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 175 results for author: <span class="mathjax">Baek, J</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/" aria-role="search"> <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="Baek, J"> </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=Baek%2C+J&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="Baek, J"> <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=Baek%2C+J&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> </ul> </nav> <ol class="breathe-horizontal" start="1"> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.07274">arXiv:2411.07274</a> <span> [<a href="https://arxiv.org/pdf/2411.07274">pdf</a>, <a href="https://arxiv.org/ps/2411.07274">ps</a>, <a href="https://arxiv.org/format/2411.07274">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Combinatorics">math.CO</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Metric Geometry">math.MG</span> </div> </div> <p class="title is-5 mathjax"> A note on the Erd艖s conjecture about square packing </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jineon Baek</a>, <a href="/search/?searchtype=author&query=Koizumi%2C+J">Junnosuke Koizumi</a>, <a href="/search/?searchtype=author&query=Ueoro%2C+T">Takahiro Ueoro</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="2411.07274v2-abstract-short" style="display: inline;"> Let $f(n)$ denote the maximum total length of the sides of $n$ squares packed inside a unit square. Erd艖s conjectured that $f(k^2+1)=k$. We show that the conjecture is true if we assume that the sides of the squares are parallel to the sides of the unit square. </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.07274v2-abstract-full" style="display: none;"> Let $f(n)$ denote the maximum total length of the sides of $n$ squares packed inside a unit square. Erd艖s conjectured that $f(k^2+1)=k$. We show that the conjecture is true if we assume that the sides of the squares are parallel to the sides of the unit square. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.07274v2-abstract-full').style.display = 'none'; document.getElementById('2411.07274v2-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 11 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">5 pages. Comments welcome!</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Report number:</span> RIKEN-iTHEMS-Report-24 <span class="has-text-black-bis has-text-weight-semibold">MSC Class:</span> 52C15; 52C10 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.06738">arXiv:2411.06738</a> <span> [<a href="https://arxiv.org/pdf/2411.06738">pdf</a>, <a href="https://arxiv.org/format/2411.06738">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Image and Video Processing">eess.IV</span> </div> </div> <p class="title is-5 mathjax"> 360-Degree Video Super Resolution and Quality Enhancement Challenge: Methods and Results </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Telili%2C+A">Ahmed Telili</a>, <a href="/search/?searchtype=author&query=Hamidouche%2C+W">Wassim Hamidouche</a>, <a href="/search/?searchtype=author&query=Farhat%2C+I">Ibrahim Farhat</a>, <a href="/search/?searchtype=author&query=Amirpour%2C+H">Hadi Amirpour</a>, <a href="/search/?searchtype=author&query=Timmerer%2C+C">Christian Timmerer</a>, <a href="/search/?searchtype=author&query=Khadraoui%2C+I">Ibrahim Khadraoui</a>, <a href="/search/?searchtype=author&query=Lu%2C+J">Jiajie Lu</a>, <a href="/search/?searchtype=author&query=Van+Le%2C+T">The Van Le</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jeonneung Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+J+Y">Jin Young Lee</a>, <a href="/search/?searchtype=author&query=Wei%2C+Y">Yiying Wei</a>, <a href="/search/?searchtype=author&query=Sun%2C+X">Xiaopeng Sun</a>, <a href="/search/?searchtype=author&query=Gao%2C+Y">Yu Gao</a>, <a href="/search/?searchtype=author&query=Huangl%2C+J">JianCheng Huangl</a>, <a href="/search/?searchtype=author&query=Zhong%2C+Y">Yujie Zhong</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="2411.06738v1-abstract-short" style="display: inline;"> Omnidirectional (360-degree) video is rapidly gaining popularity due to advancements in immersive technologies like virtual reality (VR) and extended reality (XR). However, real-time streaming of such videos, especially in live mobile scenarios like unmanned aerial vehicles (UAVs), is challenged by limited bandwidth and strict latency constraints. Traditional methods, such as compression and adapt… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06738v1-abstract-full').style.display = 'inline'; document.getElementById('2411.06738v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.06738v1-abstract-full" style="display: none;"> Omnidirectional (360-degree) video is rapidly gaining popularity due to advancements in immersive technologies like virtual reality (VR) and extended reality (XR). However, real-time streaming of such videos, especially in live mobile scenarios like unmanned aerial vehicles (UAVs), is challenged by limited bandwidth and strict latency constraints. Traditional methods, such as compression and adaptive resolution, help but often compromise video quality and introduce artifacts that degrade the viewer experience. Additionally, the unique spherical geometry of 360-degree video presents challenges not encountered in traditional 2D video. To address these issues, we initiated the 360-degree Video Super Resolution and Quality Enhancement Challenge. This competition encourages participants to develop efficient machine learning solutions to enhance the quality of low-bitrate compressed 360-degree videos, with two tracks focusing on 2x and 4x super-resolution (SR). In this paper, we outline the challenge framework, detailing the two competition tracks and highlighting the SR solutions proposed by the top-performing models. We assess these models within a unified framework, considering quality enhancement, bitrate gain, and computational efficiency. This challenge aims to drive innovation in real-time 360-degree video streaming, improving the quality and accessibility of immersive visual experiences. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06738v1-abstract-full').style.display = 'none'; document.getElementById('2411.06738v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 11 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">14 pages, 9 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/2411.06387">arXiv:2411.06387</a> <span> [<a href="https://arxiv.org/pdf/2411.06387">pdf</a>, <a href="https://arxiv.org/format/2411.06387">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="Computation and Language">cs.CL</span> </div> </div> <p class="title is-5 mathjax"> Self-Training Meets Consistency: Improving LLMs' Reasoning With Consistency-Driven Rationale Evaluation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Lee%2C+J">Jaehyeok Lee</a>, <a href="/search/?searchtype=author&query=Sakaguchi%2C+K">Keisuke Sakaguchi</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</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="2411.06387v3-abstract-short" style="display: inline;"> Self-training approach for large language models (LLMs) improves reasoning abilities by training the models on their self-generated rationales. Previous approaches have labeled rationales that produce correct answers for a given question as appropriate for training. However, a single measure risks misjudging rationale quality, leading the models to learn flawed reasoning patterns. To address this… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06387v3-abstract-full').style.display = 'inline'; document.getElementById('2411.06387v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.06387v3-abstract-full" style="display: none;"> Self-training approach for large language models (LLMs) improves reasoning abilities by training the models on their self-generated rationales. Previous approaches have labeled rationales that produce correct answers for a given question as appropriate for training. However, a single measure risks misjudging rationale quality, leading the models to learn flawed reasoning patterns. To address this issue, we propose CREST (Consistency-driven Rationale Evaluation for Self-Training), a self-training framework that further evaluates each rationale through follow-up questions and leverages this evaluation to guide its training. Specifically, we introduce two methods: (1) filtering out rationales that frequently result in incorrect answers on follow-up questions and (2) preference learning based on mixed preferences from rationale evaluation results of both original and follow-up questions. Experiments on three question-answering datasets using open LLMs show that CREST not only improves the logical robustness and correctness of rationales but also improves reasoning abilities compared to previous self-training approaches. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06387v3-abstract-full').style.display = 'none'; document.getElementById('2411.06387v3-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 10 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">Under review</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.06071">arXiv:2411.06071</a> <span> [<a href="https://arxiv.org/pdf/2411.06071">pdf</a>, <a href="https://arxiv.org/format/2411.06071">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"> GlocalCLIP: Object-agnostic Global-Local Prompt Learning for Zero-shot Anomaly Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Ham%2C+J">Jiyul Ham</a>, <a href="/search/?searchtype=author&query=Jung%2C+Y">Yonggon Jung</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jun-Geol Baek</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="2411.06071v1-abstract-short" style="display: inline;"> Zero-shot anomaly detection (ZSAD) is crucial for detecting abnormal patterns in target datasets without using training samples, specifically in scenarios where there are distributional differences between the target domain and training data or where data scarcity arises because of restricted access. Although recently pretrained vision-language models demonstrate strong zero-shot performance acros… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06071v1-abstract-full').style.display = 'inline'; document.getElementById('2411.06071v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.06071v1-abstract-full" style="display: none;"> Zero-shot anomaly detection (ZSAD) is crucial for detecting abnormal patterns in target datasets without using training samples, specifically in scenarios where there are distributional differences between the target domain and training data or where data scarcity arises because of restricted access. Although recently pretrained vision-language models demonstrate strong zero-shot performance across various visual tasks, they focus on learning class semantics, which makes their direct application to ZSAD challenging. To address this scenario, we propose GlocalCLIP, which uniquely separates global and local prompts and jointly optimizes them. This approach enables the object-agnostic glocal semantic prompt design to effectively capture general normal and anomalous patterns without dependency on specific objects in the image. We refine the text prompts for more precise adjustments by utilizing deep-text prompt tuning in the text encoder. In the vision encoder, we apply V-V attention layers to capture detailed local image features. Finally, we introduce glocal contrastive learning to improve the complementary learning of global and local prompts, effectively detecting abnormal patterns across various domains. The generalization performance of GlocalCLIP in ZSAD was demonstrated on 15 real-world datasets from both the industrial and medical domains, achieving superior performance compared to existing methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06071v1-abstract-full').style.display = 'none'; document.getElementById('2411.06071v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 9 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">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">28 pages, 33 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/2410.22375">arXiv:2410.22375</a> <span> [<a href="https://arxiv.org/pdf/2410.22375">pdf</a>, <a href="https://arxiv.org/format/2410.22375">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> <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"> Rethinking Code Refinement: Learning to Judge Code Efficiency </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Seo%2C+M">Minju Seo</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.22375v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) have demonstrated impressive capabilities in understanding and generating codes. Due to these capabilities, many recent methods are proposed to automatically refine the codes with LLMs. However, we should rethink that the refined codes (from LLMs and even humans) are not always more efficient than their original versions. On the other hand, running two different versio… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22375v1-abstract-full').style.display = 'inline'; document.getElementById('2410.22375v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.22375v1-abstract-full" style="display: none;"> Large Language Models (LLMs) have demonstrated impressive capabilities in understanding and generating codes. Due to these capabilities, many recent methods are proposed to automatically refine the codes with LLMs. However, we should rethink that the refined codes (from LLMs and even humans) are not always more efficient than their original versions. On the other hand, running two different versions of codes and comparing them every time is not ideal and time-consuming. Therefore, in this work, we propose a novel method based on the code language model that is trained to judge the efficiency between two different codes (generated across humans and machines) by either classifying the superior one or predicting the relative improvement. We validate our method on multiple programming languages with multiple refinement steps, demonstrating that the proposed method can effectively distinguish between more and less efficient versions of code. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.22375v1-abstract-full').style.display = 'none'; document.getElementById('2410.22375v1-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> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.17250">arXiv:2410.17250</a> <span> [<a href="https://arxiv.org/pdf/2410.17250">pdf</a>, <a href="https://arxiv.org/format/2410.17250">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="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> JMMMU: A Japanese Massive Multi-discipline Multimodal Understanding Benchmark for Culture-aware Evaluation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Onohara%2C+S">Shota Onohara</a>, <a href="/search/?searchtype=author&query=Miyai%2C+A">Atsuyuki Miyai</a>, <a href="/search/?searchtype=author&query=Imajuku%2C+Y">Yuki Imajuku</a>, <a href="/search/?searchtype=author&query=Egashira%2C+K">Kazuki Egashira</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jeonghun Baek</a>, <a href="/search/?searchtype=author&query=Yue%2C+X">Xiang Yue</a>, <a href="/search/?searchtype=author&query=Neubig%2C+G">Graham Neubig</a>, <a href="/search/?searchtype=author&query=Aizawa%2C+K">Kiyoharu Aizawa</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="2410.17250v1-abstract-short" style="display: inline;"> Accelerating research on Large Multimodal Models (LMMs) in non-English languages is crucial for enhancing user experiences across broader populations. In this paper, we introduce JMMMU (Japanese MMMU), the first large-scale Japanese benchmark designed to evaluate LMMs on expert-level tasks based on the Japanese cultural context. To facilitate comprehensive culture-aware evaluation, JMMMU features… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.17250v1-abstract-full').style.display = 'inline'; document.getElementById('2410.17250v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.17250v1-abstract-full" style="display: none;"> Accelerating research on Large Multimodal Models (LMMs) in non-English languages is crucial for enhancing user experiences across broader populations. In this paper, we introduce JMMMU (Japanese MMMU), the first large-scale Japanese benchmark designed to evaluate LMMs on expert-level tasks based on the Japanese cultural context. To facilitate comprehensive culture-aware evaluation, JMMMU features two complementary subsets: (i) culture-agnostic (CA) subset, where the culture-independent subjects (e.g., Math) are selected and translated into Japanese, enabling one-to-one comparison with its English counterpart MMMU; and (ii) culture-specific (CS) subset, comprising newly crafted subjects that reflect Japanese cultural context. Using the CA subset, we observe performance drop in many LMMs when evaluated in Japanese, which is purely attributable to language variation. Using the CS subset, we reveal their inadequate Japanese cultural understanding. Further, by combining both subsets, we identify that some LMMs perform well on the CA subset but not on the CS subset, exposing a shallow understanding of the Japanese language that lacks depth in cultural understanding. We hope this work will not only help advance LMM performance in Japanese but also serve as a guideline to create high-standard, culturally diverse benchmarks for multilingual LMM development. The project page is https://mmmu-japanese-benchmark.github.io/JMMMU/. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.17250v1-abstract-full').style.display = 'none'; document.getElementById('2410.17250v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 22 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://mmmu-japanese-benchmark.github.io/JMMMU/</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2410.02729">arXiv:2410.02729</a> <span> [<a href="https://arxiv.org/pdf/2410.02729">pdf</a>, <a href="https://arxiv.org/format/2410.02729">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> </div> </div> <p class="title is-5 mathjax"> Unified Multi-Modal Interleaved Document Representation for Information Retrieval </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Lee%2C+J">Jaewoo Lee</a>, <a href="/search/?searchtype=author&query=Ko%2C+J">Joonho Ko</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2410.02729v1-abstract-short" style="display: inline;"> Information Retrieval (IR) methods aim to identify relevant documents in response to a given query, which have gained remarkable attention due to their successful application in various natural language tasks. However, existing approaches typically consider only the textual information within the documents, which overlooks the fact that documents can contain multiple modalities, including texts, i… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.02729v1-abstract-full').style.display = 'inline'; document.getElementById('2410.02729v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.02729v1-abstract-full" style="display: none;"> Information Retrieval (IR) methods aim to identify relevant documents in response to a given query, which have gained remarkable attention due to their successful application in various natural language tasks. However, existing approaches typically consider only the textual information within the documents, which overlooks the fact that documents can contain multiple modalities, including texts, images, and tables. Further, they often segment each long document into multiple discrete passages for embedding, preventing them from capturing the overall document context and interactions between paragraphs. We argue that these two limitations lead to suboptimal document representations for retrieval. In this work, to address them, we aim to produce more comprehensive and nuanced document representations by holistically embedding documents interleaved with different modalities. Specifically, we achieve this by leveraging the capability of recent vision-language models that enable the processing and integration of text, images, and tables into a unified format and representation. Moreover, to mitigate the information loss from segmenting documents into passages, instead of representing and retrieving passages individually, we further merge the representations of segmented passages into one single document representation, while we additionally introduce a reranking strategy to decouple and identify the relevant passage within the document if necessary. Then, through extensive experiments on diverse information retrieval scenarios considering both the textual and multimodal queries, we show that our approach substantially outperforms relevant baselines, thanks to the consideration of the multimodal information interleaved within the documents in a unified way. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.02729v1-abstract-full').style.display = 'none'; document.getElementById('2410.02729v1-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 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">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/2410.00328">arXiv:2410.00328</a> <span> [<a href="https://arxiv.org/pdf/2410.00328">pdf</a>, <a href="https://arxiv.org/format/2410.00328">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Performance">cs.PF</span> </div> </div> <p class="title is-5 mathjax"> Tuning Fast Memory Size based on Modeling of Page Migration for Tiered Memory </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Chen%2C+S">Shangye Chen</a>, <a href="/search/?searchtype=author&query=Huang%2C+J">Jin Huang</a>, <a href="/search/?searchtype=author&query=Yang%2C+S">Shuangyan Yang</a>, <a href="/search/?searchtype=author&query=Liu%2C+J">Jie Liu</a>, <a href="/search/?searchtype=author&query=Li%2C+H">Huaicheng Li</a>, <a href="/search/?searchtype=author&query=Nikolopoulos%2C+D">Dimitrios Nikolopoulos</a>, <a href="/search/?searchtype=author&query=Ryu%2C+J">Junhee Ryu</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinho Baek</a>, <a href="/search/?searchtype=author&query=Shin%2C+K">Kwangsik Shin</a>, <a href="/search/?searchtype=author&query=Li%2C+D">Dong Li</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="2410.00328v1-abstract-short" style="display: inline;"> Tiered memory, built upon a combination of fast memory and slow memory, provides a cost-effective solution to meet ever-increasing requirements from emerging applications for large memory capacity. Reducing the size of fast memory is valuable to improve memory utilization in production and reduce production costs because fast memory tends to be expensive. However, deciding the fast memory size is… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.00328v1-abstract-full').style.display = 'inline'; document.getElementById('2410.00328v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2410.00328v1-abstract-full" style="display: none;"> Tiered memory, built upon a combination of fast memory and slow memory, provides a cost-effective solution to meet ever-increasing requirements from emerging applications for large memory capacity. Reducing the size of fast memory is valuable to improve memory utilization in production and reduce production costs because fast memory tends to be expensive. However, deciding the fast memory size is challenging because there is a complex interplay between application characterization and the overhead of page migration used to mitigate the impact of limited fast memory capacity. In this paper, we introduce a system, Tuna, to decide fast memory size based on modeling of page migration. Tuna uses micro-benchmarking to model the impact of page migration on application performance using three metrics. Tuna decides the fast memory size based on offline modeling results and limited information on workload telemetry. Evaluating with common big-memory applications and using 5% as the performance loss target, we show that Tuna in combination with a page management system (TPP) saves fast memory by 8.5% on average (up to 16%). This is in contrast to the 5% saving in fast memory reported by Microsoft Pond for the same workloads (BFS and SSSP) and the same performance loss target. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2410.00328v1-abstract-full').style.display = 'none'; document.getElementById('2410.00328v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 30 September, 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/2408.15180">arXiv:2408.15180</a> <span> [<a href="https://arxiv.org/pdf/2408.15180">pdf</a>, <a href="https://arxiv.org/ps/2408.15180">ps</a>, <a href="https://arxiv.org/format/2408.15180">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="Rings and Algebras">math.RA</span> </div> </div> <p class="title is-5 mathjax"> Formalizing Mason-Stothers Theorem and its Corollaries in Lean 4 </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jineon Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+S">Seewoo Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2408.15180v1-abstract-short" style="display: inline;"> The ABC conjecture implies many conjectures and theorems in number theory, including the celebrated Fermat's Last Theorem. Mason-Stothers Theorem is a function field analogue of the ABC conjecture that admits a much more elementary proof with many interesting consequences, including a polynomial version of Fermat's Last Theorem. While years of dedicated effort are expected for a full formalization… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.15180v1-abstract-full').style.display = 'inline'; document.getElementById('2408.15180v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2408.15180v1-abstract-full" style="display: none;"> The ABC conjecture implies many conjectures and theorems in number theory, including the celebrated Fermat's Last Theorem. Mason-Stothers Theorem is a function field analogue of the ABC conjecture that admits a much more elementary proof with many interesting consequences, including a polynomial version of Fermat's Last Theorem. While years of dedicated effort are expected for a full formalization of Fermat's Last Theorem, the simple proof of Mason-Stothers Theorem and its corollaries calls for an immediate formalization. We formalize an elementary proof of by Snyder in Lean 4, and also formalize many consequences of Mason-Stothers, including (i) non-solvability of Fermat-Cartan equations in polynomials, (ii) non-parametrizability of a certain elliptic curve, and (iii) Davenport's Theorem. We compare our work to existing formalizations of Mason-Stothers by Eberl in Isabelle and Wagemaker in Lean 3 respectively. Our formalization is based on the mathlib4 library of Lean 4, and is currently being ported back to mathlib4. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.15180v1-abstract-full').style.display = 'none'; document.getElementById('2408.15180v1-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 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2408.10107">arXiv:2408.10107</a> <span> [<a href="https://arxiv.org/pdf/2408.10107">pdf</a>, <a href="https://arxiv.org/format/2408.10107">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">stat.ML</span> </div> </div> <p class="title is-5 mathjax"> Perturb-and-Compare Approach for Detecting Out-of-Distribution Samples in Constrained Access Environments </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Lee%2C+H">Heeyoung Lee</a>, <a href="/search/?searchtype=author&query=Byun%2C+H">Hoyoon Byun</a>, <a href="/search/?searchtype=author&query=Oh%2C+C">Changdae Oh</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a>, <a href="/search/?searchtype=author&query=Song%2C+K">Kyungwoo 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="2408.10107v1-abstract-short" style="display: inline;"> Accessing machine learning models through remote APIs has been gaining prevalence following the recent trend of scaling up model parameters for increased performance. Even though these models exhibit remarkable ability, detecting out-of-distribution (OOD) samples remains a crucial safety concern for end users as these samples may induce unreliable outputs from the model. In this work, we propose a… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.10107v1-abstract-full').style.display = 'inline'; document.getElementById('2408.10107v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2408.10107v1-abstract-full" style="display: none;"> Accessing machine learning models through remote APIs has been gaining prevalence following the recent trend of scaling up model parameters for increased performance. Even though these models exhibit remarkable ability, detecting out-of-distribution (OOD) samples remains a crucial safety concern for end users as these samples may induce unreliable outputs from the model. In this work, we propose an OOD detection framework, MixDiff, that is applicable even when the model's parameters or its activations are not accessible to the end user. To bypass the access restriction, MixDiff applies an identical input-level perturbation to a given target sample and a similar in-distribution (ID) sample, then compares the relative difference in the model outputs of these two samples. MixDiff is model-agnostic and compatible with existing output-based OOD detection methods. We provide theoretical analysis to illustrate MixDiff's effectiveness in discerning OOD samples that induce overconfident outputs from the model and empirically demonstrate that MixDiff consistently enhances the OOD detection performance on various datasets in vision and text domains. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2408.10107v1-abstract-full').style.display = 'none'; document.getElementById('2408.10107v1-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> 19 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to European Conference on Artificial Intelligence (ECAI) 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.13942">arXiv:2407.13942</a> <span> [<a href="https://arxiv.org/pdf/2407.13942">pdf</a>, <a href="https://arxiv.org/format/2407.13942">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computers and Society">cs.CY</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="Social and Information Networks">cs.SI</span> </div> </div> <p class="title is-5 mathjax"> Harmful Suicide Content Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Park%2C+K">Kyumin Park</a>, <a href="/search/?searchtype=author&query=Baik%2C+M+J">Myung Jae Baik</a>, <a href="/search/?searchtype=author&query=Hwang%2C+Y">YeongJun Hwang</a>, <a href="/search/?searchtype=author&query=Shin%2C+Y">Yen Shin</a>, <a href="/search/?searchtype=author&query=Lee%2C+H">HoJae Lee</a>, <a href="/search/?searchtype=author&query=Lee%2C+R">Ruda Lee</a>, <a href="/search/?searchtype=author&query=Lee%2C+S+M">Sang Min Lee</a>, <a href="/search/?searchtype=author&query=Sun%2C+J+Y+H">Je Young Hannah Sun</a>, <a href="/search/?searchtype=author&query=Lee%2C+A+R">Ah Rah Lee</a>, <a href="/search/?searchtype=author&query=Yoon%2C+S+Y">Si Yeun Yoon</a>, <a href="/search/?searchtype=author&query=Lee%2C+D">Dong-ho Lee</a>, <a href="/search/?searchtype=author&query=Moon%2C+J">Jihyung Moon</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a>, <a href="/search/?searchtype=author&query=Cho%2C+K">Kyunghyun Cho</a>, <a href="/search/?searchtype=author&query=Paik%2C+J">Jong-Woo Paik</a>, <a href="/search/?searchtype=author&query=Park%2C+S">Sungjoon Park</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.13942v1-abstract-short" style="display: inline;"> Harmful suicide content on the Internet is a significant risk factor inducing suicidal thoughts and behaviors among vulnerable populations. Despite global efforts, existing resources are insufficient, specifically in high-risk regions like the Republic of Korea. Current research mainly focuses on understanding negative effects of such content or suicide risk in individuals, rather than on automati… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.13942v1-abstract-full').style.display = 'inline'; document.getElementById('2407.13942v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.13942v1-abstract-full" style="display: none;"> Harmful suicide content on the Internet is a significant risk factor inducing suicidal thoughts and behaviors among vulnerable populations. Despite global efforts, existing resources are insufficient, specifically in high-risk regions like the Republic of Korea. Current research mainly focuses on understanding negative effects of such content or suicide risk in individuals, rather than on automatically detecting the harmfulness of content. To fill this gap, we introduce a harmful suicide content detection task for classifying online suicide content into five harmfulness levels. We develop a multi-modal benchmark and a task description document in collaboration with medical professionals, and leverage large language models (LLMs) to explore efficient methods for moderating such content. Our contributions include proposing a novel detection task, a multi-modal Korean benchmark with expert annotations, and suggesting strategies using LLMs to detect illegal and harmful content. Owing to the potential harm involved, we publicize our implementations and benchmark, incorporating an ethical verification process. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.13942v1-abstract-full').style.display = 'none'; document.getElementById('2407.13942v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 2 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">30 pages, 7 figures</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.07413">arXiv:2407.07413</a> <span> [<a href="https://arxiv.org/pdf/2407.07413">pdf</a>, <a href="https://arxiv.org/format/2407.07413">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"> KpopMT: Translation Dataset with Terminology for Kpop Fandom </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kim%2C+J">JiWoo Kim</a>, <a href="/search/?searchtype=author&query=Kim%2C+Y">Yunsu Kim</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.07413v1-abstract-short" style="display: inline;"> While machines learn from existing corpora, humans have the unique capability to establish and accept new language systems. This makes human form unique language systems within social groups. Aligning with this, we focus on a gap remaining in addressing translation challenges within social groups, where in-group members utilize unique terminologies. We propose KpopMT dataset, which aims to fill th… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.07413v1-abstract-full').style.display = 'inline'; document.getElementById('2407.07413v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.07413v1-abstract-full" style="display: none;"> While machines learn from existing corpora, humans have the unique capability to establish and accept new language systems. This makes human form unique language systems within social groups. Aligning with this, we focus on a gap remaining in addressing translation challenges within social groups, where in-group members utilize unique terminologies. We propose KpopMT dataset, which aims to fill this gap by enabling precise terminology translation, choosing Kpop fandom as an initiative for social groups given its global popularity. Expert translators provide 1k English translations for Korean posts and comments, each annotated with specific terminology within social groups' language systems. We evaluate existing translation systems including GPT models on KpopMT to identify their failure cases. Results show overall low scores, underscoring the challenges of reflecting group-specific terminologies and styles in translation. We make KpopMT publicly available. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.07413v1-abstract-full').style.display = 'none'; document.getElementById('2407.07413v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 10 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">accepted to LoresMT 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2407.02736">arXiv:2407.02736</a> <span> [<a href="https://arxiv.org/pdf/2407.02736">pdf</a>, <a href="https://arxiv.org/format/2407.02736">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"> MentalAgora: A Gateway to Advanced Personalized Care in Mental Health through Multi-Agent Debating and Attribute Control </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Lee%2C+Y">Yeonji Lee</a>, <a href="/search/?searchtype=author&query=Park%2C+S">Sangjun Park</a>, <a href="/search/?searchtype=author&query=Cho%2C+K">Kyunghyun Cho</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2407.02736v1-abstract-short" style="display: inline;"> As mental health issues globally escalate, there is a tremendous need for advanced digital support systems. We introduce MentalAgora, a novel framework employing large language models enhanced by interaction between multiple agents for tailored mental health support. This framework operates through three stages: strategic debating, tailored counselor creation, and response generation, enabling the… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.02736v1-abstract-full').style.display = 'inline'; document.getElementById('2407.02736v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2407.02736v1-abstract-full" style="display: none;"> As mental health issues globally escalate, there is a tremendous need for advanced digital support systems. We introduce MentalAgora, a novel framework employing large language models enhanced by interaction between multiple agents for tailored mental health support. This framework operates through three stages: strategic debating, tailored counselor creation, and response generation, enabling the dynamic customization of responses based on individual user preferences and therapeutic needs. We conduct experiments utilizing a high-quality evaluation dataset TherapyTalk crafted with mental health professionals, shwoing that MentalAgora generates expert-aligned and user preference-enhanced responses. Our evaluations, including experiments and user studies, demonstrate that MentalAgora aligns with professional standards and effectively meets user preferences, setting a new benchmark for digital mental health interventions. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2407.02736v1-abstract-full').style.display = 'none'; document.getElementById('2407.02736v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 2 July, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.16042">arXiv:2406.16042</a> <span> [<a href="https://arxiv.org/pdf/2406.16042">pdf</a>, <a href="https://arxiv.org/format/2406.16042">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"> Pose-dIVE: Pose-Diversified Augmentation with Diffusion Model for Person Re-Identification </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kim%2C+I+H">In猫s Hyeonsu Kim</a>, <a href="/search/?searchtype=author&query=Lee%2C+J">JoungBin Lee</a>, <a href="/search/?searchtype=author&query=Jin%2C+W">Woojeong Jin</a>, <a href="/search/?searchtype=author&query=Son%2C+S">Soowon Son</a>, <a href="/search/?searchtype=author&query=Cho%2C+K">Kyusun Cho</a>, <a href="/search/?searchtype=author&query=Seo%2C+J">Junyoung Seo</a>, <a href="/search/?searchtype=author&query=Kwak%2C+M">Min-Seop Kwak</a>, <a href="/search/?searchtype=author&query=Cho%2C+S">Seokju Cho</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">JeongYeol Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+B">Byeongwon Lee</a>, <a href="/search/?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="2406.16042v2-abstract-short" style="display: inline;"> Person re-identification (Re-ID) often faces challenges due to variations in human poses and camera viewpoints, which significantly affect the appearance of individuals across images. Existing datasets frequently lack diversity and scalability in these aspects, hindering the generalization of Re-ID models to new camera systems. We propose Pose-dIVE, a novel data augmentation approach that incorpor… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16042v2-abstract-full').style.display = 'inline'; document.getElementById('2406.16042v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.16042v2-abstract-full" style="display: none;"> Person re-identification (Re-ID) often faces challenges due to variations in human poses and camera viewpoints, which significantly affect the appearance of individuals across images. Existing datasets frequently lack diversity and scalability in these aspects, hindering the generalization of Re-ID models to new camera systems. We propose Pose-dIVE, a novel data augmentation approach that incorporates sparse and underrepresented human pose and camera viewpoint examples into the training data, addressing the limited diversity in the original training data distribution. Our objective is to augment the training dataset to enable existing Re-ID models to learn features unbiased by human pose and camera viewpoint variations. To achieve this, we leverage the knowledge of pre-trained large-scale diffusion models. By conditioning the diffusion model on both the human pose and camera viewpoint concurrently through the SMPL model, we generate training data with diverse human poses and camera viewpoints. Experimental results demonstrate the effectiveness of our method in addressing human pose bias and enhancing the generalizability of Re-ID models compared to other data augmentation-based Re-ID approaches. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16042v2-abstract-full').style.display = 'none'; document.getElementById('2406.16042v2-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 October, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 23 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.16013">arXiv:2406.16013</a> <span> [<a href="https://arxiv.org/pdf/2406.16013">pdf</a>, <a href="https://arxiv.org/format/2406.16013">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> </div> </div> <p class="title is-5 mathjax"> Database-Augmented Query Representation for Information Retrieval </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Cho%2C+S">Sukmin Cho</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a>, <a href="/search/?searchtype=author&query=Park%2C+J+C">Jong C. Park</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2406.16013v1-abstract-short" style="display: inline;"> Information retrieval models that aim to search for the documents relevant to the given query have shown many successes, which have been applied to diverse tasks. However, the query provided by the user is oftentimes very short, which challenges the retrievers to correctly fetch relevant documents. To tackle this, existing studies have proposed expanding the query with a couple of additional (user… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16013v1-abstract-full').style.display = 'inline'; document.getElementById('2406.16013v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.16013v1-abstract-full" style="display: none;"> Information retrieval models that aim to search for the documents relevant to the given query have shown many successes, which have been applied to diverse tasks. However, the query provided by the user is oftentimes very short, which challenges the retrievers to correctly fetch relevant documents. To tackle this, existing studies have proposed expanding the query with a couple of additional (user-related) features related to the query. Yet, they may be suboptimal to effectively augment the query, though there is plenty of information available to augment it in a relational database. Motivated by this, we present a novel retrieval framework called Database-Augmented Query representation (DAQu), which augments the original query with various (query-related) metadata across multiple tables. In addition, as the number of features in the metadata can be very large and there is no order among them, we encode them with our graph-based set encoding strategy, which considers hierarchies of features in the database without order. We validate DAQu in diverse retrieval scenarios that can incorporate metadata from the relational database, demonstrating that ours significantly enhances overall retrieval performance, compared to existing query augmentation methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.16013v1-abstract-full').style.display = 'none'; document.getElementById('2406.16013v1-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 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.10725">arXiv:2406.10725</a> <span> [<a href="https://arxiv.org/pdf/2406.10725">pdf</a>, <a href="https://arxiv.org/format/2406.10725">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Metric Geometry">math.MG</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Optimization and Control">math.OC</span> </div> </div> <p class="title is-5 mathjax"> A Conditional Upper Bound for the Moving Sofa Problem </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jineon Baek</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2406.10725v1-abstract-short" style="display: inline;"> The moving sofa problem asks for the connected shape with the largest area $渭_{\text{max}}$ that can move around the right-angled corner of a hallway $L$ with unit width. The best bounds currently known on $渭_{\max}$ are summarized as $2.2195\ldots \leq 渭_{\max} \leq 2.37$. The lower bound $2.2195\ldots \leq 渭_{\max}$ comes from Gerver's sofa $S_G$ of area $渭_G := 2.2195\ldots$. The upper bound… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.10725v1-abstract-full').style.display = 'inline'; document.getElementById('2406.10725v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.10725v1-abstract-full" style="display: none;"> The moving sofa problem asks for the connected shape with the largest area $渭_{\text{max}}$ that can move around the right-angled corner of a hallway $L$ with unit width. The best bounds currently known on $渭_{\max}$ are summarized as $2.2195\ldots \leq 渭_{\max} \leq 2.37$. The lower bound $2.2195\ldots \leq 渭_{\max}$ comes from Gerver's sofa $S_G$ of area $渭_G := 2.2195\ldots$. The upper bound $渭_{\max} \leq 2.37$ was proved by Kallus and Romik using extensive computer assistance. It is conjectured that the equality $渭_{\max} = 渭_G$ holds at the lower bound. We develop a new approach to the moving sofa problem by approximating it as an infinite-dimensional convex quadratic optimization problem. The problem is then explicitly solved using a calculus of variation based on the Brunn-Minkowski theory. Consequently, we prove that any moving sofa satisfying a property named the injectivity condition has an area of at most $1 + 蟺^2/8 = 2.2337\dots$. The new conditional bound does not rely on any computer assistance, yet it is much closer to the lower bound $2.2195\ldots$ of Gerver than the computer-assisted upper bound $2.37$ of Kallus and Romik. Gerver's sofa $S_G$, the conjectured optimum, satisfies the injectivity condition in particular. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.10725v1-abstract-full').style.display = 'none'; document.getElementById('2406.10725v1-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 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">53 pages, 9 figures</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">MSC Class:</span> 49Q10; 52A10; 52A41 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.06929">arXiv:2406.06929</a> <span> [<a href="https://arxiv.org/pdf/2406.06929">pdf</a>, <a href="https://arxiv.org/ps/2406.06929">ps</a>, <a href="https://arxiv.org/format/2406.06929">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Science and Game Theory">cs.GT</span> </div> </div> <p class="title is-5 mathjax"> Social Learning with Bounded Rationality: Negative Reviews Persist under Newest First </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jackie Baek</a>, <a href="/search/?searchtype=author&query=Dinev%2C+A">Atanas Dinev</a>, <a href="/search/?searchtype=author&query=Lykouris%2C+T">Thodoris Lykouris</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2406.06929v2-abstract-short" style="display: inline;"> We study a model of social learning from reviews where customers are computationally limited and make purchases based on reading only the first few reviews displayed by the platform. Under this bounded rationality, we establish that the review ordering policy can have a significant impact. In particular, the popular Newest First ordering induces a negative review to persist as the most recent revi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.06929v2-abstract-full').style.display = 'inline'; document.getElementById('2406.06929v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.06929v2-abstract-full" style="display: none;"> We study a model of social learning from reviews where customers are computationally limited and make purchases based on reading only the first few reviews displayed by the platform. Under this bounded rationality, we establish that the review ordering policy can have a significant impact. In particular, the popular Newest First ordering induces a negative review to persist as the most recent review longer than a positive review. This phenomenon, which we term the Cost of Newest First, can make the long-term revenue unboundedly lower than a counterpart where reviews are exogenously drawn for each customer. We show that the impact of the Cost of Newest First can be mitigated under dynamic pricing, which allows the price to depend on the set of displayed reviews. Under the optimal dynamic pricing policy, the revenue loss is at most a factor of 2. On the way, we identify a structural property for this optimal dynamic pricing: the prices should ensure that the probability of a purchase is always the same, regardless of the state of reviews. We also study an extension of the model where customers put more weight on more recent reviews (and discount older reviews based on their time of posting), and we show that Newest First is still not the optimal ordering policy if customers discount slowly. Lastly, we corroborate our theoretical findings using a real-world review dataset. We find that the average rating of the first page of reviews is statistically significantly smaller than the overall average rating, which is in line with our theoretical results. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.06929v2-abstract-full').style.display = 'none'; document.getElementById('2406.06929v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 22 August, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 11 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">An extended abstract appeared at the Twenty-Fifth ACM Conference on Economics and Computation (EC 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/2406.06793">arXiv:2406.06793</a> <span> [<a href="https://arxiv.org/pdf/2406.06793">pdf</a>, <a href="https://arxiv.org/format/2406.06793">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"> PlanDQ: Hierarchical Plan Orchestration via D-Conductor and Q-Performer </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Chen%2C+C">Chang Chen</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Junyeob Baek</a>, <a href="/search/?searchtype=author&query=Deng%2C+F">Fei Deng</a>, <a href="/search/?searchtype=author&query=Kawaguchi%2C+K">Kenji Kawaguchi</a>, <a href="/search/?searchtype=author&query=Gulcehre%2C+C">Caglar Gulcehre</a>, <a href="/search/?searchtype=author&query=Ahn%2C+S">Sungjin Ahn</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2406.06793v1-abstract-short" style="display: inline;"> Despite the recent advancements in offline RL, no unified algorithm could achieve superior performance across a broad range of tasks. Offline \textit{value function learning}, in particular, struggles with sparse-reward, long-horizon tasks due to the difficulty of solving credit assignment and extrapolation errors that accumulates as the horizon of the task grows.~On the other hand, models that ca… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.06793v1-abstract-full').style.display = 'inline'; document.getElementById('2406.06793v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.06793v1-abstract-full" style="display: none;"> Despite the recent advancements in offline RL, no unified algorithm could achieve superior performance across a broad range of tasks. Offline \textit{value function learning}, in particular, struggles with sparse-reward, long-horizon tasks due to the difficulty of solving credit assignment and extrapolation errors that accumulates as the horizon of the task grows.~On the other hand, models that can perform well in long-horizon tasks are designed specifically for goal-conditioned tasks, which commonly perform worse than value function learning methods on short-horizon, dense-reward scenarios. To bridge this gap, we propose a hierarchical planner designed for offline RL called PlanDQ. PlanDQ incorporates a diffusion-based planner at the high level, named D-Conductor, which guides the low-level policy through sub-goals. At the low level, we used a Q-learning based approach called the Q-Performer to accomplish these sub-goals. Our experimental results suggest that PlanDQ can achieve superior or competitive performance on D4RL continuous control benchmark tasks as well as AntMaze, Kitchen, and Calvin as long-horizon tasks. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.06793v1-abstract-full').style.display = 'none'; document.getElementById('2406.06793v1-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 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.05967">arXiv:2406.05967</a> <span> [<a href="https://arxiv.org/pdf/2406.05967">pdf</a>, <a href="https://arxiv.org/format/2406.05967">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="Computation and Language">cs.CL</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"> CVQA: Culturally-diverse Multilingual Visual Question Answering Benchmark </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Romero%2C+D">David Romero</a>, <a href="/search/?searchtype=author&query=Lyu%2C+C">Chenyang Lyu</a>, <a href="/search/?searchtype=author&query=Wibowo%2C+H+A">Haryo Akbarianto Wibowo</a>, <a href="/search/?searchtype=author&query=Lynn%2C+T">Teresa Lynn</a>, <a href="/search/?searchtype=author&query=Hamed%2C+I">Injy Hamed</a>, <a href="/search/?searchtype=author&query=Kishore%2C+A+N">Aditya Nanda Kishore</a>, <a href="/search/?searchtype=author&query=Mandal%2C+A">Aishik Mandal</a>, <a href="/search/?searchtype=author&query=Dragonetti%2C+A">Alina Dragonetti</a>, <a href="/search/?searchtype=author&query=Abzaliev%2C+A">Artem Abzaliev</a>, <a href="/search/?searchtype=author&query=Tonja%2C+A+L">Atnafu Lambebo Tonja</a>, <a href="/search/?searchtype=author&query=Balcha%2C+B+F">Bontu Fufa Balcha</a>, <a href="/search/?searchtype=author&query=Whitehouse%2C+C">Chenxi Whitehouse</a>, <a href="/search/?searchtype=author&query=Salamea%2C+C">Christian Salamea</a>, <a href="/search/?searchtype=author&query=Velasco%2C+D+J">Dan John Velasco</a>, <a href="/search/?searchtype=author&query=Adelani%2C+D+I">David Ifeoluwa Adelani</a>, <a href="/search/?searchtype=author&query=Meur%2C+D+L">David Le Meur</a>, <a href="/search/?searchtype=author&query=Villa-Cueva%2C+E">Emilio Villa-Cueva</a>, <a href="/search/?searchtype=author&query=Koto%2C+F">Fajri Koto</a>, <a href="/search/?searchtype=author&query=Farooqui%2C+F">Fauzan Farooqui</a>, <a href="/search/?searchtype=author&query=Belcavello%2C+F">Frederico Belcavello</a>, <a href="/search/?searchtype=author&query=Batnasan%2C+G">Ganzorig Batnasan</a>, <a href="/search/?searchtype=author&query=Vallejo%2C+G">Gisela Vallejo</a>, <a href="/search/?searchtype=author&query=Caulfield%2C+G">Grainne Caulfield</a>, <a href="/search/?searchtype=author&query=Ivetta%2C+G">Guido Ivetta</a>, <a href="/search/?searchtype=author&query=Song%2C+H">Haiyue Song</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="2406.05967v2-abstract-short" style="display: inline;"> Visual Question Answering (VQA) is an important task in multimodal AI, and it is often used to test the ability of vision-language models to understand and reason on knowledge present in both visual and textual data. However, most of the current VQA models use datasets that are primarily focused on English and a few major world languages, with images that are typically Western-centric. While recen… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.05967v2-abstract-full').style.display = 'inline'; document.getElementById('2406.05967v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.05967v2-abstract-full" style="display: none;"> Visual Question Answering (VQA) is an important task in multimodal AI, and it is often used to test the ability of vision-language models to understand and reason on knowledge present in both visual and textual data. However, most of the current VQA models use datasets that are primarily focused on English and a few major world languages, with images that are typically Western-centric. While recent efforts have tried to increase the number of languages covered on VQA datasets, they still lack diversity in low-resource languages. More importantly, although these datasets often extend their linguistic range via translation or some other approaches, they usually keep images the same, resulting in narrow cultural representation. To address these limitations, we construct CVQA, a new Culturally-diverse multilingual Visual Question Answering benchmark, designed to cover a rich set of languages and cultures, where we engage native speakers and cultural experts in the data collection process. As a result, CVQA includes culturally-driven images and questions from across 30 countries on four continents, covering 31 languages with 13 scripts, providing a total of 10k questions. We then benchmark several Multimodal Large Language Models (MLLMs) on CVQA, and show that the dataset is challenging for the current state-of-the-art models. This benchmark can serve as a probing evaluation suite for assessing the cultural capability and bias of multimodal models and hopefully encourage more research efforts toward increasing cultural awareness and linguistic diversity in this field. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.05967v2-abstract-full').style.display = 'none'; document.getElementById('2406.05967v2-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 9 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2406.05761">arXiv:2406.05761</a> <span> [<a href="https://arxiv.org/pdf/2406.05761">pdf</a>, <a href="https://arxiv.org/format/2406.05761">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"> The BiGGen Bench: A Principled Benchmark for Fine-grained Evaluation of Language Models with Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kim%2C+S">Seungone Kim</a>, <a href="/search/?searchtype=author&query=Suk%2C+J">Juyoung Suk</a>, <a href="/search/?searchtype=author&query=Cho%2C+J+Y">Ji Yong Cho</a>, <a href="/search/?searchtype=author&query=Longpre%2C+S">Shayne Longpre</a>, <a href="/search/?searchtype=author&query=Kim%2C+C">Chaeeun Kim</a>, <a href="/search/?searchtype=author&query=Yoon%2C+D">Dongkeun Yoon</a>, <a href="/search/?searchtype=author&query=Son%2C+G">Guijin Son</a>, <a href="/search/?searchtype=author&query=Cho%2C+Y">Yejin Cho</a>, <a href="/search/?searchtype=author&query=Shafayat%2C+S">Sheikh Shafayat</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Park%2C+S+H">Sue Hyun Park</a>, <a href="/search/?searchtype=author&query=Hwang%2C+H">Hyeonbin Hwang</a>, <a href="/search/?searchtype=author&query=Jo%2C+J">Jinkyung Jo</a>, <a href="/search/?searchtype=author&query=Cho%2C+H">Hyowon Cho</a>, <a href="/search/?searchtype=author&query=Shin%2C+H">Haebin Shin</a>, <a href="/search/?searchtype=author&query=Lee%2C+S">Seongyun Lee</a>, <a href="/search/?searchtype=author&query=Oh%2C+H">Hanseok Oh</a>, <a href="/search/?searchtype=author&query=Lee%2C+N">Noah Lee</a>, <a href="/search/?searchtype=author&query=Ho%2C+N">Namgyu Ho</a>, <a href="/search/?searchtype=author&query=Joo%2C+S+J">Se June Joo</a>, <a href="/search/?searchtype=author&query=Ko%2C+M">Miyoung Ko</a>, <a href="/search/?searchtype=author&query=Lee%2C+Y">Yoonjoo Lee</a>, <a href="/search/?searchtype=author&query=Chae%2C+H">Hyungjoo Chae</a>, <a href="/search/?searchtype=author&query=Shin%2C+J">Jamin Shin</a>, <a href="/search/?searchtype=author&query=Jang%2C+J">Joel Jang</a> , et al. (7 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="2406.05761v1-abstract-short" style="display: inline;"> As language models (LMs) become capable of handling a wide range of tasks, their evaluation is becoming as challenging as their development. Most generation benchmarks currently assess LMs using abstract evaluation criteria like helpfulness and harmlessness, which often lack the flexibility and granularity of human assessment. Additionally, these benchmarks tend to focus disproportionately on spec… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.05761v1-abstract-full').style.display = 'inline'; document.getElementById('2406.05761v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2406.05761v1-abstract-full" style="display: none;"> As language models (LMs) become capable of handling a wide range of tasks, their evaluation is becoming as challenging as their development. Most generation benchmarks currently assess LMs using abstract evaluation criteria like helpfulness and harmlessness, which often lack the flexibility and granularity of human assessment. Additionally, these benchmarks tend to focus disproportionately on specific capabilities such as instruction following, leading to coverage bias. To overcome these limitations, we introduce the BiGGen Bench, a principled generation benchmark designed to thoroughly evaluate nine distinct capabilities of LMs across 77 diverse tasks. A key feature of the BiGGen Bench is its use of instance-specific evaluation criteria, closely mirroring the nuanced discernment of human evaluation. We apply this benchmark to assess 103 frontier LMs using five evaluator LMs. Our code, data, and evaluation results are all publicly available at https://github.com/prometheus-eval/prometheus-eval/tree/main/BiGGen-Bench. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2406.05761v1-abstract-full').style.display = 'none'; document.getElementById('2406.05761v1-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 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Work in Progress</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2405.01860">arXiv:2405.01860</a> <span> [<a href="https://arxiv.org/pdf/2405.01860">pdf</a>, <a href="https://arxiv.org/ps/2405.01860">ps</a>, <a href="https://arxiv.org/format/2405.01860">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="General Topology">math.GN</span> </div> </div> <p class="title is-5 mathjax"> Characterizing Lipschitz images of injective metric spaces </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=B%C4%85k%2C+J">Judyta B膮k</a>, <a href="/search/?searchtype=author&query=Banakh%2C+T">Taras Banakh</a>, <a href="/search/?searchtype=author&query=Garbuli%C5%84ska-W%C4%99grzyn%2C+J">Joanna Garbuli艅ska-W臋grzyn</a>, <a href="/search/?searchtype=author&query=Nowak%2C+M">Magdalena Nowak</a>, <a href="/search/?searchtype=author&query=Pop%C5%82awski%2C+M">Micha艂 Pop艂awski</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2405.01860v2-abstract-short" style="display: inline;"> A metric space $X$ is {\em injective} if every non-expanding map $f:B\to X$ defined on a subspace $B$ of a metric space $A$ can be extended to a non-expanding map $\bar f:A\to X$. We prove that a metric space $X$ is a Lipschitz image of an injective metric space if and only if $X$ is Lipschitz connected in the sense that for every points $x,y\in X$, there exists a Lipschitz map $f:[0,1]\to X$ such… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.01860v2-abstract-full').style.display = 'inline'; document.getElementById('2405.01860v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2405.01860v2-abstract-full" style="display: none;"> A metric space $X$ is {\em injective} if every non-expanding map $f:B\to X$ defined on a subspace $B$ of a metric space $A$ can be extended to a non-expanding map $\bar f:A\to X$. We prove that a metric space $X$ is a Lipschitz image of an injective metric space if and only if $X$ is Lipschitz connected in the sense that for every points $x,y\in X$, there exists a Lipschitz map $f:[0,1]\to X$ such that $f(0)=x$ and $f(1)=y$. In this case the metric space $X$ carries a well-defined intrinsic metric. A metric space $X$ is a Lipschitz image of a compact injective metric space if and only if $X$ is compact, Lipschitz connected and its intrinsic metric is totally bounded. A metric space $X$ is a Lipschitz image of a separable injective metric space if and only if $X$ is a Lipschitz image of the Urysohn universal metric space if and only if $X$ is analytic, Lipschitz connected and its intrinsic metric is separable. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2405.01860v2-abstract-full').style.display = 'none'; document.getElementById('2405.01860v2-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 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 3 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> May 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">MSC Class:</span> Primary:54E35; 54E40; Secondary: 51F30; 54C55; 54E45; 54E50; 54F15 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2404.07738">arXiv:2404.07738</a> <span> [<a href="https://arxiv.org/pdf/2404.07738">pdf</a>, <a href="https://arxiv.org/format/2404.07738">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"> ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Jauhar%2C+S+K">Sujay Kumar Jauhar</a>, <a href="/search/?searchtype=author&query=Cucerzan%2C+S">Silviu Cucerzan</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2404.07738v1-abstract-short" style="display: inline;"> Scientific Research, vital for improving human life, is hindered by its inherent complexity, slow pace, and the need for specialized experts. To enhance its productivity, we propose a ResearchAgent, a large language model-powered research idea writing agent, which automatically generates problems, methods, and experiment designs while iteratively refining them based on scientific literature. Speci… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.07738v1-abstract-full').style.display = 'inline'; document.getElementById('2404.07738v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2404.07738v1-abstract-full" style="display: none;"> Scientific Research, vital for improving human life, is hindered by its inherent complexity, slow pace, and the need for specialized experts. To enhance its productivity, we propose a ResearchAgent, a large language model-powered research idea writing agent, which automatically generates problems, methods, and experiment designs while iteratively refining them based on scientific literature. Specifically, starting with a core paper as the primary focus to generate ideas, our ResearchAgent is augmented not only with relevant publications through connecting information over an academic graph but also entities retrieved from an entity-centric knowledge store based on their underlying concepts, mined and shared across numerous papers. In addition, mirroring the human approach to iteratively improving ideas with peer discussions, we leverage multiple ReviewingAgents that provide reviews and feedback iteratively. Further, they are instantiated with human preference-aligned large language models whose criteria for evaluation are derived from actual human judgments. We experimentally validate our ResearchAgent on scientific publications across multiple disciplines, showcasing its effectiveness in generating novel, clear, and valid research ideas based on human and model-based evaluation results. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.07738v1-abstract-full').style.display = 'none'; document.getElementById('2404.07738v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 11 April, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> April 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2404.02949">arXiv:2404.02949</a> <span> [<a href="https://arxiv.org/pdf/2404.02949">pdf</a>, <a href="https://arxiv.org/format/2404.02949">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"> The SaTML '24 CNN Interpretability Competition: New Innovations for Concept-Level Interpretability </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Casper%2C+S">Stephen Casper</a>, <a href="/search/?searchtype=author&query=Yun%2C+J">Jieun Yun</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Joonhyuk Baek</a>, <a href="/search/?searchtype=author&query=Jung%2C+Y">Yeseong Jung</a>, <a href="/search/?searchtype=author&query=Kim%2C+M">Minhwan Kim</a>, <a href="/search/?searchtype=author&query=Kwon%2C+K">Kiwan Kwon</a>, <a href="/search/?searchtype=author&query=Park%2C+S">Saerom Park</a>, <a href="/search/?searchtype=author&query=Moore%2C+H">Hayden Moore</a>, <a href="/search/?searchtype=author&query=Shriver%2C+D">David Shriver</a>, <a href="/search/?searchtype=author&query=Connor%2C+M">Marissa Connor</a>, <a href="/search/?searchtype=author&query=Grimes%2C+K">Keltin Grimes</a>, <a href="/search/?searchtype=author&query=Nicolson%2C+A">Angus Nicolson</a>, <a href="/search/?searchtype=author&query=Tagade%2C+A">Arush Tagade</a>, <a href="/search/?searchtype=author&query=Rumbelow%2C+J">Jessica Rumbelow</a>, <a href="/search/?searchtype=author&query=Nguyen%2C+H+M">Hieu Minh Nguyen</a>, <a href="/search/?searchtype=author&query=Hadfield-Menell%2C+D">Dylan Hadfield-Menell</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2404.02949v1-abstract-short" style="display: inline;"> Interpretability techniques are valuable for helping humans understand and oversee AI systems. The SaTML 2024 CNN Interpretability Competition solicited novel methods for studying convolutional neural networks (CNNs) at the ImageNet scale. The objective of the competition was to help human crowd-workers identify trojans in CNNs. This report showcases the methods and results of four featured compet… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.02949v1-abstract-full').style.display = 'inline'; document.getElementById('2404.02949v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2404.02949v1-abstract-full" style="display: none;"> Interpretability techniques are valuable for helping humans understand and oversee AI systems. The SaTML 2024 CNN Interpretability Competition solicited novel methods for studying convolutional neural networks (CNNs) at the ImageNet scale. The objective of the competition was to help human crowd-workers identify trojans in CNNs. This report showcases the methods and results of four featured competition entries. It remains challenging to help humans reliably diagnose trojans via interpretability tools. However, the competition's entries have contributed new techniques and set a new record on the benchmark from Casper et al., 2023. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.02949v1-abstract-full').style.display = 'none'; document.getElementById('2404.02949v1-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 April, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> April 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">Competition for SaTML 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/2404.02143">arXiv:2404.02143</a> <span> [<a href="https://arxiv.org/pdf/2404.02143">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Medical Physics">physics.med-ph</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Signal Processing">eess.SP</span> </div> </div> <p class="title is-5 mathjax"> Multiparametric quantification and visualization of liver fat using ultrasound </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jihye Baek</a>, <a href="/search/?searchtype=author&query=Kaffas%2C+A+E">Ahmed El Kaffas</a>, <a href="/search/?searchtype=author&query=Kamaya%2C+A">Aya Kamaya</a>, <a href="/search/?searchtype=author&query=Hoyt%2C+K">Kenneth Hoyt</a>, <a href="/search/?searchtype=author&query=Parker%2C+K+J">Kevin J. Parker</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2404.02143v1-abstract-short" style="display: inline;"> Objectives- Several ultrasound measures have shown promise for assessment of steatosis compared to traditional B-scan, however clinicians may be required to integrate information across the parameters. Here, we propose an integrated multiparametric approach, enabling simple clinical assessment of key information from combined ultrasound parameters. Methods- We have measured 13 parameters related t… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.02143v1-abstract-full').style.display = 'inline'; document.getElementById('2404.02143v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2404.02143v1-abstract-full" style="display: none;"> Objectives- Several ultrasound measures have shown promise for assessment of steatosis compared to traditional B-scan, however clinicians may be required to integrate information across the parameters. Here, we propose an integrated multiparametric approach, enabling simple clinical assessment of key information from combined ultrasound parameters. Methods- We have measured 13 parameters related to ultrasound and shear wave elastography. These were measured in 30 human subjects under a study of liver fat. The 13 individual measures are assessed for their predictive value using independent magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) measurements as a reference standard. In addition, a comprehensive and fine-grain analysis is made of all possible combinations of sub-sets of these parameters to determine if any subset can be efficiently combined to predict fat fraction. Results- We found that as few as four key parameters related to ultrasound propagation are sufficient to generate a linear multiparametric parameter with a correlation against MRI-PDFF values of greater than 0.93. This optimal combination was found to have a classification area under the curve (AUC) approaching 1.0 when applying a threshold for separating steatosis grade zero from higher classes. Furthermore, a strategy is developed for applying local estimates of fat content as a color overlay to produce a visual impression of the extent and distribution of fat within the liver. Conclusion- In principle, this approach can be applied to most clinical ultrasound systems to provide the clinician and patient with a rapid and inexpensive estimate of liver fat content. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2404.02143v1-abstract-full').style.display = 'none'; document.getElementById('2404.02143v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 2 April, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> April 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2403.14403">arXiv:2403.14403</a> <span> [<a href="https://arxiv.org/pdf/2403.14403">pdf</a>, <a href="https://arxiv.org/format/2403.14403">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"> Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexity </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Cho%2C+S">Sukmin Cho</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a>, <a href="/search/?searchtype=author&query=Park%2C+J+C">Jong C. Park</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2403.14403v2-abstract-short" style="display: inline;"> Retrieval-Augmented Large Language Models (LLMs), which incorporate the non-parametric knowledge from external knowledge bases into LLMs, have emerged as a promising approach to enhancing response accuracy in several tasks, such as Question-Answering (QA). However, even though there are various approaches dealing with queries of different complexities, they either handle simple queries with unnece… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.14403v2-abstract-full').style.display = 'inline'; document.getElementById('2403.14403v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2403.14403v2-abstract-full" style="display: none;"> Retrieval-Augmented Large Language Models (LLMs), which incorporate the non-parametric knowledge from external knowledge bases into LLMs, have emerged as a promising approach to enhancing response accuracy in several tasks, such as Question-Answering (QA). However, even though there are various approaches dealing with queries of different complexities, they either handle simple queries with unnecessary computational overhead or fail to adequately address complex multi-step queries; yet, not all user requests fall into only one of the simple or complex categories. In this work, we propose a novel adaptive QA framework, that can dynamically select the most suitable strategy for (retrieval-augmented) LLMs from the simplest to the most sophisticated ones based on the query complexity. Also, this selection process is operationalized with a classifier, which is a smaller LM trained to predict the complexity level of incoming queries with automatically collected labels, obtained from actual predicted outcomes of models and inherent inductive biases in datasets. This approach offers a balanced strategy, seamlessly adapting between the iterative and single-step retrieval-augmented LLMs, as well as the no-retrieval methods, in response to a range of query complexities. We validate our model on a set of open-domain QA datasets, covering multiple query complexities, and show that ours enhances the overall efficiency and accuracy of QA systems, compared to relevant baselines including the adaptive retrieval approaches. Code is available at: https://github.com/starsuzi/Adaptive-RAG. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2403.14403v2-abstract-full').style.display = 'none'; document.getElementById('2403.14403v2-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 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 21 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> March 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">NAACL 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/2402.13482">arXiv:2402.13482</a> <span> [<a href="https://arxiv.org/pdf/2402.13482">pdf</a>, <a href="https://arxiv.org/format/2402.13482">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"> Retrieval-Augmented Data Augmentation for Low-Resource Domain Tasks </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Seo%2C+M">Minju Seo</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Thorne%2C+J">James Thorne</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.13482v1-abstract-short" style="display: inline;"> Despite large successes of recent language models on diverse tasks, they suffer from severe performance degeneration in low-resource settings with limited training data available. Many existing works tackle this problem by generating synthetic data from the training data and then training models on them, recently using Large Language Models (LLMs). However, in low-resource settings, the amount of… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.13482v1-abstract-full').style.display = 'inline'; document.getElementById('2402.13482v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.13482v1-abstract-full" style="display: none;"> Despite large successes of recent language models on diverse tasks, they suffer from severe performance degeneration in low-resource settings with limited training data available. Many existing works tackle this problem by generating synthetic data from the training data and then training models on them, recently using Large Language Models (LLMs). However, in low-resource settings, the amount of seed data samples to use for data augmentation is very small, which makes generated samples suboptimal and less diverse. To tackle this challenge, we propose a novel method that augments training data by incorporating a wealth of examples from other datasets, along with the given training data. Specifically, we first retrieve the relevant instances from other datasets, such as their input-output pairs or contexts, based on their similarities with the given seed data, and then prompt LLMs to generate new samples with the contextual information within and across the original and retrieved samples. This approach can ensure that the generated data is not only relevant but also more diverse than what could be achieved using the limited seed data alone. We validate our proposed Retrieval-Augmented Data Augmentation (RADA) framework on multiple datasets under low-resource settings of training and test-time data augmentation scenarios, on which it outperforms existing LLM-powered data augmentation baselines. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.13482v1-abstract-full').style.display = 'none'; document.getElementById('2402.13482v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2402.02778">arXiv:2402.02778</a> <span> [<a href="https://arxiv.org/pdf/2402.02778">pdf</a>, <a href="https://arxiv.org/format/2402.02778">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Astrophysics of Galaxies">astro-ph.GA</span> </div> </div> <p class="title is-5 mathjax"> Detection of extragalactic anomalous microwave emission in NGC 2903 using KVN single-dish observations </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Poojon%2C+P">Panomporn Poojon</a>, <a href="/search/?searchtype=author&query=Chung%2C+A">Aeree Chung</a>, <a href="/search/?searchtype=author&query=Hoang%2C+T">Thiem Hoang</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Junhyun Baek</a>, <a href="/search/?searchtype=author&query=Nakanishi%2C+H">Hiroyuki Nakanishi</a>, <a href="/search/?searchtype=author&query=Hirota%2C+T">Tomoya Hirota</a>, <a href="/search/?searchtype=author&query=Tsai%2C+C">Chao-Wei Tsai</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2402.02778v1-abstract-short" style="display: inline;"> We present the results of the single-dish observations using the Korean VLBI Network to search for anomalous microwave emission (AME) in nearby galaxies. The targets were selected from MApping the dense moLecular gAs in the sTrongest stAr-formiNg Galaxies (MALATANG), a legacy survey project of the James Clerk Maxwell Telescope. The MALATANG galaxies are good representatives of local galaxies with… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.02778v1-abstract-full').style.display = 'inline'; document.getElementById('2402.02778v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2402.02778v1-abstract-full" style="display: none;"> We present the results of the single-dish observations using the Korean VLBI Network to search for anomalous microwave emission (AME) in nearby galaxies. The targets were selected from MApping the dense moLecular gAs in the sTrongest stAr-formiNg Galaxies (MALATANG), a legacy survey project of the James Clerk Maxwell Telescope. The MALATANG galaxies are good representatives of local galaxies with enhanced nuclear activity associated with star formation and/or AGN, providing IR-bright galaxy samples; thus, they are good candidates for AME hosts. Combining with the ancillary data, we investigated the radio-IR spectral energy distribution (SED), while searching for the AME signals in five galaxies. The AME in NGC 2903 was well detected at a significant confidence level, whereas that in NGC 2146 and M82 was marginal. NGC 1068 and Arp 299 indicated no significant hints, and we provided the upper limits for the AME. The best-fit SED exhibited local peaks of the AME components at higher frequencies and with stronger peak fluxes than those in the previous studies. This suggested the origin of AME being denser environments such as molecular clouds or photodissociation regions rather than warm neutral/ionized medium as commonly suggested by previous studies. Further, our AME-detected targets were observed to exhibit higher specific star-formation rates than the other extragalactic AME hosts. Furthermore, AME favored starburst galaxies among our sample rather than AGN hosts. Consequently, this might imply that AGNs are excessively harsh environments for tiny dust to survive. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2402.02778v1-abstract-full').style.display = 'none'; document.getElementById('2402.02778v1-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, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">20 pages, 10 figures, accepted for publication in ApJ</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.10404">arXiv:2401.10404</a> <span> [<a href="https://arxiv.org/pdf/2401.10404">pdf</a>, <a href="https://arxiv.org/format/2401.10404">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"> Inflation with Diffusion: Efficient Temporal Adaptation for Text-to-Video Super-Resolution </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Yuan%2C+X">Xin Yuan</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinoo Baek</a>, <a href="/search/?searchtype=author&query=Xu%2C+K">Keyang Xu</a>, <a href="/search/?searchtype=author&query=Tov%2C+O">Omer Tov</a>, <a href="/search/?searchtype=author&query=Fei%2C+H">Hongliang Fei</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.10404v1-abstract-short" style="display: inline;"> We propose an efficient diffusion-based text-to-video super-resolution (SR) tuning approach that leverages the readily learned capacity of pixel level image diffusion model to capture spatial information for video generation. To accomplish this goal, we design an efficient architecture by inflating the weightings of the text-to-image SR model into our video generation framework. Additionally, we i… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.10404v1-abstract-full').style.display = 'inline'; document.getElementById('2401.10404v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.10404v1-abstract-full" style="display: none;"> We propose an efficient diffusion-based text-to-video super-resolution (SR) tuning approach that leverages the readily learned capacity of pixel level image diffusion model to capture spatial information for video generation. To accomplish this goal, we design an efficient architecture by inflating the weightings of the text-to-image SR model into our video generation framework. Additionally, we incorporate a temporal adapter to ensure temporal coherence across video frames. We investigate different tuning approaches based on our inflated architecture and report trade-offs between computational costs and super-resolution quality. Empirical evaluation, both quantitative and qualitative, on the Shutterstock video dataset, demonstrates that our approach is able to perform text-to-video SR generation with good visual quality and temporal consistency. To evaluate temporal coherence, we also present visualizations in video format in https://drive.google.com/drive/folders/1YVc-KMSJqOrEUdQWVaI-Yfu8Vsfu_1aO?usp=sharing . <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.10404v1-abstract-full').style.display = 'none'; document.getElementById('2401.10404v1-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 January, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">WACV'24 workshop</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2401.08544">arXiv:2401.08544</a> <span> [<a href="https://arxiv.org/pdf/2401.08544">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Numerical Analysis">math.NA</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"> N-Adaptive Ritz Method: A Neural Network Enriched Partition of Unity for Boundary Value Problems </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jonghyuk Baek</a>, <a href="/search/?searchtype=author&query=Wang%2C+Y">Yanran Wang</a>, <a href="/search/?searchtype=author&query=Chen%2C+J+S">J. S. Chen</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2401.08544v1-abstract-short" style="display: inline;"> Conventional finite element methods are known to be tedious in adaptive refinements due to their conformal regularity requirements. Further, the enrichment functions for adaptive refinements are often not readily available in general applications. This work introduces a novel neural network-enriched Partition of Unity (NN-PU) approach for solving boundary value problems via artificial neural netwo… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.08544v1-abstract-full').style.display = 'inline'; document.getElementById('2401.08544v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2401.08544v1-abstract-full" style="display: none;"> Conventional finite element methods are known to be tedious in adaptive refinements due to their conformal regularity requirements. Further, the enrichment functions for adaptive refinements are often not readily available in general applications. This work introduces a novel neural network-enriched Partition of Unity (NN-PU) approach for solving boundary value problems via artificial neural networks with a potential energy-based loss function minimization. The flexibility and adaptivity of the NN function space are utilized to capture complex solution patterns that the conventional Galerkin methods fail to capture. The NN enrichment is constructed by combining pre-trained feature-encoded NN blocks with an additional untrained NN block. The pre-trained NN blocks learn specific local features during the offline stage, enabling efficient enrichment of the approximation space during the online stage through the Ritz-type energy minimization. The NN enrichment is introduced under the Partition of Unity (PU) framework, ensuring convergence of the proposed method. The proposed NN-PU approximation and feature-encoded transfer learning forms an adaptive approximation framework, termed the neural-refinement (n-refinement), for solving boundary value problems. Demonstrated by solving various elasticity problems, the proposed method offers accurate solutions while notably reducing the computational cost compared to the conventional adaptive refinement in the mesh-based methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2401.08544v1-abstract-full').style.display = 'none'; document.getElementById('2401.08544v1-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, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">66 pages, 41 figures, 7 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/2312.14492">arXiv:2312.14492</a> <span> [<a href="https://arxiv.org/pdf/2312.14492">pdf</a>, <a href="https://arxiv.org/format/2312.14492">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"> Context Enhanced Transformer for Single Image Object Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=An%2C+S">Seungjun An</a>, <a href="/search/?searchtype=author&query=Park%2C+S">Seonghoon Park</a>, <a href="/search/?searchtype=author&query=Kim%2C+G">Gyeongnyeon Kim</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jeongyeol Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+B">Byeongwon Lee</a>, <a href="/search/?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="2312.14492v2-abstract-short" style="display: inline;"> With the increasing importance of video data in real-world applications, there is a rising need for efficient object detection methods that utilize temporal information. While existing video object detection (VOD) techniques employ various strategies to address this challenge, they typically depend on locally adjacent frames or randomly sampled images within a clip. Although recent Transformer-bas… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.14492v2-abstract-full').style.display = 'inline'; document.getElementById('2312.14492v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.14492v2-abstract-full" style="display: none;"> With the increasing importance of video data in real-world applications, there is a rising need for efficient object detection methods that utilize temporal information. While existing video object detection (VOD) techniques employ various strategies to address this challenge, they typically depend on locally adjacent frames or randomly sampled images within a clip. Although recent Transformer-based VOD methods have shown promising results, their reliance on multiple inputs and additional network complexity to incorporate temporal information limits their practical applicability. In this paper, we propose a novel approach to single image object detection, called Context Enhanced TRansformer (CETR), by incorporating temporal context into DETR using a newly designed memory module. To efficiently store temporal information, we construct a class-wise memory that collects contextual information across data. Additionally, we present a classification-based sampling technique to selectively utilize the relevant memory for the current image. In the testing, We introduce a test-time memory adaptation method that updates individual memory functions by considering the test distribution. Experiments with CityCam and ImageNet VID datasets exhibit the efficiency of the framework on various video systems. The project page and code will be made available at: https://ku-cvlab.github.io/CETR. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.14492v2-abstract-full').style.display = 'none'; document.getElementById('2312.14492v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 22 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page: https://ku-cvlab.github.io/CETR</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2312.10806">arXiv:2312.10806</a> <span> [<a href="https://arxiv.org/pdf/2312.10806">pdf</a>, <a href="https://arxiv.org/format/2312.10806">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"> Cross-Lingual Learning in Multilingual Scene Text Recognition </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jeonghun Baek</a>, <a href="/search/?searchtype=author&query=Matsui%2C+Y">Yusuke Matsui</a>, <a href="/search/?searchtype=author&query=Aizawa%2C+K">Kiyoharu Aizawa</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.10806v1-abstract-short" style="display: inline;"> In this paper, we investigate cross-lingual learning (CLL) for multilingual scene text recognition (STR). CLL transfers knowledge from one language to another. We aim to find the condition that exploits knowledge from high-resource languages for improving performance in low-resource languages. To do so, we first examine if two general insights about CLL discussed in previous works are applied to m… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.10806v1-abstract-full').style.display = 'inline'; document.getElementById('2312.10806v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.10806v1-abstract-full" style="display: none;"> In this paper, we investigate cross-lingual learning (CLL) for multilingual scene text recognition (STR). CLL transfers knowledge from one language to another. We aim to find the condition that exploits knowledge from high-resource languages for improving performance in low-resource languages. To do so, we first examine if two general insights about CLL discussed in previous works are applied to multilingual STR: (1) Joint learning with high- and low-resource languages may reduce performance on low-resource languages, and (2) CLL works best between typologically similar languages. Through extensive experiments, we show that two general insights may not be applied to multilingual STR. After that, we show that the crucial condition for CLL is the dataset size of high-resource languages regardless of the kind of high-resource languages. Our code, data, and models are available at https://github.com/ku21fan/CLL-STR. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.10806v1-abstract-full').style.display = 'none'; document.getElementById('2312.10806v1-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, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted at ICASSP2024, 5 pages, 2 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/2312.09567">arXiv:2312.09567</a> <span> [<a href="https://arxiv.org/pdf/2312.09567">pdf</a>, <a href="https://arxiv.org/format/2312.09567">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Astrophysics of Galaxies">astro-ph.GA</span> </div> </div> <p class="title is-5 mathjax"> Discovery of a large-scale H I plume in the NGC 7194 Group </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Pak%2C+M">Mina Pak</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Junhyun Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+J+H">Joon Hyeop Lee</a>, <a href="/search/?searchtype=author&query=Chung%2C+A">Aeree Chung</a>, <a href="/search/?searchtype=author&query=Owers%2C+M">Matt Owers</a>, <a href="/search/?searchtype=author&query=Jeong%2C+H">Hyunjin Jeong</a>, <a href="/search/?searchtype=author&query=Sung%2C+E">Eon-Chang Sung</a>, <a href="/search/?searchtype=author&query=Sheen%2C+Y">Yun-Kyeong Sheen</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2312.09567v1-abstract-short" style="display: inline;"> We present the discovery of a new H I structure in the NGC 7194 group from the observations using the Karl G. Jansky Very Large Array. NGC 7194 group is a nearby (z ~ 0.027) small galaxy group with five quiescent members. The observations reveal a 200 kpc-long H I plume that spans the entire group with a total mass of M$_{HI}$ = 3.4 x 10$^{10}$ M$_{\odot}$. The line-of-sight velocity of the H I ga… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.09567v1-abstract-full').style.display = 'inline'; document.getElementById('2312.09567v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2312.09567v1-abstract-full" style="display: none;"> We present the discovery of a new H I structure in the NGC 7194 group from the observations using the Karl G. Jansky Very Large Array. NGC 7194 group is a nearby (z ~ 0.027) small galaxy group with five quiescent members. The observations reveal a 200 kpc-long H I plume that spans the entire group with a total mass of M$_{HI}$ = 3.4 x 10$^{10}$ M$_{\odot}$. The line-of-sight velocity of the H I gas gradually increases from south (7200 km s$^{-1}$) to north (8200 km $^{-1}$), and the local velocity dispersion is up to 70 km s$^{-1}$. The structure is not spatially coincident with any member galaxies but it shows close associations with a number of blue star-forming knots. Intragroup H I gas is not rare, but this particular structure is still one of the unusual cases in the sense that it does not show any clear connection with sizable galaxies in the group. We discuss the potential origins of this large-scale H I gas in the NGC 7194 group and its relation with the intergalactic star-forming knots. We propose that this HI feature could have originated from tidal interactions among group members or the infall of a late-type galaxy into the group. Alternatively, it might be leftover gas from flyby intruders. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2312.09567v1-abstract-full').style.display = 'none'; document.getElementById('2312.09567v1-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 December, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">9 pages, 3 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/2311.16524">arXiv:2311.16524</a> <span> [<a href="https://arxiv.org/pdf/2311.16524">pdf</a>, <a href="https://arxiv.org/format/2311.16524">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="Machine Learning">cs.LG</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/10.1007/978-3-031-43999-5_36">10.1007/978-3-031-43999-5_36 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> 3D Teeth Reconstruction from Panoramic Radiographs using Neural Implicit Functions </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Park%2C+S">Sihwa Park</a>, <a href="/search/?searchtype=author&query=Kim%2C+S">Seongjun Kim</a>, <a href="/search/?searchtype=author&query=Song%2C+I">In-Seok Song</a>, <a href="/search/?searchtype=author&query=Baek%2C+S+J">Seung Jun Baek</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2311.16524v1-abstract-short" style="display: inline;"> Panoramic radiography is a widely used imaging modality in dental practice and research. However, it only provides flattened 2D images, which limits the detailed assessment of dental structures. In this paper, we propose Occudent, a framework for 3D teeth reconstruction from panoramic radiographs using neural implicit functions, which, to the best of our knowledge, is the first work to do so. For… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.16524v1-abstract-full').style.display = 'inline'; document.getElementById('2311.16524v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2311.16524v1-abstract-full" style="display: none;"> Panoramic radiography is a widely used imaging modality in dental practice and research. However, it only provides flattened 2D images, which limits the detailed assessment of dental structures. In this paper, we propose Occudent, a framework for 3D teeth reconstruction from panoramic radiographs using neural implicit functions, which, to the best of our knowledge, is the first work to do so. For a given point in 3D space, the implicit function estimates whether the point is occupied by a tooth, and thus implicitly determines the boundaries of 3D tooth shapes. Firstly, Occudent applies multi-label segmentation to the input panoramic radiograph. Next, tooth shape embeddings as well as tooth class embeddings are generated from the segmentation outputs, which are fed to the reconstruction network. A novel module called Conditional eXcitation (CX) is proposed in order to effectively incorporate the combined shape and class embeddings into the implicit function. The performance of Occudent is evaluated using both quantitative and qualitative measures. Importantly, Occudent is trained and validated with actual panoramic radiographs as input, distinct from recent works which used synthesized images. Experiments demonstrate the superiority of Occudent over state-of-the-art methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.16524v1-abstract-full').style.display = 'none'; document.getElementById('2311.16524v1-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 November, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">12 pages, 2 figures, accepted to International Conference on Medical Image Computing and Computer-Assisted Intervention MICCAI 2023</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2311.08590">arXiv:2311.08590</a> <span> [<a href="https://arxiv.org/pdf/2311.08590">pdf</a>, <a href="https://arxiv.org/format/2311.08590">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"> PEMA: An Offsite-Tunable Plug-in External Memory Adaptation for Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kim%2C+H">HyunJin Kim</a>, <a href="/search/?searchtype=author&query=Kim%2C+Y+J">Young Jin Kim</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2311.08590v3-abstract-short" style="display: inline;"> Pre-trained language models (PLMs) show impressive performance in various downstream NLP tasks. However, pre-training large language models demands substantial memory and training compute. Furthermore, due to the substantial resources required, many PLM weights are confidential. Consequently, users are compelled to share their data with model owners for fine-tuning specific tasks. To overcome the… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.08590v3-abstract-full').style.display = 'inline'; document.getElementById('2311.08590v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2311.08590v3-abstract-full" style="display: none;"> Pre-trained language models (PLMs) show impressive performance in various downstream NLP tasks. However, pre-training large language models demands substantial memory and training compute. Furthermore, due to the substantial resources required, many PLM weights are confidential. Consequently, users are compelled to share their data with model owners for fine-tuning specific tasks. To overcome the limitations, we introduce Plug-in External Memory Adaptation (PEMA), a Parameter-Efficient Fine-Tuning (PEFT) method, enabling PLM fine-tuning without requiring access to all the weights. PEMA integrates with context representations from test data during inference to perform downstream tasks. It uses external memory to store PLM-generated context representations mapped with target tokens. Our method utilizes weight matrices of LoRA-like bottlenecked adapter in the PLM's final layer to enhance efficiency. Our approach also includes Gradual Unrolling, a novel interpolation strategy to improve generation quality. We validate PEMA's effectiveness through experiments on syntactic and real datasets for machine translation and style transfer. Our findings show that PEMA outperforms other PEFT approaches in memory and latency efficiency for training, and also excels in maintaining sentence meaning and generating appropriate language and styles. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.08590v3-abstract-full').style.display = 'none'; document.getElementById('2311.08590v3-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 29 March, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 14 November, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to NAACL 2024</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2311.06318">arXiv:2311.06318</a> <span> [<a href="https://arxiv.org/pdf/2311.06318">pdf</a>, <a href="https://arxiv.org/format/2311.06318">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> <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="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Knowledge-Augmented Large Language Models for Personalized Contextual Query Suggestion </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Chandrasekaran%2C+N">Nirupama Chandrasekaran</a>, <a href="/search/?searchtype=author&query=Cucerzan%2C+S">Silviu Cucerzan</a>, <a href="/search/?searchtype=author&query=herring%2C+A">Allen herring</a>, <a href="/search/?searchtype=author&query=Jauhar%2C+S+K">Sujay Kumar Jauhar</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2311.06318v2-abstract-short" style="display: inline;"> Large Language Models (LLMs) excel at tackling various natural language tasks. However, due to the significant costs involved in re-training or fine-tuning them, they remain largely static and difficult to personalize. Nevertheless, a variety of applications could benefit from generations that are tailored to users' preferences, goals, and knowledge. Among them is web search, where knowing what a… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.06318v2-abstract-full').style.display = 'inline'; document.getElementById('2311.06318v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2311.06318v2-abstract-full" style="display: none;"> Large Language Models (LLMs) excel at tackling various natural language tasks. However, due to the significant costs involved in re-training or fine-tuning them, they remain largely static and difficult to personalize. Nevertheless, a variety of applications could benefit from generations that are tailored to users' preferences, goals, and knowledge. Among them is web search, where knowing what a user is trying to accomplish, what they care about, and what they know can lead to improved search experiences. In this work, we propose a novel and general approach that augments an LLM with relevant context from users' interaction histories with a search engine in order to personalize its outputs. Specifically, we construct an entity-centric knowledge store for each user based on their search and browsing activities on the web, which is then leveraged to provide contextually relevant LLM prompt augmentations. This knowledge store is light-weight, since it only produces user-specific aggregate projections of interests and knowledge onto public knowledge graphs, and leverages existing search log infrastructure, thereby mitigating the privacy, compliance, and scalability concerns associated with building deep user profiles for personalization. We validate our approach on the task of contextual query suggestion, which requires understanding not only the user's current search context but also what they historically know and care about. Through a number of experiments based on human evaluation, we show that our approach is significantly better than several other LLM-powered baselines, generating query suggestions that are contextually more relevant, personalized, and useful. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2311.06318v2-abstract-full').style.display = 'none'; document.getElementById('2311.06318v2-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> 19 February, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 9 November, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">The Web Conference (WWW) 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/2310.17857">arXiv:2310.17857</a> <span> [<a href="https://arxiv.org/pdf/2310.17857">pdf</a>, <a href="https://arxiv.org/format/2310.17857">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"> From Values to Opinions: Predicting Human Behaviors and Stances Using Value-Injected Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kang%2C+D">Dongjun Kang</a>, <a href="/search/?searchtype=author&query=Park%2C+J">Joonsuk Park</a>, <a href="/search/?searchtype=author&query=Jo%2C+Y">Yohan Jo</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.17857v1-abstract-short" style="display: inline;"> Being able to predict people's opinions on issues and behaviors in realistic scenarios can be helpful in various domains, such as politics and marketing. However, conducting large-scale surveys like the European Social Survey to solicit people's opinions on individual issues can incur prohibitive costs. Leveraging prior research showing influence of core human values on individual decisions and ac… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.17857v1-abstract-full').style.display = 'inline'; document.getElementById('2310.17857v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.17857v1-abstract-full" style="display: none;"> Being able to predict people's opinions on issues and behaviors in realistic scenarios can be helpful in various domains, such as politics and marketing. However, conducting large-scale surveys like the European Social Survey to solicit people's opinions on individual issues can incur prohibitive costs. Leveraging prior research showing influence of core human values on individual decisions and actions, we propose to use value-injected large language models (LLM) to predict opinions and behaviors. To this end, we present Value Injection Method (VIM), a collection of two methods -- argument generation and question answering -- designed to inject targeted value distributions into LLMs via fine-tuning. We then conduct a series of experiments on four tasks to test the effectiveness of VIM and the possibility of using value-injected LLMs to predict opinions and behaviors of people. We find that LLMs value-injected with variations of VIM substantially outperform the baselines. Also, the results suggest that opinions and behaviors can be better predicted using value-injected LLMs than the baseline approaches. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.17857v1-abstract-full').style.display = 'none'; document.getElementById('2310.17857v1-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 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">EMNLP 2023 main paper accepted</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.16446">arXiv:2310.16446</a> <span> [<a href="https://arxiv.org/pdf/2310.16446">pdf</a>, <a href="https://arxiv.org/format/2310.16446">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"> Diversity Enhanced Narrative Question Generation for Storybooks </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Yoon%2C+H">Hokeun Yoon</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.16446v1-abstract-short" style="display: inline;"> Question generation (QG) from a given context can enhance comprehension, engagement, assessment, and overall efficacy in learning or conversational environments. Despite recent advancements in QG, the challenge of enhancing or measuring the diversity of generated questions often remains unaddressed. In this paper, we introduce a multi-question generation model (mQG), which is capable of generating… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.16446v1-abstract-full').style.display = 'inline'; document.getElementById('2310.16446v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.16446v1-abstract-full" style="display: none;"> Question generation (QG) from a given context can enhance comprehension, engagement, assessment, and overall efficacy in learning or conversational environments. Despite recent advancements in QG, the challenge of enhancing or measuring the diversity of generated questions often remains unaddressed. In this paper, we introduce a multi-question generation model (mQG), which is capable of generating multiple, diverse, and answerable questions by focusing on context and questions. To validate the answerability of the generated questions, we employ a SQuAD2.0 fine-tuned question answering model, classifying the questions as answerable or not. We train and evaluate mQG on the FairytaleQA dataset, a well-structured QA dataset based on storybooks, with narrative questions. We further apply a zero-shot adaptation on the TellMeWhy and SQuAD1.1 datasets. mQG shows promising results across various evaluation metrics, among strong baselines. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.16446v1-abstract-full').style.display = 'none'; document.getElementById('2310.16446v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 25 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted to EMNLP 2023</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.13307">arXiv:2310.13307</a> <span> [<a href="https://arxiv.org/pdf/2310.13307">pdf</a>, <a href="https://arxiv.org/format/2310.13307">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="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Test-Time Self-Adaptive Small Language Models for Question Answering </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Cho%2C+S">Sukmin Cho</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a>, <a href="/search/?searchtype=author&query=Park%2C+J+C">Jong C. Park</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.13307v1-abstract-short" style="display: inline;"> Recent instruction-finetuned large language models (LMs) have achieved notable performances in various tasks, such as question-answering (QA). However, despite their ability to memorize a vast amount of general knowledge across diverse tasks, they might be suboptimal on specific tasks due to their limited capacity to transfer and adapt knowledge to target tasks. Moreover, further finetuning LMs wi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.13307v1-abstract-full').style.display = 'inline'; document.getElementById('2310.13307v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.13307v1-abstract-full" style="display: none;"> Recent instruction-finetuned large language models (LMs) have achieved notable performances in various tasks, such as question-answering (QA). However, despite their ability to memorize a vast amount of general knowledge across diverse tasks, they might be suboptimal on specific tasks due to their limited capacity to transfer and adapt knowledge to target tasks. Moreover, further finetuning LMs with labeled datasets is often infeasible due to their absence, but it is also questionable if we can transfer smaller LMs having limited knowledge only with unlabeled test data. In this work, we show and investigate the capabilities of smaller self-adaptive LMs, only with unlabeled test data. In particular, we first stochastically generate multiple answers, and then ensemble them while filtering out low-quality samples to mitigate noise from inaccurate labels. Our proposed self-adaption strategy demonstrates significant performance improvements on benchmark QA datasets with higher robustness across diverse prompts, enabling LMs to stay stable. Code is available at: https://github.com/starsuzi/T-SAS. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.13307v1-abstract-full').style.display = 'none'; document.getElementById('2310.13307v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">EMNLP Findings 2023</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.12836">arXiv:2310.12836</a> <span> [<a href="https://arxiv.org/pdf/2310.12836">pdf</a>, <a href="https://arxiv.org/format/2310.12836">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="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Knowledge-Augmented Language Model Verification </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Kang%2C+M">Minki Kang</a>, <a href="/search/?searchtype=author&query=Park%2C+J+C">Jong C. Park</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.12836v1-abstract-short" style="display: inline;"> Recent Language Models (LMs) have shown impressive capabilities in generating texts with the knowledge internalized in parameters. Yet, LMs often generate the factually incorrect responses to the given queries, since their knowledge may be inaccurate, incomplete, and outdated. To address this problem, previous works propose to augment LMs with the knowledge retrieved from an external knowledge sou… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.12836v1-abstract-full').style.display = 'inline'; document.getElementById('2310.12836v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.12836v1-abstract-full" style="display: none;"> Recent Language Models (LMs) have shown impressive capabilities in generating texts with the knowledge internalized in parameters. Yet, LMs often generate the factually incorrect responses to the given queries, since their knowledge may be inaccurate, incomplete, and outdated. To address this problem, previous works propose to augment LMs with the knowledge retrieved from an external knowledge source. However, such approaches often show suboptimal text generation performance due to two reasons: 1) the model may fail to retrieve the knowledge relevant to the given query, or 2) the model may not faithfully reflect the retrieved knowledge in the generated text. To overcome these, we propose to verify the output and the knowledge of the knowledge-augmented LMs with a separate verifier, which is a small LM that is trained to detect those two types of errors through instruction-finetuning. Then, when the verifier recognizes an error, we can rectify it by either retrieving new knowledge or generating new text. Further, we use an ensemble of the outputs from different instructions with a single verifier to enhance the reliability of the verification processes. We validate the effectiveness of the proposed verification steps on multiple question answering benchmarks, whose results show that the proposed verifier effectively identifies retrieval and generation errors, allowing LMs to provide more factually correct outputs. Our code is available at https://github.com/JinheonBaek/KALMV. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.12836v1-abstract-full').style.display = 'none'; document.getElementById('2310.12836v1-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> 19 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">EMNLP 2023</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.12412">arXiv:2310.12412</a> <span> [<a href="https://arxiv.org/pdf/2310.12412">pdf</a>, <a href="https://arxiv.org/format/2310.12412">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Plasma Physics">physics.plasm-ph</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Applied Physics">physics.app-ph</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/10.1088/1361-6595/ad2b7c">10.1088/1361-6595/ad2b7c <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Experimental characterization of thermionic surface cooling in thermionic discharge </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Bak%2C+J">Junhwi Bak</a>, <a href="/search/?searchtype=author&query=Tropina%2C+A">Albina Tropina</a>, <a href="/search/?searchtype=author&query=Creel%2C+J">James Creel</a>, <a href="/search/?searchtype=author&query=Miles%2C+R+B">Richard B. Miles</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.12412v1-abstract-short" style="display: inline;"> In this work, the thermionic cooling effect during thermionic discharges with parallel plate electrodes at 1 Torr is investigated. Time-resolved observation of electron emission and surface temperature is realized in addition to the typical steady state characterization. Surface cooling by the electron emission, initiated by plasma ignition, is directly captured at its onset and an estimated cooli… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.12412v1-abstract-full').style.display = 'inline'; document.getElementById('2310.12412v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.12412v1-abstract-full" style="display: none;"> In this work, the thermionic cooling effect during thermionic discharges with parallel plate electrodes at 1 Torr is investigated. Time-resolved observation of electron emission and surface temperature is realized in addition to the typical steady state characterization. Surface cooling by the electron emission, initiated by plasma ignition, is directly captured at its onset and an estimated cooling capacity of 1.6 \pm 0.2 MW/m^2 is observed. The present work provides experimental evidence of considerable surface cooling achieved by thermionic cooling. This result indicates that thermionic cooling can be a promising thermal protection method at elevated temperatures, such as those encountered by hypersonic vehicle leading edges in flight. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.12412v1-abstract-full').style.display = 'none'; document.getElementById('2310.12412v1-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 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">14 pages, 21 figures. Submitted on 17 September 2022</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> Plasma Sources Sci. Technol. 33, 034001 (2024) </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.09687">arXiv:2310.09687</a> <span> [<a href="https://arxiv.org/pdf/2310.09687">pdf</a>, <a href="https://arxiv.org/format/2310.09687">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="Computers and Society">cs.CY</span> </div> </div> <p class="title is-5 mathjax"> When Collaborative Filtering is not Collaborative: Unfairness of PCA for Recommendations </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Liu%2C+D">David Liu</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jackie Baek</a>, <a href="/search/?searchtype=author&query=Eliassi-Rad%2C+T">Tina Eliassi-Rad</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.09687v1-abstract-short" style="display: inline;"> We study the fairness of dimensionality reduction methods for recommendations. We focus on the established method of principal component analysis (PCA), which identifies latent components and produces a low-rank approximation via the leading components while discarding the trailing components. Prior works have defined notions of "fair PCA"; however, these definitions do not answer the following qu… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.09687v1-abstract-full').style.display = 'inline'; document.getElementById('2310.09687v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.09687v1-abstract-full" style="display: none;"> We study the fairness of dimensionality reduction methods for recommendations. We focus on the established method of principal component analysis (PCA), which identifies latent components and produces a low-rank approximation via the leading components while discarding the trailing components. Prior works have defined notions of "fair PCA"; however, these definitions do not answer the following question: what makes PCA unfair? We identify two underlying mechanisms of PCA that induce unfairness at the item level. The first negatively impacts less popular items, due to the fact that less popular items rely on trailing latent components to recover their values. The second negatively impacts the highly popular items, since the leading PCA components specialize in individual popular items instead of capturing similarities between items. To address these issues, we develop a polynomial-time algorithm, Item-Weighted PCA, a modification of PCA that uses item-specific weights in the objective. On a stylized class of matrices, we prove that Item-Weighted PCA using a specific set of weights minimizes a popularity-normalized error metric. Our evaluations on real-world datasets show that Item-Weighted PCA not only improves overall recommendation quality by up to $0.1$ item-level AUC-ROC but also improves on both popular and less popular items. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.09687v1-abstract-full').style.display = 'none'; document.getElementById('2310.09687v1-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 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.06928">arXiv:2310.06928</a> <span> [<a href="https://arxiv.org/pdf/2310.06928">pdf</a>, <a href="https://arxiv.org/format/2310.06928">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Astrophysics of Galaxies">astro-ph.GA</span> </div> </div> <p class="title is-5 mathjax"> Unraveling the Complex Structure of AGN-driven Outflows. VI. Strong Ionized Outflows in Type 1 AGNs and the Outflow Size-Luminosity Relation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Kim%2C+C">Changseok Kim</a>, <a href="/search/?searchtype=author&query=Woo%2C+J">Jong-Hak Woo</a>, <a href="/search/?searchtype=author&query=Luo%2C+R">Rongxin Luo</a>, <a href="/search/?searchtype=author&query=Chung%2C+A">Aeree Chung</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Junhyun Baek</a>, <a href="/search/?searchtype=author&query=Le%2C+H+A+N">Huynh Anh N. Le</a>, <a href="/search/?searchtype=author&query=Son%2C+D">Donghoon Son</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.06928v2-abstract-short" style="display: inline;"> We present spatially resolved gas kinematics, ionization, and energetics of 11 type 1 and 5 type 2 active galactic nuclei (AGNs) with strong ionized gas outflows at z $<0.3$ using Gemini Multi-Object Spectrograph Integral Field Unit (GMOS-IFU) data. We find a strongly blueshifted region in [OIII] velocity maps, representing an approaching cone in biconical outflows, and blueshifted and redshifted… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.06928v2-abstract-full').style.display = 'inline'; document.getElementById('2310.06928v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.06928v2-abstract-full" style="display: none;"> We present spatially resolved gas kinematics, ionization, and energetics of 11 type 1 and 5 type 2 active galactic nuclei (AGNs) with strong ionized gas outflows at z $<0.3$ using Gemini Multi-Object Spectrograph Integral Field Unit (GMOS-IFU) data. We find a strongly blueshifted region in [OIII] velocity maps, representing an approaching cone in biconical outflows, and blueshifted and redshifted regions in H$伪$ velocity maps, which show gravitationally rotating kinematics. AGN photoionization is dominant in the central region of most targets, and some of them also show ring-like structures of LINER or composite that surround the AGN-dominated center. Following our previous studies, we kinematically determine outflow sizes by the ratio between [OIII] and stellar velocity dispersion. Outflow sizes of type 1 AGNs follow the same kinematic outflow size-[OIII] luminosity relation obtained from the type 2 IFU sample in Kang & Woo and Luo (updated slope $0.29\pm0.04$), while they are limited to the central kpc scales, indicating the lack of global impact of outflows on the interstellar medium. Small mass outflow rates and large star formation rates of the combined sample support that there is no evidence of rapid star formation quenching by outflows, which is consistent with the delayed AGN feedback. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.06928v2-abstract-full').style.display = 'none'; document.getElementById('2310.06928v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 12 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 10 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">27 pages, 18 figures, Accepted for publication in ApJ</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.03052">arXiv:2310.03052</a> <span> [<a href="https://arxiv.org/pdf/2310.03052">pdf</a>, <a href="https://arxiv.org/format/2310.03052">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="Neural and Evolutionary Computing">cs.NE</span> </div> </div> <p class="title is-5 mathjax"> Memoria: Resolving Fateful Forgetting Problem through Human-Inspired Memory Architecture </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Park%2C+S">Sangjun Park</a>, <a href="/search/?searchtype=author&query=Bak%2C+J">JinYeong Bak</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.03052v3-abstract-short" style="display: inline;"> Making neural networks remember over the long term has been a longstanding issue. Although several external memory techniques have been introduced, most focus on retaining recent information in the short term. Regardless of its importance, information tends to be fatefully forgotten over time. We present Memoria, a memory system for artificial neural networks, drawing inspiration from humans and a… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.03052v3-abstract-full').style.display = 'inline'; document.getElementById('2310.03052v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.03052v3-abstract-full" style="display: none;"> Making neural networks remember over the long term has been a longstanding issue. Although several external memory techniques have been introduced, most focus on retaining recent information in the short term. Regardless of its importance, information tends to be fatefully forgotten over time. We present Memoria, a memory system for artificial neural networks, drawing inspiration from humans and applying various neuroscientific and psychological theories. The experimental results prove the effectiveness of Memoria in the diverse tasks of sorting, language modeling, and classification, surpassing conventional techniques. Engram analysis reveals that Memoria exhibits the primacy, recency, and temporal contiguity effects which are characteristics of human memory. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.03052v3-abstract-full').style.display = 'none'; document.getElementById('2310.03052v3-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> 8 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 4 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">ICML 2024 Spotlight. 29 pages, 15 figures, 11 tables</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> Proceedings of the 41st International Conference on Machine Learning, PMLR 235:39587-39615, 2024 </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2310.02580">arXiv:2310.02580</a> <span> [<a href="https://arxiv.org/pdf/2310.02580">pdf</a>, <a href="https://arxiv.org/ps/2310.02580">ps</a>, <a href="https://arxiv.org/format/2310.02580">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Quantum Physics">quant-ph</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Quantum Gases">cond-mat.quant-gas</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/10.1103/PhysRevLett.132.240803">10.1103/PhysRevLett.132.240803 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Self-consistent many-body metrology </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baak%2C+J">Jae-Gyun Baak</a>, <a href="/search/?searchtype=author&query=Fischer%2C+U+R">Uwe R. Fischer</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2310.02580v3-abstract-short" style="display: inline;"> We investigate performing classical and quantum metrology and parameter estimation by using interacting trapped bosons, which we theoretically treat by a self-consistent many-body approach of the multiconfigurational Hartree type. Focusing on a tilted double-well geometry, we compare a self-consistently determined and monitored two-mode truncation, with dynamically changing orbitals, to the conven… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.02580v3-abstract-full').style.display = 'inline'; document.getElementById('2310.02580v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2310.02580v3-abstract-full" style="display: none;"> We investigate performing classical and quantum metrology and parameter estimation by using interacting trapped bosons, which we theoretically treat by a self-consistent many-body approach of the multiconfigurational Hartree type. Focusing on a tilted double-well geometry, we compare a self-consistently determined and monitored two-mode truncation, with dynamically changing orbitals, to the conventional two-mode approach of fixed orbitals, where only Fock space coefficients evolve in time. We demonstrate that, as a consequence, various metrological quantities associated to a concrete measurement such as the classical Fisher information and the maximum likelihood estimator are deeply affected by the orbitals' change during the quantum evolution. Self-consistency of the quantum many-body dynamics of interacting trapped ultracold gases thus fundamentally affects the attainable parameter estimation accuracy of a given metrological protocol. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2310.02580v3-abstract-full').style.display = 'none'; document.getElementById('2310.02580v3-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 May, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 4 October, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> October 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">6+7 pages, 4+4 figures; version as accepted by PRL</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> Phys. Rev. Lett. 132, 240803 (2024) </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2309.13519">arXiv:2309.13519</a> <span> [<a href="https://arxiv.org/pdf/2309.13519">pdf</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> <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"> Data-Driven Modeling of an Unsaturated Bentonite Buffer Model Test Under High Temperatures Using an Enhanced Axisymmetric Reproducing Kernel Particle Method </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jonghyuk Baek</a>, <a href="/search/?searchtype=author&query=Wang%2C+Y">Yanran Wang</a>, <a href="/search/?searchtype=author&query=He%2C+X">Xiaolong He</a>, <a href="/search/?searchtype=author&query=Lu%2C+Y">Yu Lu</a>, <a href="/search/?searchtype=author&query=McCartney%2C+J+S">John S. McCartney</a>, <a href="/search/?searchtype=author&query=Chen%2C+J+S">J. S. Chen</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="2309.13519v1-abstract-short" style="display: inline;"> In deep geological repositories for high level nuclear waste with close canister spacings, bentonite buffers can experience temperatures higher than 100 掳C. In this range of extreme temperatures, phenomenological constitutive laws face limitations in capturing the thermo-hydro-mechanical (THM) behavior of the bentonite, since the pre-defined functional constitutive laws often lack generality and f… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2309.13519v1-abstract-full').style.display = 'inline'; document.getElementById('2309.13519v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2309.13519v1-abstract-full" style="display: none;"> In deep geological repositories for high level nuclear waste with close canister spacings, bentonite buffers can experience temperatures higher than 100 掳C. In this range of extreme temperatures, phenomenological constitutive laws face limitations in capturing the thermo-hydro-mechanical (THM) behavior of the bentonite, since the pre-defined functional constitutive laws often lack generality and flexibility to capture a wide range of complex coupling phenomena as well as the effects of stress state and path dependency. In this work, a deep neural network (DNN)-based soil-water retention curve (SWRC) of bentonite is introduced and integrated into a Reproducing Kernel Particle Method (RKPM) for conducting THM simulations of the bentonite buffer. The DNN-SWRC model incorporates temperature as an additional input variable, allowing it to learn the relationship between suction and degree of saturation under the general non-isothermal condition, which is difficult to represent using a phenomenological SWRC. For effective modeling of the tank-scale test, new axisymmetric Reproducing Kernel basis functions enriched with singular Dirichlet enforcement representing heater placement and an effective convective heat transfer coefficient representing thin-layer composite tank construction are developed. The proposed method is demonstrated through the modeling of a tank-scale experiment involving a cylindrical layer of MX-80 bentonite exposed to central heating. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2309.13519v1-abstract-full').style.display = 'none'; document.getElementById('2309.13519v1-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 September, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> September 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">51 pages, 19 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/2308.12565">arXiv:2308.12565</a> <span> [<a href="https://arxiv.org/pdf/2308.12565">pdf</a>, <a href="https://arxiv.org/format/2308.12565">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Astrophysics of Galaxies">astro-ph.GA</span> </div> </div> <p class="title is-5 mathjax"> AMUSE-antlia I: Nuclear X-ray properties of early-type galaxies in a dynamically young galaxy cluster </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Hu%2C+Z">Zhensong Hu</a>, <a href="/search/?searchtype=author&query=Su%2C+Y">Yuanyuan Su</a>, <a href="/search/?searchtype=author&query=Li%2C+Z">Zhiyuan Li</a>, <a href="/search/?searchtype=author&query=Hess%2C+K+M">Kelley M. Hess</a>, <a href="/search/?searchtype=author&query=Kraft%2C+R+P">Ralph P. Kraft</a>, <a href="/search/?searchtype=author&query=Forman%2C+W+R">William R. Forman</a>, <a href="/search/?searchtype=author&query=Nulsen%2C+P+E+J">Paul E. J. Nulsen</a>, <a href="/search/?searchtype=author&query=Sridhar%2C+S+S">Sarrvesh S. Sridhar</a>, <a href="/search/?searchtype=author&query=Stroe%2C+A">Andra Stroe</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Junhyun Baek</a>, <a href="/search/?searchtype=author&query=Chung%2C+A">Aeree Chung</a>, <a href="/search/?searchtype=author&query=Grupe%2C+D">Dirk Grupe</a>, <a href="/search/?searchtype=author&query=Chen%2C+H">Hao Chen</a>, <a href="/search/?searchtype=author&query=Irwin%2C+J+A">Jimmy A. Irwin</a>, <a href="/search/?searchtype=author&query=Jones%2C+C">Christine Jones</a>, <a href="/search/?searchtype=author&query=Randall%2C+S+W">Scott W. Randall</a>, <a href="/search/?searchtype=author&query=Roediger%2C+E">Elke Roediger</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="2308.12565v1-abstract-short" style="display: inline;"> To understand the formation and growth of supermassive black holes (SMBHs) and their co-evolution with host galaxies, it is essential to know the impact of environment on the activity of active galactic nuclei (AGN). We present new Chandra X-ray observations of nuclear emission from member galaxies in the Antlia cluster, the nearest non-cool core and the nearest merging galaxy cluster, residing at… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2308.12565v1-abstract-full').style.display = 'inline'; document.getElementById('2308.12565v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2308.12565v1-abstract-full" style="display: none;"> To understand the formation and growth of supermassive black holes (SMBHs) and their co-evolution with host galaxies, it is essential to know the impact of environment on the activity of active galactic nuclei (AGN). We present new Chandra X-ray observations of nuclear emission from member galaxies in the Antlia cluster, the nearest non-cool core and the nearest merging galaxy cluster, residing at D = 35.2 Mpc. Its inner region, centered on two dominant galaxies NGC 3268 and NGC 3258, has been mapped with three deep Chandra ACIS-I pointings. Nuclear X-ray sources are detected in 7/84 (8.3%) early-type galaxies (ETG) and 2/8 (25%) late-type galaxies with a median detection limit of 8x10^38 erg/s. All nuclear X-ray sources but one have a corresponding radio continuum source detected by MeerKAT at the L-band. Nuclear X-ray sources detected in early-type galaxies are considered as the genuine X-ray counterpart of low-luminosity AGN. When restricted to a detection limit of logLx(erg/s) > 38.9 and a stellar mass of 10 < log Ms(Msun) <11.6, 6/11 (54.5%) ETG are found to contain an X-ray AGN in Antlia, exceeding the AGN occupation fraction of 7/39 (18.0%) and 2/12 (16.7%) in the more relaxed, cool core clusters, Virgo and Fornax, respectively, and rivaling that of the AMUSE-Field ETG of 27/49 (55.1%). Furthermore, more than half of the X-ray AGN in Antlia are hosted by its younger subcluster, centered on NGC 3258. We believe that this is because SMBH activity is enhanced in a dynamically young cluster compared to relatively relaxed clusters. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2308.12565v1-abstract-full').style.display = 'none'; document.getElementById('2308.12565v1-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 August, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">14 pages, 8 figures, 2 tables, accepted for publication in The Astrophysical Journal</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2308.07741">arXiv:2308.07741</a> <span> [<a href="https://arxiv.org/pdf/2308.07741">pdf</a>, <a href="https://arxiv.org/format/2308.07741">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="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Real Robot Challenge 2022: Learning Dexterous Manipulation from Offline Data in the Real World </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=G%C3%BCrtler%2C+N">Nico G眉rtler</a>, <a href="/search/?searchtype=author&query=Widmaier%2C+F">Felix Widmaier</a>, <a href="/search/?searchtype=author&query=Sancaktar%2C+C">Cansu Sancaktar</a>, <a href="/search/?searchtype=author&query=Blaes%2C+S">Sebastian Blaes</a>, <a href="/search/?searchtype=author&query=Kolev%2C+P">Pavel Kolev</a>, <a href="/search/?searchtype=author&query=Bauer%2C+S">Stefan Bauer</a>, <a href="/search/?searchtype=author&query=W%C3%BCthrich%2C+M">Manuel W眉thrich</a>, <a href="/search/?searchtype=author&query=Wulfmeier%2C+M">Markus Wulfmeier</a>, <a href="/search/?searchtype=author&query=Riedmiller%2C+M">Martin Riedmiller</a>, <a href="/search/?searchtype=author&query=Allshire%2C+A">Arthur Allshire</a>, <a href="/search/?searchtype=author&query=Wang%2C+Q">Qiang Wang</a>, <a href="/search/?searchtype=author&query=McCarthy%2C+R">Robert McCarthy</a>, <a href="/search/?searchtype=author&query=Kim%2C+H">Hangyeol Kim</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jongchan Baek</a>, <a href="/search/?searchtype=author&query=Kwon%2C+W">Wookyong Kwon</a>, <a href="/search/?searchtype=author&query=Qian%2C+S">Shanliang Qian</a>, <a href="/search/?searchtype=author&query=Toshimitsu%2C+Y">Yasunori Toshimitsu</a>, <a href="/search/?searchtype=author&query=Michelis%2C+M+Y">Mike Yan Michelis</a>, <a href="/search/?searchtype=author&query=Kazemipour%2C+A">Amirhossein Kazemipour</a>, <a href="/search/?searchtype=author&query=Raayatsanati%2C+A">Arman Raayatsanati</a>, <a href="/search/?searchtype=author&query=Zheng%2C+H">Hehui Zheng</a>, <a href="/search/?searchtype=author&query=Cangan%2C+B+G">Barnabas Gavin Cangan</a>, <a href="/search/?searchtype=author&query=Sch%C3%B6lkopf%2C+B">Bernhard Sch枚lkopf</a>, <a href="/search/?searchtype=author&query=Martius%2C+G">Georg Martius</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="2308.07741v3-abstract-short" style="display: inline;"> Experimentation on real robots is demanding in terms of time and costs. For this reason, a large part of the reinforcement learning (RL) community uses simulators to develop and benchmark algorithms. However, insights gained in simulation do not necessarily translate to real robots, in particular for tasks involving complex interactions with the environment. The Real Robot Challenge 2022 therefore… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2308.07741v3-abstract-full').style.display = 'inline'; document.getElementById('2308.07741v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2308.07741v3-abstract-full" style="display: none;"> Experimentation on real robots is demanding in terms of time and costs. For this reason, a large part of the reinforcement learning (RL) community uses simulators to develop and benchmark algorithms. However, insights gained in simulation do not necessarily translate to real robots, in particular for tasks involving complex interactions with the environment. The Real Robot Challenge 2022 therefore served as a bridge between the RL and robotics communities by allowing participants to experiment remotely with a real robot - as easily as in simulation. In the last years, offline reinforcement learning has matured into a promising paradigm for learning from pre-collected datasets, alleviating the reliance on expensive online interactions. We therefore asked the participants to learn two dexterous manipulation tasks involving pushing, grasping, and in-hand orientation from provided real-robot datasets. An extensive software documentation and an initial stage based on a simulation of the real set-up made the competition particularly accessible. By giving each team plenty of access budget to evaluate their offline-learned policies on a cluster of seven identical real TriFinger platforms, we organized an exciting competition for machine learners and roboticists alike. In this work we state the rules of the competition, present the methods used by the winning teams and compare their results with a benchmark of state-of-the-art offline RL algorithms on the challenge datasets. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2308.07741v3-abstract-full').style.display = 'none'; document.getElementById('2308.07741v3-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 November, 2023; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 15 August, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> August 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Typo in author list fixed</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2307.01937">arXiv:2307.01937</a> <span> [<a href="https://arxiv.org/pdf/2307.01937">pdf</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> <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="Numerical Analysis">math.NA</span> </div> </div> <p class="title is-5 mathjax"> A Neural Network-Based Enrichment of Reproducing Kernel Approximation for Modeling Brittle Fracture </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jonghyuk Baek</a>, <a href="/search/?searchtype=author&query=Chen%2C+J">Jiun-Shyan Chen</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="2307.01937v1-abstract-short" style="display: inline;"> Numerical modeling of localizations is a challenging task due to the evolving rough solution in which the localization paths are not predefined. Despite decades of efforts, there is a need for innovative discretization-independent computational methods to predict the evolution of localizations. In this work, an improved version of the neural network-enhanced Reproducing Kernel Particle Method (NN-… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2307.01937v1-abstract-full').style.display = 'inline'; document.getElementById('2307.01937v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2307.01937v1-abstract-full" style="display: none;"> Numerical modeling of localizations is a challenging task due to the evolving rough solution in which the localization paths are not predefined. Despite decades of efforts, there is a need for innovative discretization-independent computational methods to predict the evolution of localizations. In this work, an improved version of the neural network-enhanced Reproducing Kernel Particle Method (NN-RKPM) is proposed for modeling brittle fracture. In the proposed method, a background reproducing kernel (RK) approximation defined on a coarse and uniform discretization is enriched by a neural network (NN) approximation under a Partition of Unity framework. In the NN approximation, the deep neural network automatically locates and inserts regularized discontinuities in the function space. The NN-based enrichment functions are then patched together with RK approximation functions using RK as a Partition of Unity patching function. The optimum NN parameters defining the location, orientation, and displacement distribution across location together with RK approximation coefficients are obtained via the energy-based loss function minimization. To regularize the NN-RK approximation, a constraint on the spatial gradient of the parametric coordinates is imposed in the loss function. Analysis of the convergence properties shows that the solution convergence of the proposed method is guaranteed. The effectiveness of the proposed method is demonstrated by a series of numerical examples involving damage propagation and branching. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2307.01937v1-abstract-full').style.display = 'none'; document.getElementById('2307.01937v1-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 July, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> July 2023. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2306.09533">arXiv:2306.09533</a> <span> [<a href="https://arxiv.org/pdf/2306.09533">pdf</a>, <a href="https://arxiv.org/ps/2306.09533">ps</a>, <a href="https://arxiv.org/format/2306.09533">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Combinatorics">math.CO</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Metric Geometry">math.MG</span> </div> </div> <p class="title is-5 mathjax"> $n^2 + 1$ unit equilateral triangles cannot cover an equilateral triangle of side $> n$ if all triangles have parallel sides </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Baek%2C+J">Jineon Baek</a>, <a href="/search/?searchtype=author&query=Lee%2C+S">Seewoo Lee</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2306.09533v3-abstract-short" style="display: inline;"> Conway and Soifer showed that an equilateral triangle $T$ of side $n + \varepsilon$ with sufficiently small $\varepsilon > 0$ can be covered by $n^2 + 2$ unit equilateral triangles. They conjectured that it is impossible to cover $T$ with $n^2 + 1$ unit equilateral triangles no matter how small $\varepsilon$ is. We show that if we require all sides of the unit equilateral triangles to be paralle… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2306.09533v3-abstract-full').style.display = 'inline'; document.getElementById('2306.09533v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2306.09533v3-abstract-full" style="display: none;"> Conway and Soifer showed that an equilateral triangle $T$ of side $n + \varepsilon$ with sufficiently small $\varepsilon > 0$ can be covered by $n^2 + 2$ unit equilateral triangles. They conjectured that it is impossible to cover $T$ with $n^2 + 1$ unit equilateral triangles no matter how small $\varepsilon$ is. We show that if we require all sides of the unit equilateral triangles to be parallel to the sides of $T$ (e.g. $\bigtriangleup$ and $\bigtriangledown$), then it is impossible to cover $T$ of side $n + \varepsilon$ with $n^2 + 1$ unit equilateral triangles for any $\varepsilon > 0$. As the coverings of $T$ by Conway and Soifer only involve triangles with sides parallel to $T$, our result determines the exact minimum number $n^2+2$ of unit equilateral triangles with all sides parallel to $T$ that cover $T$. We also determine the largest value $\varepsilon = 1/(n + 1)$ (resp. $\varepsilon = 1 / n$) of $\varepsilon$ such that the equilateral triangle $T$ of side $n + \varepsilon$ can be covered by $n^2+2$ (resp. $n^2 + 3$) unit equilateral triangles with sides parallel to $T$, where the first case is achieved by the construction of Conway and Soifer. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2306.09533v3-abstract-full').style.display = 'none'; document.getElementById('2306.09533v3-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 June, 2024; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 15 June, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">8 pages, 7 figures</span> </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">MSC Class:</span> 52C15 (Primary) 05B40 (Secondary) </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2306.04293">arXiv:2306.04293</a> <span> [<a href="https://arxiv.org/pdf/2306.04293">pdf</a>, <a href="https://arxiv.org/format/2306.04293">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="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"> Phrase Retrieval for Open-Domain Conversational Question Answering with Conversational Dependency Modeling via Contrastive Learning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/?searchtype=author&query=Jeong%2C+S">Soyeong Jeong</a>, <a href="/search/?searchtype=author&query=Baek%2C+J">Jinheon Baek</a>, <a href="/search/?searchtype=author&query=Hwang%2C+S+J">Sung Ju Hwang</a>, <a href="/search/?searchtype=author&query=Park%2C+J+C">Jong C. Park</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="2306.04293v1-abstract-short" style="display: inline;"> Open-Domain Conversational Question Answering (ODConvQA) aims at answering questions through a multi-turn conversation based on a retriever-reader pipeline, which retrieves passages and then predicts answers with them. However, such a pipeline approach not only makes the reader vulnerable to the errors propagated from the retriever, but also demands additional effort to develop both the retriever… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2306.04293v1-abstract-full').style.display = 'inline'; document.getElementById('2306.04293v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2306.04293v1-abstract-full" style="display: none;"> Open-Domain Conversational Question Answering (ODConvQA) aims at answering questions through a multi-turn conversation based on a retriever-reader pipeline, which retrieves passages and then predicts answers with them. However, such a pipeline approach not only makes the reader vulnerable to the errors propagated from the retriever, but also demands additional effort to develop both the retriever and the reader, which further makes it slower since they are not runnable in parallel. In this work, we propose a method to directly predict answers with a phrase retrieval scheme for a sequence of words, reducing the conventional two distinct subtasks into a single one. Also, for the first time, we study its capability for ODConvQA tasks. However, simply adopting it is largely problematic, due to the dependencies between previous and current turns in a conversation. To address this problem, we further introduce a novel contrastive learning strategy, making sure to reflect previous turns when retrieving the phrase for the current context, by maximizing representational similarities of consecutive turns in a conversation while minimizing irrelevant conversational contexts. We validate our model on two ODConvQA datasets, whose experimental results show that it substantially outperforms the relevant baselines with the retriever-reader. Code is available at: https://github.com/starsuzi/PRO-ConvQA. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2306.04293v1-abstract-full').style.display = 'none'; document.getElementById('2306.04293v1-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 June, 2023; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> June 2023. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Findings of ACL 2023</span> </p> </li> </ol> <nav class="pagination is-small is-centered breathe-horizontal" role="navigation" aria-label="pagination"> <a href="" class="pagination-previous is-invisible">Previous </a> <a href="/search/?searchtype=author&query=Baek%2C+J&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Baek%2C+J&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> </ul> </nav> <div class="is-hidden-tablet">  <span class="help" style="display: inline-block;"><a href="https://github.com/arXiv/arxiv-search/releases">Search v0.5.6 released 2020-02-24</a>  </span> </div> </div> </main> <footer> <div class="columns is-desktop" role="navigation" aria-label="Secondary">  <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/about">About</a></li> <li><a href="https://info.arxiv.org/help">Help</a></li> </ul> </div> <div class="column"> <ul class="nav-spaced"> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>contact arXiv</title><desc>Click here to contact arXiv</desc><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg> <a href="https://info.arxiv.org/help/contact.html"> Contact</a> </li> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>subscribe to arXiv mailings</title><desc>Click here to subscribe</desc><path d="M476 3.2L12.5 270.6c-18.1 10.4-15.8 35.6 2.2 43.2L121 358.4l287.3-253.2c5.5-4.9 13.3 2.6 8.6 8.3L176 407v80.5c0 23.6 28.5 32.9 42.5 15.8L282 426l124.6 52.2c14.2 6 30.4-2.9 33-18.2l72-432C515 7.8 493.3-6.8 476 3.2z"/></svg> <a href="https://info.arxiv.org/help/subscribe"> Subscribe</a> </li> </ul> </div> </div> </div>   <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/license/index.html">Copyright</a></li> <li><a href="https://info.arxiv.org/help/policies/privacy_policy.html">Privacy Policy</a></li> </ul> </div> <div class="column sorry-app-links"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/web_accessibility.html">Web Accessibility Assistance</a></li> <li> <p class="help"> <a class="a11y-main-link" href="https://status.arxiv.org" target="_blank">arXiv Operational Status <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 256 512" class="icon filter-dark_grey" role="presentation"><path d="M224.3 273l-136 136c-9.4 9.4-24.6 9.4-33.9 0l-22.6-22.6c-9.4-9.4-9.4-24.6 0-33.9l96.4-96.4-96.4-96.4c-9.4-9.4-9.4-24.6 0-33.9L54.3 103c9.4-9.4 24.6-9.4 33.9 0l136 136c9.5 9.4 9.5 24.6.1 34z"/></svg></a><br> Get status notifications via <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/email/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg>email</a> or <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/slack/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon filter-black" role="presentation"><path d="M94.12 315.1c0 25.9-21.16 47.06-47.06 47.06S0 341 0 315.1c0-25.9 21.16-47.06 47.06-47.06h47.06v47.06zm23.72 0c0-25.9 21.16-47.06 47.06-47.06s47.06 21.16 47.06 47.06v117.84c0 25.9-21.16 47.06-47.06 47.06s-47.06-21.16-47.06-47.06V315.1zm47.06-188.98c-25.9 0-47.06-21.16-47.06-47.06S139 32 164.9 32s47.06 21.16 47.06 47.06v47.06H164.9zm0 23.72c25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06H47.06C21.16 243.96 0 222.8 0 196.9s21.16-47.06 47.06-47.06H164.9zm188.98 47.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06h-47.06V196.9zm-23.72 0c0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06V79.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06V196.9zM283.1 385.88c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06v-47.06h47.06zm0-23.72c-25.9 0-47.06-21.16-47.06-47.06 0-25.9 21.16-47.06 47.06-47.06h117.84c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06H283.1z"/></svg>slack</a> </p> </li> </ul> </div> </div> </div>  </div> </footer> <script src="https://static.arxiv.org/static/base/1.0.0a5/js/member_acknowledgement.js"></script> </body> </html>

CINXE.COM

Search | arXiv e-print repository