this work, we theoretically show that, compared to Stepwise ICL, the transformer gains better error correction ability and more accurate predictions if the reasoning from earlier steps (Coherent CoT) is integrated. Given that this coherent reasoning changes the behavior of the transformer, we further investigate the sensitivity of the transformer with Coherent CoT when the demonstration examples are corrupted at the inference stage. Our theoretical results indicate that the transformer is more sensitive to errors in intermediate reasoning steps than the final outcome. Building upon this observation, we propose an improvement on CoT by incorporating both correct and incorrect reasoning paths in the demonstration. Our experiments validate the effectiveness of the proposed approach. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.16540v1-abstract-full').style.display = 'none'; document.getElementById('2410.16540v1-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> 21 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/2410.16146">arXiv:2410.16146</a> <span> [<a href="https://arxiv.org/pdf/2410.16146">pdf</a>, <a href="https://arxiv.org/format/2410.16146">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="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Towards Combating Frequency Simplicity-biased Learning for Domain Generalization </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=He%2C+X">Xilin He</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jingyu Hu</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+Q">Qinliang Lin</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cheng Luo</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+W">Weicheng Xie</a>, <a href="/search/cs?searchtype=author&query=Song%2C+S">Siyang Song</a>, <a href="/search/cs?searchtype=author&query=Khan%2C+M+H">Muhammad Haris Khan</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+L">Linlin Shen</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="2410.16146v1-abstract-short" style="display: inline;"> Domain generalization methods aim to learn transferable knowledge from source domains that can generalize well to unseen target domains. Recent studies show that neural networks frequently suffer from a simplicity-biased learning behavior which leads to over-reliance on specific frequency sets, namely as frequency shortcuts, instead of semantic information, resulting in poor generalization perform… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.16146v1-abstract-full').style.display = 'inline'; document.getElementById('2410.16146v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.16146v1-abstract-full" style="display: none;"> Domain generalization methods aim to learn transferable knowledge from source domains that can generalize well to unseen target domains. Recent studies show that neural networks frequently suffer from a simplicity-biased learning behavior which leads to over-reliance on specific frequency sets, namely as frequency shortcuts, instead of semantic information, resulting in poor generalization performance. Despite previous data augmentation techniques successfully enhancing generalization performances, they intend to apply more frequency shortcuts, thereby causing hallucinations of generalization improvement. In this paper, we aim to prevent such learning behavior of applying frequency shortcuts from a data-driven perspective. Given the theoretical justification of models' biased learning behavior on different spatial frequency components, which is based on the dataset frequency properties, we argue that the learning behavior on various frequency components could be manipulated by changing the dataset statistical structure in the Fourier domain. Intuitively, as frequency shortcuts are hidden in the dominant and highly dependent frequencies of dataset structure, dynamically perturbating the over-reliance frequency components could prevent the application of frequency shortcuts. To this end, we propose two effective data augmentation modules designed to collaboratively and adaptively adjust the frequency characteristic of the dataset, aiming to dynamically influence the learning behavior of the model and ultimately serving as a strategy to mitigate shortcut learning. Code is available at AdvFrequency (https://github.com/C0notSilly/AdvFrequency). <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.16146v1-abstract-full').style.display = 'none'; document.getElementById('2410.16146v1-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> 21 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 NeurIPS 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/2410.15817">arXiv:2410.15817</a> <span> [<a href="https://arxiv.org/pdf/2410.15817">pdf</a>, <a href="https://arxiv.org/format/2410.15817">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computational Engineering, Finance, and Science">cs.CE</span> </div> </div> <p class="title is-5 mathjax"> Large Language Models Empower Personalized Valuation in Auction </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sun%2C+J">Jie Sun</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+T">Tianyu Zhang</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+H">Houcheng Jiang</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+K">Kexin Huang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Chi Luo</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+J">Junkang Wu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+J">Jiancan Wu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+A">An Zhang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+X">Xiang Wang</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="2410.15817v1-abstract-short" style="display: inline;"> Auctions, a fundamental economic mechanism, encompass the valuation of goods or services and the competitive bidding algorithms within a specific framework, serving to uncover the true market value. However, current research predominantly focuses on the bidding algorithms within a given auction mechanism, often overlooking the advantages of incorporating individual bidders' unique preferences and… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.15817v1-abstract-full').style.display = 'inline'; document.getElementById('2410.15817v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.15817v1-abstract-full" style="display: none;"> Auctions, a fundamental economic mechanism, encompass the valuation of goods or services and the competitive bidding algorithms within a specific framework, serving to uncover the true market value. However, current research predominantly focuses on the bidding algorithms within a given auction mechanism, often overlooking the advantages of incorporating individual bidders' unique preferences and the semantic information related to the items into the valuation process. Our analysis, both theoretical and empirical, shows that imprecise or noisy valuations can significantly affect the overall utility for participants. To bridge this gap, we propose a personalized valuation framework, namely \textbf{S}emantic-enhanced \textbf{P}ersonalized \textbf{V}aluation in \textbf{A}uction (\ours), which integrates Large Language Models (LLMs) to incorporate semantic information into each bidder's unique valuation process. Specifically, SPVA employs a two-stage approach: it first fine-tunes LLMs to encode bidder preferences in personalized valuations, and then constructs a Vickrey auction environment integrated with a bidding algorithm to demonstrate that SPVA's more accurate valuations result in higher profits. Additionally, we have developed a semantic-enhanced dataset comprising over 23,000 samples and introduced new personalized evaluation metrics that reflect both bidder preferences and profit. Through simulations of various auction scenarios, our method demonstrates its ability to provide accurate valuations and capture bidder preferences, affirming the method's effectiveness in real-world auction settings. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.15817v1-abstract-full').style.display = 'none'; document.getElementById('2410.15817v1-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> 21 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">14 pages, 5 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/2410.13439">arXiv:2410.13439</a> <span> [<a href="https://arxiv.org/pdf/2410.13439">pdf</a>, <a href="https://arxiv.org/format/2410.13439">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> <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"> Similarity-Dissimilarity Loss for Multi-label Supervised Contrastive Learning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Huang%2C+G">Guangming Huang</a>, <a href="/search/cs?searchtype=author&query=Long%2C+Y">Yunfei Long</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+C">Cunjin Luo</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+S">Sheng 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="2410.13439v2-abstract-short" style="display: inline;"> Supervised contrastive learning has been explored in making use of label information for multi-label classification, but determining positive samples in multi-label scenario remains challenging. Previous studies have examined strategies for identifying positive samples, considering label overlap proportion between anchors and samples. However, they ignore various relations between given anchors an… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.13439v2-abstract-full').style.display = 'inline'; document.getElementById('2410.13439v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.13439v2-abstract-full" style="display: none;"> Supervised contrastive learning has been explored in making use of label information for multi-label classification, but determining positive samples in multi-label scenario remains challenging. Previous studies have examined strategies for identifying positive samples, considering label overlap proportion between anchors and samples. However, they ignore various relations between given anchors and samples, as well as how to dynamically adjust the weights in contrastive loss functions based on different relations, leading to great ambiguity. In this paper, we introduce five distinct relations between multi-label samples and propose a Similarity-Dissimilarity Loss with contrastive learning for multi-label classification. Our loss function re-weights the loss by computing the similarity and dissimilarity between positive samples and a given anchor based on the introduced relations. We mainly conduct experiments for multi-label text classification on MIMIC datasets, then further extend the evaluation on MS-COCO. The Experimental results show that our proposed loss effectively improves the performance on all encoders under supervised contrastive learning paradigm, demonstrating its effectiveness and robustness. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.13439v2-abstract-full').style.display = 'none'; document.getElementById('2410.13439v2-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> 23 March, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 17 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2024. </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=Luo%2C+C&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> <li> <a href="/search/?searchtype=author&query=Luo%2C+C&start=200" class="pagination-link " aria-label="Page 5" aria-current="page">5 </a> </li> <li><span class="pagination-ellipsis">…</span></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