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 419 results for author: <span class="mathjax">Qiu, X</span> </h1> </div> <div class="level-right is-hidden-mobile">  <span class="help" style="display: inline-block;"><a href="https://github.com/arXiv/arxiv-search/releases">Search v0.5.6 released 2020-02-24</a>  </span> </div> </div> <div class="content"> <form method="GET" action="/search/cs" aria-role="search"> Searching in archive <strong>cs</strong>. <a href="/search/?searchtype=author&query=Qiu%2C+X">Search in all archives.</a> <div class="field has-addons-tablet"> <div class="control is-expanded"> <label for="query" class="hidden-label">Search term or terms</label> <input class="input is-medium" id="query" name="query" placeholder="Search term..." type="text" value="Qiu, X"> </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=Qiu%2C+X&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="Qiu, X"> <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=Qiu%2C+X&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=200" class="pagination-link " aria-label="Page 5" aria-current="page">5 </a> </li> <li><span class="pagination-ellipsis">…</span></li> </ul> </nav> <ol class="breathe-horizontal" start="1"> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.19754">arXiv:2502.19754</a> <span> [<a href="https://arxiv.org/pdf/2502.19754">pdf</a>, <a href="https://arxiv.org/format/2502.19754">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"> Finding Local Diffusion Schr枚dinger Bridge using Kolmogorov-Arnold Network </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xingyu Qiu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+M">Mengying Yang</a>, <a href="/search/cs?searchtype=author&query=Ma%2C+X">Xinghua Ma</a>, <a href="/search/cs?searchtype=author&query=Li%2C+F">Fanding Li</a>, <a href="/search/cs?searchtype=author&query=Liang%2C+D">Dong Liang</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+G">Gongning Luo</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+W">Wei Wang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+K">Kuanquan Wang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+S">Shuo 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="2502.19754v1-abstract-short" style="display: inline;"> In image generation, Schr枚dinger Bridge (SB)-based methods theoretically enhance the efficiency and quality compared to the diffusion models by finding the least costly path between two distributions. However, they are computationally expensive and time-consuming when applied to complex image data. The reason is that they focus on fitting globally optimal paths in high-dimensional spaces, directly… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.19754v1-abstract-full').style.display = 'inline'; document.getElementById('2502.19754v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.19754v1-abstract-full" style="display: none;"> In image generation, Schr枚dinger Bridge (SB)-based methods theoretically enhance the efficiency and quality compared to the diffusion models by finding the least costly path between two distributions. However, they are computationally expensive and time-consuming when applied to complex image data. The reason is that they focus on fitting globally optimal paths in high-dimensional spaces, directly generating images as next step on the path using complex networks through self-supervised training, which typically results in a gap with the global optimum. Meanwhile, most diffusion models are in the same path subspace generated by weights $f_A(t)$ and $f_B(t)$, as they follow the paradigm ($x_t = f_A(t)x_{Img} + f_B(t)蔚$). To address the limitations of SB-based methods, this paper proposes for the first time to find local Diffusion Schr枚dinger Bridges (LDSB) in the diffusion path subspace, which strengthens the connection between the SB problem and diffusion models. Specifically, our method optimizes the diffusion paths using Kolmogorov-Arnold Network (KAN), which has the advantage of resistance to forgetting and continuous output. The experiment shows that our LDSB significantly improves the quality and efficiency of image generation using the same pre-trained denoising network and the KAN for optimising is only less than 0.1MB. The FID metric is reduced by \textbf{more than 15\%}, especially with a reduction of 48.50\% when NFE of DDIM is $5$ for the CelebA dataset. Code is available at https://github.com/Qiu-XY/LDSB. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.19754v1-abstract-full').style.display = 'none'; document.getElementById('2502.19754v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">16 pages, 10 figures, to be published in CVPR 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.19279">arXiv:2502.19279</a> <span> [<a href="https://arxiv.org/pdf/2502.19279">pdf</a>, <a href="https://arxiv.org/format/2502.19279">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"> CritiQ: Mining Data Quality Criteria from Human Preferences </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Guo%2C+H">Honglin Guo</a>, <a href="/search/cs?searchtype=author&query=Lv%2C+K">Kai Lv</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Liang%2C+T">Tianyi Liang</a>, <a href="/search/cs?searchtype=author&query=Xi%2C+Z">Zhiheng Xi</a>, <a href="/search/cs?searchtype=author&query=Song%2C+D">Demin Song</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Q">Qiuyinzhe Zhang</a>, <a href="/search/cs?searchtype=author&query=Sun%2C+Y">Yu Sun</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+K">Kai Chen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Gui%2C+T">Tao Gui</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.19279v1-abstract-short" style="display: inline;"> Language model heavily depends on high-quality data for optimal performance. Existing approaches rely on manually designed heuristics, the perplexity of existing models, training classifiers, or careful prompt engineering, which require significant expert experience and human annotation effort while introduce biases. We introduce CritiQ, a novel data selection method that automatically mines crite… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.19279v1-abstract-full').style.display = 'inline'; document.getElementById('2502.19279v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.19279v1-abstract-full" style="display: none;"> Language model heavily depends on high-quality data for optimal performance. Existing approaches rely on manually designed heuristics, the perplexity of existing models, training classifiers, or careful prompt engineering, which require significant expert experience and human annotation effort while introduce biases. We introduce CritiQ, a novel data selection method that automatically mines criteria from human preferences for data quality with only $\sim$30 human-annotated pairs and performs efficient data selection. The main component, CritiQ Flow, employs a manager agent to evolve quality criteria and worker agents to make pairwise judgments. We build a knowledge base that extracts quality criteria from previous work to boost CritiQ Flow. Compared to perplexity- and classifier- based methods, verbal criteria are more interpretable and possess reusable value. After deriving the criteria, we train the CritiQ Scorer to give quality scores and perform efficient data selection. We demonstrate the effectiveness of our method in the code, math, and logic domains, achieving high accuracy on human-annotated test sets. To validate the quality of the selected data, we continually train Llama 3.1 models and observe improved performance on downstream tasks compared to uniform sampling. Ablation studies validate the benefits of the knowledge base and the reflection process. We analyze how criteria evolve and the effectiveness of majority voting. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.19279v1-abstract-full').style.display = 'none'; document.getElementById('2502.19279v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.18917">arXiv:2502.18917</a> <span> [<a href="https://arxiv.org/pdf/2502.18917">pdf</a>, <a href="https://arxiv.org/format/2502.18917">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Programming Languages">cs.PL</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Software Engineering">cs.SE</span> </div> </div> <p class="title is-5 mathjax"> ClassInvGen: Class Invariant Synthesis using Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sun%2C+C">Chuyue Sun</a>, <a href="/search/cs?searchtype=author&query=Agashe%2C+V">Viraj Agashe</a>, <a href="/search/cs?searchtype=author&query=Chakraborty%2C+S">Saikat Chakraborty</a>, <a href="/search/cs?searchtype=author&query=Taneja%2C+J">Jubi Taneja</a>, <a href="/search/cs?searchtype=author&query=Barrett%2C+C">Clark Barrett</a>, <a href="/search/cs?searchtype=author&query=Dill%2C+D">David Dill</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaokang Qiu</a>, <a href="/search/cs?searchtype=author&query=Lahiri%2C+S+K">Shuvendu K. Lahiri</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.18917v1-abstract-short" style="display: inline;"> Formal program specifications in the form of preconditions, postconditions, and class invariants have several benefits for the construction and maintenance of programs. They not only aid in program understanding due to their unambiguous semantics but can also be enforced dynamically (or even statically when the language supports a formal verifier). However, synthesizing high-quality specifications… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.18917v1-abstract-full').style.display = 'inline'; document.getElementById('2502.18917v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.18917v1-abstract-full" style="display: none;"> Formal program specifications in the form of preconditions, postconditions, and class invariants have several benefits for the construction and maintenance of programs. They not only aid in program understanding due to their unambiguous semantics but can also be enforced dynamically (or even statically when the language supports a formal verifier). However, synthesizing high-quality specifications in an underlying programming language is limited by the expressivity of the specifications or the need to express them in a declarative manner. Prior work has demonstrated the potential of large language models (LLMs) for synthesizing high-quality method pre/postconditions for Python and Java, but does not consider class invariants. In this work, we describe ClassInvGen, a method for co-generating executable class invariants and test inputs to produce high-quality class invariants for a mainstream language such as C++, leveraging LLMs' ability to synthesize pure functions. We show that ClassInvGen outperforms a pure LLM-based technique to generate specifications (from code) as well as prior data-driven invariant inference techniques such as Daikon. We contribute a benchmark of standard C++ data structures along with a harness that can help measure both the correctness and completeness of generated specifications using tests and mutants. We also demonstrate its applicability to real-world code by performing a case study on several classes within a widely used and high-integrity C++ codebase. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.18917v1-abstract-full').style.display = 'none'; document.getElementById('2502.18917v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.17129">arXiv:2502.17129</a> <span> [<a href="https://arxiv.org/pdf/2502.17129">pdf</a>, <a href="https://arxiv.org/format/2502.17129">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"> Thus Spake Long-Context Large Language Model </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Liu%2C+X">Xiaoran Liu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+R">Ruixiao Li</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+M">Mianqiu Huang</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Z">Zhigeng Liu</a>, <a href="/search/cs?searchtype=author&query=Song%2C+Y">Yuerong Song</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=He%2C+S">Siyang He</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Q">Qiqi Wang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+L">Linlin Li</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Q">Qun Liu</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Y">Yaqian Zhou</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.17129v1-abstract-short" style="display: inline;"> Long context is an important topic in Natural Language Processing (NLP), running through the development of NLP architectures, and offers immense opportunities for Large Language Models (LLMs) giving LLMs the lifelong learning potential akin to humans. Unfortunately, the pursuit of a long context is accompanied by numerous obstacles. Nevertheless, long context remains a core competitive advantage… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.17129v1-abstract-full').style.display = 'inline'; document.getElementById('2502.17129v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.17129v1-abstract-full" style="display: none;"> Long context is an important topic in Natural Language Processing (NLP), running through the development of NLP architectures, and offers immense opportunities for Large Language Models (LLMs) giving LLMs the lifelong learning potential akin to humans. Unfortunately, the pursuit of a long context is accompanied by numerous obstacles. Nevertheless, long context remains a core competitive advantage for LLMs. In the past two years, the context length of LLMs has achieved a breakthrough extension to millions of tokens. Moreover, the research on long-context LLMs has expanded from length extrapolation to a comprehensive focus on architecture, infrastructure, training, and evaluation technologies. Inspired by the symphonic poem, Thus Spake Zarathustra, we draw an analogy between the journey of extending the context of LLM and the attempts of humans to transcend its mortality. In this survey, We will illustrate how LLM struggles between the tremendous need for a longer context and its equal need to accept the fact that it is ultimately finite. To achieve this, we give a global picture of the lifecycle of long-context LLMs from four perspectives: architecture, infrastructure, training, and evaluation, showcasing the full spectrum of long-context technologies. At the end of this survey, we will present 10 unanswered questions currently faced by long-context LLMs. We hope this survey can serve as a systematic introduction to the research on long-context LLMs. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.17129v1-abstract-full').style.display = 'none'; document.getElementById('2502.17129v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 24 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">a global picture of the lifecycle of long-context LLMs from four perspectives: architecture, infrastructure, training, and evaluation</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.16906">arXiv:2502.16906</a> <span> [<a href="https://arxiv.org/pdf/2502.16906">pdf</a>, <a href="https://arxiv.org/format/2502.16906">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"> AutoLogi: Automated Generation of Logic Puzzles for Evaluating Reasoning Abilities of Large Language Models </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhu%2C+Q">Qin Zhu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+F">Fei Huang</a>, <a href="/search/cs?searchtype=author&query=Peng%2C+R">Runyu Peng</a>, <a href="/search/cs?searchtype=author&query=Lu%2C+K">Keming Lu</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+B">Bowen Yu</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+Q">Qinyuan Cheng</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+J">Junyang Lin</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.16906v1-abstract-short" style="display: inline;"> While logical reasoning evaluation of Large Language Models (LLMs) has attracted significant attention, existing benchmarks predominantly rely on multiple-choice formats that are vulnerable to random guessing, leading to overestimated performance and substantial performance fluctuations. To obtain more accurate assessments of models' reasoning capabilities, we propose an automated method for synth… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16906v1-abstract-full').style.display = 'inline'; document.getElementById('2502.16906v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.16906v1-abstract-full" style="display: none;"> While logical reasoning evaluation of Large Language Models (LLMs) has attracted significant attention, existing benchmarks predominantly rely on multiple-choice formats that are vulnerable to random guessing, leading to overestimated performance and substantial performance fluctuations. To obtain more accurate assessments of models' reasoning capabilities, we propose an automated method for synthesizing open-ended logic puzzles, and use it to develop a bilingual benchmark, AutoLogi. Our approach features program-based verification and controllable difficulty levels, enabling more reliable evaluation that better distinguishes models' reasoning abilities. Extensive evaluation of eight modern LLMs shows that AutoLogi can better reflect true model capabilities, with performance scores spanning from 35% to 73% compared to the narrower range of 21% to 37% on the source multiple-choice dataset. Beyond benchmark creation, this synthesis method can generate high-quality training data by incorporating program verifiers into the rejection sampling process, enabling systematic enhancement of LLMs' reasoning capabilities across diverse datasets. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16906v1-abstract-full').style.display = 'none'; document.getElementById('2502.16906v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 24 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.16587">arXiv:2502.16587</a> <span> [<a href="https://arxiv.org/pdf/2502.16587">pdf</a>, <a href="https://arxiv.org/format/2502.16587">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> Human2Robot: Learning Robot Actions from Paired Human-Robot Videos </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xie%2C+S">Sicheng Xie</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+H">Haidong Cao</a>, <a href="/search/cs?searchtype=author&query=Weng%2C+Z">Zejia Weng</a>, <a href="/search/cs?searchtype=author&query=Xing%2C+Z">Zhen Xing</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+S">Shiwei Shen</a>, <a href="/search/cs?searchtype=author&query=Leng%2C+J">Jiaqi Leng</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Fu%2C+Y">Yanwei Fu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.16587v1-abstract-short" style="display: inline;"> Distilling knowledge from human demonstrations is a promising way for robots to learn and act. Existing work often overlooks the differences between humans and robots, producing unsatisfactory results. In this paper, we study how perfectly aligned human-robot pairs benefit robot learning. Capitalizing on VR-based teleportation, we introduce H\&R, a third-person dataset with 2,600 episodes, each of… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16587v1-abstract-full').style.display = 'inline'; document.getElementById('2502.16587v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.16587v1-abstract-full" style="display: none;"> Distilling knowledge from human demonstrations is a promising way for robots to learn and act. Existing work often overlooks the differences between humans and robots, producing unsatisfactory results. In this paper, we study how perfectly aligned human-robot pairs benefit robot learning. Capitalizing on VR-based teleportation, we introduce H\&R, a third-person dataset with 2,600 episodes, each of which captures the fine-grained correspondence between human hands and robot gripper. Inspired by the recent success of diffusion models, we introduce Human2Robot, an end-to-end diffusion framework that formulates learning from human demonstrates as a generative task. Human2Robot fully explores temporal dynamics in human videos to generate robot videos and predict actions at the same time. Through comprehensive evaluations of 8 seen, changed and unseen tasks in real-world settings, we demonstrate that Human2Robot can not only generate high-quality robot videos but also excel in seen tasks and generalize to unseen objects, backgrounds and even new tasks effortlessly. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.16587v1-abstract-full').style.display = 'none'; document.getElementById('2502.16587v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.15770">arXiv:2502.15770</a> <span> [<a href="https://arxiv.org/pdf/2502.15770">pdf</a>, <a href="https://arxiv.org/format/2502.15770">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Software Engineering">cs.SE</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> </div> </div> <p class="title is-5 mathjax"> Performance Review on LLM for solving leetcode problems </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wang%2C+L">Lun Wang</a>, <a href="/search/cs?searchtype=author&query=Shi%2C+C">Chuanqi Shi</a>, <a href="/search/cs?searchtype=author&query=Du%2C+S">Shaoshui Du</a>, <a href="/search/cs?searchtype=author&query=Tao%2C+Y">Yiyi Tao</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+Y">Yixian Shen</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+H">Hang Zheng</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinyu Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.15770v1-abstract-short" style="display: inline;"> This paper presents a comprehensive performance evaluation of Large Language Models (LLMs) in solving programming challenges from Leetcode, a widely used platform for algorithm practice and technical interviews. We began by crawling the Leetcode website to collect a diverse set of problems encompassing various difficulty levels and topics. Using this dataset, we generated solutions with multiple L… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.15770v1-abstract-full').style.display = 'inline'; document.getElementById('2502.15770v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.15770v1-abstract-full" style="display: none;"> This paper presents a comprehensive performance evaluation of Large Language Models (LLMs) in solving programming challenges from Leetcode, a widely used platform for algorithm practice and technical interviews. We began by crawling the Leetcode website to collect a diverse set of problems encompassing various difficulty levels and topics. Using this dataset, we generated solutions with multiple LLMs, including GPT-4 and GPT-3.5-turbo (ChatGPT-turbo). The generated solutions were systematically evaluated for correctness and efficiency. We employed the pass@k metric to assess the success rates within a given number of attempts and analyzed the runtime performance of the solutions. Our results highlight the strengths and limitations of current LLMs [10] in code generation and problem-solving tasks, providing insights into their potential applications and areas for improvement in automated programming assistance. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.15770v1-abstract-full').style.display = 'none'; document.getElementById('2502.15770v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 16 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.15694">arXiv:2502.15694</a> <span> [<a href="https://arxiv.org/pdf/2502.15694">pdf</a>, <a href="https://arxiv.org/format/2502.15694">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="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> <p class="title is-5 mathjax"> Image Fusion for Cross-Domain Sequential Recommendation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wu%2C+W">Wangyu Wu</a>, <a href="/search/cs?searchtype=author&query=Song%2C+S">Siqi Song</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xianglin Qiu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xiaowei Huang</a>, <a href="/search/cs?searchtype=author&query=Ma%2C+F">Fei Ma</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+J">Jimin Xiao</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.15694v2-abstract-short" style="display: inline;"> Cross-Domain Sequential Recommendation (CDSR) aims to predict future user interactions based on historical interactions across multiple domains. The key challenge in CDSR is effectively capturing cross-domain user preferences by fully leveraging both intra-sequence and inter-sequence item interactions. In this paper, we propose a novel method, Image Fusion for Cross-Domain Sequential Recommendatio… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.15694v2-abstract-full').style.display = 'inline'; document.getElementById('2502.15694v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.15694v2-abstract-full" style="display: none;"> Cross-Domain Sequential Recommendation (CDSR) aims to predict future user interactions based on historical interactions across multiple domains. The key challenge in CDSR is effectively capturing cross-domain user preferences by fully leveraging both intra-sequence and inter-sequence item interactions. In this paper, we propose a novel method, Image Fusion for Cross-Domain Sequential Recommendation (IFCDSR), which incorporates item image information to better capture visual preferences. Our approach integrates a frozen CLIP model to generate image embeddings, enriching original item embeddings with visual data from both intra-sequence and inter-sequence interactions. Additionally, we employ a multiple attention layer to capture cross-domain interests, enabling joint learning of single-domain and cross-domain user preferences. To validate the effectiveness of IFCDSR, we re-partitioned four e-commerce datasets and conducted extensive experiments. Results demonstrate that IFCDSR significantly outperforms existing methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.15694v2-abstract-full').style.display = 'none'; document.getElementById('2502.15694v2-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 26 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 30 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.14837">arXiv:2502.14837</a> <span> [<a href="https://arxiv.org/pdf/2502.14837">pdf</a>, <a href="https://arxiv.org/format/2502.14837">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"> Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Ji%2C+T">Tao Ji</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+B">Bin Guo</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Y">Yuanbin Wu</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+L">Lixing Shen</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Z">Zhan Chen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Q">Qi Zhang</a>, <a href="/search/cs?searchtype=author&query=Gui%2C+T">Tao Gui</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.14837v1-abstract-short" style="display: inline;"> Multi-head Latent Attention (MLA) is an innovative architecture proposed by DeepSeek, designed to ensure efficient and economical inference by significantly compressing the Key-Value (KV) cache into a latent vector. Compared to MLA, standard LLMs employing Multi-Head Attention (MHA) and its variants such as Grouped-Query Attention (GQA) exhibit significant cost disadvantages. Enabling well-trained… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.14837v1-abstract-full').style.display = 'inline'; document.getElementById('2502.14837v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.14837v1-abstract-full" style="display: none;"> Multi-head Latent Attention (MLA) is an innovative architecture proposed by DeepSeek, designed to ensure efficient and economical inference by significantly compressing the Key-Value (KV) cache into a latent vector. Compared to MLA, standard LLMs employing Multi-Head Attention (MHA) and its variants such as Grouped-Query Attention (GQA) exhibit significant cost disadvantages. Enabling well-trained LLMs (e.g., Llama) to rapidly adapt to MLA without pre-training from scratch is both meaningful and challenging. This paper proposes the first data-efficient fine-tuning method for transitioning from MHA to MLA (MHA2MLA), which includes two key components: for partial-RoPE, we remove RoPE from dimensions of queries and keys that contribute less to the attention scores, for low-rank approximation, we introduce joint SVD approximations based on the pre-trained parameters of keys and values. These carefully designed strategies enable MHA2MLA to recover performance using only a small fraction (0.3% to 0.6%) of the data, significantly reducing inference costs while seamlessly integrating with compression techniques such as KV cache quantization. For example, the KV cache size of Llama2-7B is reduced by 92.19%, with only a 0.5% drop in LongBench performance. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.14837v1-abstract-full').style.display = 'none'; document.getElementById('2502.14837v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 20 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">16 pages, 8 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/2502.12264">arXiv:2502.12264</a> <span> [<a href="https://arxiv.org/pdf/2502.12264">pdf</a>, <a href="https://arxiv.org/ps/2502.12264">ps</a>, <a href="https://arxiv.org/format/2502.12264">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Theoretical Economics">econ.TH</span> <span class="tag is-small is-grey 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="Computer Science and Game Theory">cs.GT</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"> Multi-dimensional Test Design </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaoyun Qiu</a>, <a href="/search/cs?searchtype=author&query=Shan%2C+L">Liren Shan</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.12264v1-abstract-short" style="display: inline;"> How should one jointly design tests and the arrangement of agencies to administer these tests (testing procedure)? To answer this question, we analyze a model where a principal must use multiple tests to screen an agent with a multi-dimensional type, knowing that the agent can change his type at a cost. We identify a new tradeoff between setting difficult tests and using a difficult testing proced… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12264v1-abstract-full').style.display = 'inline'; document.getElementById('2502.12264v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.12264v1-abstract-full" style="display: none;"> How should one jointly design tests and the arrangement of agencies to administer these tests (testing procedure)? To answer this question, we analyze a model where a principal must use multiple tests to screen an agent with a multi-dimensional type, knowing that the agent can change his type at a cost. We identify a new tradeoff between setting difficult tests and using a difficult testing procedure. We compare two settings: (1) the agent only misrepresents his type (manipulation) and (2) the agent improves his actual type (investment). Examples include interviews, regulations, and data classification. We show that in the manipulation setting, stringent tests combined with an easy procedure, i.e., offering tests sequentially in a fixed order, is optimal. In contrast, in the investment setting, non-stringent tests with a difficult procedure, i.e., offering tests simultaneously, is optimal; however, under mild conditions offering them sequentially in a random order may be as good. Our results suggest that whether the agent manipulates or invests in his type determines which arrangement of agencies is optimal. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12264v1-abstract-full').style.display = 'none'; document.getElementById('2502.12264v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.12215">arXiv:2502.12215</a> <span> [<a href="https://arxiv.org/pdf/2502.12215">pdf</a>, <a href="https://arxiv.org/format/2502.12215">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"> Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities? </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zeng%2C+Z">Zhiyuan Zeng</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+Q">Qinyuan Cheng</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+Z">Zhangyue Yin</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Y">Yunhua Zhou</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.12215v1-abstract-short" style="display: inline;"> The advent of test-time scaling in large language models (LLMs), exemplified by OpenAI's o1 series, has advanced reasoning capabilities by scaling computational resource allocation during inference. While successors like QwQ, Deepseek-R1 (R1) and LIMO replicate these advancements, whether these models truly possess test-time scaling capabilities remains underexplored. This study found that longer… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12215v1-abstract-full').style.display = 'inline'; document.getElementById('2502.12215v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.12215v1-abstract-full" style="display: none;"> The advent of test-time scaling in large language models (LLMs), exemplified by OpenAI's o1 series, has advanced reasoning capabilities by scaling computational resource allocation during inference. While successors like QwQ, Deepseek-R1 (R1) and LIMO replicate these advancements, whether these models truly possess test-time scaling capabilities remains underexplored. This study found that longer CoTs of these o1-like models do not consistently enhance accuracy; in fact, correct solutions are often shorter than incorrect ones for the same questions. Further investigation shows this phenomenon is closely related to models' self-revision capabilities - longer CoTs contain more self-revisions, which often lead to performance degradation. We then compare sequential and parallel scaling strategies on QwQ, R1 and LIMO, finding that parallel scaling achieves better coverage and scalability. Based on these insights, we propose Shortest Majority Vote, a method that combines parallel scaling strategies with CoT length characteristics, significantly improving models' test-time scalability compared to conventional majority voting approaches. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.12215v1-abstract-full').style.display = 'none'; document.getElementById('2502.12215v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.11520">arXiv:2502.11520</a> <span> [<a href="https://arxiv.org/pdf/2502.11520">pdf</a>, <a href="https://arxiv.org/format/2502.11520">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"> AURORA:Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verification </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Tan%2C+X">Xiaoyu Tan</a>, <a href="/search/cs?searchtype=author&query=Yao%2C+T">Tianchu Yao</a>, <a href="/search/cs?searchtype=author&query=Qu%2C+C">Chao Qu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+B">Bin Li</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+M">Minghao Yang</a>, <a href="/search/cs?searchtype=author&query=Lu%2C+D">Dakuan Lu</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+H">Haozhe Wang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xihe Qiu</a>, <a href="/search/cs?searchtype=author&query=Chu%2C+W">Wei Chu</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+Y">Yinghui Xu</a>, <a href="/search/cs?searchtype=author&query=Qi%2C+Y">Yuan Qi</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.11520v1-abstract-short" style="display: inline;"> The reasoning capabilities of advanced large language models (LLMs) like o1 have revolutionized artificial intelligence applications. Nevertheless, evaluating and optimizing complex reasoning processes remain significant challenges due to diverse policy distributions and the inherent limitations of human effort and accuracy. In this paper, we present AURORA, a novel automated framework for trainin… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11520v1-abstract-full').style.display = 'inline'; document.getElementById('2502.11520v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.11520v1-abstract-full" style="display: none;"> The reasoning capabilities of advanced large language models (LLMs) like o1 have revolutionized artificial intelligence applications. Nevertheless, evaluating and optimizing complex reasoning processes remain significant challenges due to diverse policy distributions and the inherent limitations of human effort and accuracy. In this paper, we present AURORA, a novel automated framework for training universal process reward models (PRMs) using ensemble prompting and reverse verification. The framework employs a two-phase approach: First, it uses diverse prompting strategies and ensemble methods to perform automated annotation and evaluation of processes, ensuring robust assessments for reward learning. Second, it leverages practical reference answers for reverse verification, enhancing the model's ability to validate outputs and improving training accuracy. To assess the framework's performance, we extend beyond the existing ProcessBench benchmark by introducing UniversalBench, which evaluates reward predictions across full trajectories under diverse policy distribtion with long Chain-of-Thought (CoT) outputs. Experimental results demonstrate that AURORA enhances process evaluation accuracy, improves PRMs' accuracy for diverse policy distributions and long-CoT responses. The project will be open-sourced at https://auroraprm.github.io/. The Universal-PRM-7B is available at https://huggingface.co/infly/Universal-PRM-7B. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11520v1-abstract-full').style.display = 'none'; document.getElementById('2502.11520v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">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/2502.11476">arXiv:2502.11476</a> <span> [<a href="https://arxiv.org/pdf/2502.11476">pdf</a>, <a href="https://arxiv.org/format/2502.11476">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"> FastMCTS: A Simple Sampling Strategy for Data Synthesis </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Li%2C+P">Peiji Li</a>, <a href="/search/cs?searchtype=author&query=Lv%2C+K">Kai Lv</a>, <a href="/search/cs?searchtype=author&query=Shao%2C+Y">Yunfan Shao</a>, <a href="/search/cs?searchtype=author&query=Ma%2C+Y">Yichuan Ma</a>, <a href="/search/cs?searchtype=author&query=Li%2C+L">Linyang Li</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+X">Xiaoqing Zheng</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.11476v1-abstract-short" style="display: inline;"> Synthetic high-quality multi-step reasoning data can significantly enhance the performance of large language models on various tasks. However, most existing methods rely on rejection sampling, which generates trajectories independently and suffers from inefficiency and imbalanced sampling across problems of varying difficulty. In this work, we introduce FastMCTS, an innovative data synthesis strat… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11476v1-abstract-full').style.display = 'inline'; document.getElementById('2502.11476v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.11476v1-abstract-full" style="display: none;"> Synthetic high-quality multi-step reasoning data can significantly enhance the performance of large language models on various tasks. However, most existing methods rely on rejection sampling, which generates trajectories independently and suffers from inefficiency and imbalanced sampling across problems of varying difficulty. In this work, we introduce FastMCTS, an innovative data synthesis strategy inspired by Monte Carlo Tree Search. FastMCTS provides a more efficient sampling method for multi-step reasoning data, offering step-level evaluation signals and promoting balanced sampling across problems of different difficulty levels. Experiments on both English and Chinese reasoning datasets demonstrate that FastMCTS generates over 30\% more correct reasoning paths compared to rejection sampling as the number of generated tokens scales up. Furthermore, under comparable synthetic data budgets, models trained on FastMCTS-generated data outperform those trained on rejection sampling data by 3.9\% across multiple benchmarks. As a lightweight sampling strategy, FastMCTS offers a practical and efficient alternative for synthesizing high-quality reasoning data. Our code will be released soon. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11476v1-abstract-full').style.display = 'none'; document.getElementById('2502.11476v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">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/2502.11460">arXiv:2502.11460</a> <span> [<a href="https://arxiv.org/pdf/2502.11460">pdf</a>, <a href="https://arxiv.org/format/2502.11460">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="Software Engineering">cs.SE</span> </div> </div> <p class="title is-5 mathjax"> UnitCoder: Scalable Iterative Code Synthesis with Unit Test Guidance </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Ma%2C+Y">Yichuan Ma</a>, <a href="/search/cs?searchtype=author&query=Shao%2C+Y">Yunfan Shao</a>, <a href="/search/cs?searchtype=author&query=Li%2C+P">Peiji Li</a>, <a href="/search/cs?searchtype=author&query=Song%2C+D">Demin Song</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Li%2C+L">Linyang Li</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+K">Kai 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="2502.11460v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, yet code generation remains a major challenge. Current approaches for obtaining high-quality code data primarily focus on (i) collecting large-scale pre-training data and (ii) synthesizing instruction data through prompt engineering with powerful models. While pre-training data faces quality consistency issues… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11460v1-abstract-full').style.display = 'inline'; document.getElementById('2502.11460v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.11460v1-abstract-full" style="display: none;"> Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, yet code generation remains a major challenge. Current approaches for obtaining high-quality code data primarily focus on (i) collecting large-scale pre-training data and (ii) synthesizing instruction data through prompt engineering with powerful models. While pre-training data faces quality consistency issues, instruction-based synthesis suffers from limited instruction diversity and inherent biases of LLMs. To address this gap, we introduce UnitCoder, a systematic pipeline leveraging model-generated unit tests to both guide and validate the code generation process. Combined with large-scale package-based retrieval from pre-training corpus, we generate a dataset of 500K+ verifiable programs containing diverse API calls. Evaluations on multiple Python benchmarks (BigCodeBench, HumanEval, MBPP) demonstrate that models fine-tuned on our synthetic data exhibit consistent performance improvements. Notably, Llama3.1-8B and InternLM2.5-7B improve from 31\% and 28\% to 40\% and 39\% success rates on BigCodeBench, respectively. Our work presents a scalable approach that leverages model-generated unit tests to guide the synthesis of high-quality code data from pre-training corpora, demonstrating the potential for producing diverse and high-quality post-training data at scale. All code and data will be released (https://github.com). <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.11460v1-abstract-full').style.display = 'none'; document.getElementById('2502.11460v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">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/2502.10721">arXiv:2502.10721</a> <span> [<a href="https://arxiv.org/pdf/2502.10721">pdf</a>, <a href="https://arxiv.org/format/2502.10721">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> A Comprehensive Survey of Deep Learning for Multivariate Time Series Forecasting: A Channel Strategy Perspective </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiangfei Qiu</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+H">Hanyin Cheng</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+X">Xingjian Wu</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jilin Hu</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+C">Chenjuan Guo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.10721v1-abstract-short" style="display: inline;"> Multivariate Time Series Forecasting (MTSF) plays a crucial role across diverse fields, ranging from economic, energy, to traffic. In recent years, deep learning has demonstrated outstanding performance in MTSF tasks. In MTSF, modeling the correlations among different channels is critical, as leveraging information from other related channels can significantly improve the prediction accuracy of a… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.10721v1-abstract-full').style.display = 'inline'; document.getElementById('2502.10721v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.10721v1-abstract-full" style="display: none;"> Multivariate Time Series Forecasting (MTSF) plays a crucial role across diverse fields, ranging from economic, energy, to traffic. In recent years, deep learning has demonstrated outstanding performance in MTSF tasks. In MTSF, modeling the correlations among different channels is critical, as leveraging information from other related channels can significantly improve the prediction accuracy of a specific channel. This study systematically reviews the channel modeling strategies for time series and proposes a taxonomy organized into three hierarchical levels: the strategy perspective, the mechanism perspective, and the characteristic perspective. On this basis, we provide a structured analysis of these methods and conduct an in-depth examination of the advantages and limitations of different channel strategies. Finally, we summarize and discuss some future research directions to provide useful research guidance. Moreover, we maintain an up-to-date Github repository (https://github.com/decisionintelligence/CS4TS) which includes all the papers discussed in the survey. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.10721v1-abstract-full').style.display = 'none'; document.getElementById('2502.10721v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.07218">arXiv:2502.07218</a> <span> [<a href="https://arxiv.org/pdf/2502.07218">pdf</a>, <a href="https://arxiv.org/format/2502.07218">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"> LUNAR: LLM Unlearning via Neural Activation Redirection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Shen%2C+W+F">William F. Shen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinchi Qiu</a>, <a href="/search/cs?searchtype=author&query=Kurmanji%2C+M">Meghdad Kurmanji</a>, <a href="/search/cs?searchtype=author&query=Iacob%2C+A">Alex Iacob</a>, <a href="/search/cs?searchtype=author&query=Sani%2C+L">Lorenzo Sani</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yihong Chen</a>, <a href="/search/cs?searchtype=author&query=Cancedda%2C+N">Nicola Cancedda</a>, <a href="/search/cs?searchtype=author&query=Lane%2C+N+D">Nicholas D. Lane</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.07218v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) benefit from training on ever larger amounts of textual data, but as a result, they increasingly incur the risk of leaking private information. The ability to selectively remove knowledge from LLMs is, therefore, a highly desirable capability. In this paper, we propose LUNAR, a novel unlearning methodology grounded in the Linear Representation Hypothesis. LUNAR operate… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.07218v1-abstract-full').style.display = 'inline'; document.getElementById('2502.07218v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.07218v1-abstract-full" style="display: none;"> Large Language Models (LLMs) benefit from training on ever larger amounts of textual data, but as a result, they increasingly incur the risk of leaking private information. The ability to selectively remove knowledge from LLMs is, therefore, a highly desirable capability. In this paper, we propose LUNAR, a novel unlearning methodology grounded in the Linear Representation Hypothesis. LUNAR operates by redirecting the representations of unlearned data to regions that trigger the model's inherent ability to express its inability to answer. LUNAR achieves state-of-the-art unlearning performance while significantly enhancing the controllability of the unlearned model during inference. Specifically, LUNAR achieves between 2.9x to 11.7x improvements on combined "unlearning efficacy" and "model utility" score ("Deviation Score") on the PISTOL dataset across various base models. We also demonstrate, through quantitative analysis and qualitative examples, LUNAR's superior controllability in generating coherent and contextually aware responses, mitigating undesired side effects of existing methods. Moreover, we demonstrate that LUNAR is robust against white-box adversarial attacks and versatile in handling real-world scenarios, such as processing sequential unlearning requests. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.07218v1-abstract-full').style.display = 'none'; document.getElementById('2502.07218v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 10 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.05694">arXiv:2502.05694</a> <span> [<a href="https://arxiv.org/pdf/2502.05694">pdf</a>, <a href="https://arxiv.org/format/2502.05694">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"> Zero-Shot End-to-End Relation Extraction in Chinese: A Comparative Study of Gemini, LLaMA and ChatGPT </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Du%2C+S">Shaoshuai Du</a>, <a href="/search/cs?searchtype=author&query=Tao%2C+Y">Yiyi Tao</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+Y">Yixian Shen</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+H">Hang Zhang</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+Y">Yanxin Shen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinyu Qiu</a>, <a href="/search/cs?searchtype=author&query=Shi%2C+C">Chuanqi Shi</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.05694v1-abstract-short" style="display: inline;"> This study investigates the performance of various large language models (LLMs) on zero-shot end-to-end relation extraction (RE) in Chinese, a task that integrates entity recognition and relation extraction without requiring annotated data. While LLMs show promise for RE, most prior work focuses on English or assumes pre-annotated entities, leaving their effectiveness in Chinese RE largely unexplo… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05694v1-abstract-full').style.display = 'inline'; document.getElementById('2502.05694v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.05694v1-abstract-full" style="display: none;"> This study investigates the performance of various large language models (LLMs) on zero-shot end-to-end relation extraction (RE) in Chinese, a task that integrates entity recognition and relation extraction without requiring annotated data. While LLMs show promise for RE, most prior work focuses on English or assumes pre-annotated entities, leaving their effectiveness in Chinese RE largely unexplored. To bridge this gap, we evaluate ChatGPT, Gemini, and LLaMA based on accuracy, efficiency, and adaptability. ChatGPT demonstrates the highest overall performance, balancing precision and recall, while Gemini achieves the fastest inference speed, making it suitable for real-time applications. LLaMA underperforms in both accuracy and latency, highlighting the need for further adaptation. Our findings provide insights into the strengths and limitations of LLMs for zero-shot Chinese RE, shedding light on trade-offs between accuracy and efficiency. This study serves as a foundation for future research aimed at improving LLM adaptability to complex linguistic tasks in Chinese NLP. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05694v1-abstract-full').style.display = 'none'; document.getElementById('2502.05694v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.05206">arXiv:2502.05206</a> <span> [<a href="https://arxiv.org/pdf/2502.05206">pdf</a>, <a href="https://arxiv.org/format/2502.05206">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Cryptography and Security">cs.CR</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="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="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> Safety at Scale: A Comprehensive Survey of Large Model Safety </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Ma%2C+X">Xingjun Ma</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+Y">Yifeng Gao</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yixu Wang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+R">Ruofan Wang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+X">Xin Wang</a>, <a href="/search/cs?searchtype=author&query=Sun%2C+Y">Ye Sun</a>, <a href="/search/cs?searchtype=author&query=Ding%2C+Y">Yifan Ding</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+H">Hengyuan Xu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yunhao Chen</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+Y">Yunhan Zhao</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+H">Hanxun Huang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yige Li</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jiaming Zhang</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+X">Xiang Zheng</a>, <a href="/search/cs?searchtype=author&query=Bai%2C+Y">Yang Bai</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jingfeng Zhang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yiming Li</a>, <a href="/search/cs?searchtype=author&query=Sun%2C+J">Jun Sun</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+C">Cong Wang</a>, <a href="/search/cs?searchtype=author&query=Gu%2C+J">Jindong Gu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+B">Baoyuan Wu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+S">Siheng Chen</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+T">Tianwei Zhang</a> , et al. (19 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="2502.05206v2-abstract-short" style="display: inline;"> The rapid advancement of large models, driven by their exceptional abilities in learning and generalization through large-scale pre-training, has reshaped the landscape of Artificial Intelligence (AI). These models are now foundational to a wide range of applications, including conversational AI, recommendation systems, autonomous driving, content generation, medical diagnostics, and scientific di… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05206v2-abstract-full').style.display = 'inline'; document.getElementById('2502.05206v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.05206v2-abstract-full" style="display: none;"> The rapid advancement of large models, driven by their exceptional abilities in learning and generalization through large-scale pre-training, has reshaped the landscape of Artificial Intelligence (AI). These models are now foundational to a wide range of applications, including conversational AI, recommendation systems, autonomous driving, content generation, medical diagnostics, and scientific discovery. However, their widespread deployment also exposes them to significant safety risks, raising concerns about robustness, reliability, and ethical implications. This survey provides a systematic review of current safety research on large models, covering Vision Foundation Models (VFMs), Large Language Models (LLMs), Vision-Language Pre-training (VLP) models, Vision-Language Models (VLMs), Diffusion Models (DMs), and large-model-based Agents. Our contributions are summarized as follows: (1) We present a comprehensive taxonomy of safety threats to these models, including adversarial attacks, data poisoning, backdoor attacks, jailbreak and prompt injection attacks, energy-latency attacks, data and model extraction attacks, and emerging agent-specific threats. (2) We review defense strategies proposed for each type of attacks if available and summarize the commonly used datasets and benchmarks for safety research. (3) Building on this, we identify and discuss the open challenges in large model safety, emphasizing the need for comprehensive safety evaluations, scalable and effective defense mechanisms, and sustainable data practices. More importantly, we highlight the necessity of collective efforts from the research community and international collaboration. Our work can serve as a useful reference for researchers and practitioners, fostering the ongoing development of comprehensive defense systems and platforms to safeguard AI models. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05206v2-abstract-full').style.display = 'none'; document.getElementById('2502.05206v2-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 2 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">47 pages, 3 figures, 11 tables GitHub: https://github.com/xingjunm/Awesome-Large-Model-Safety</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.05173">arXiv:2502.05173</a> <span> [<a href="https://arxiv.org/pdf/2502.05173">pdf</a>, <a href="https://arxiv.org/format/2502.05173">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"> VideoRoPE: What Makes for Good Video Rotary Position Embedding? </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wei%2C+X">Xilin Wei</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+X">Xiaoran Liu</a>, <a href="/search/cs?searchtype=author&query=Zang%2C+Y">Yuhang Zang</a>, <a href="/search/cs?searchtype=author&query=Dong%2C+X">Xiaoyi Dong</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+P">Pan Zhang</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+Y">Yuhang Cao</a>, <a href="/search/cs?searchtype=author&query=Tong%2C+J">Jian Tong</a>, <a href="/search/cs?searchtype=author&query=Duan%2C+H">Haodong Duan</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+J">Jiaqi Wang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+D">Dahua Lin</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.05173v1-abstract-short" style="display: inline;"> While Rotary Position Embedding (RoPE) and its variants are widely adopted for their long-context capabilities, the extension of the 1D RoPE to video, with its complex spatio-temporal structure, remains an open challenge. This work first introduces a comprehensive analysis that identifies four key characteristics essential for the effective adaptation of RoPE to video, which have not been fully co… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05173v1-abstract-full').style.display = 'inline'; document.getElementById('2502.05173v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.05173v1-abstract-full" style="display: none;"> While Rotary Position Embedding (RoPE) and its variants are widely adopted for their long-context capabilities, the extension of the 1D RoPE to video, with its complex spatio-temporal structure, remains an open challenge. This work first introduces a comprehensive analysis that identifies four key characteristics essential for the effective adaptation of RoPE to video, which have not been fully considered in prior work. As part of our analysis, we introduce a challenging V-NIAH-D (Visual Needle-In-A-Haystack with Distractors) task, which adds periodic distractors into V-NIAH. The V-NIAH-D task demonstrates that previous RoPE variants, lacking appropriate temporal dimension allocation, are easily misled by distractors. Based on our analysis, we introduce \textbf{VideoRoPE}, with a \textit{3D structure} designed to preserve spatio-temporal relationships. VideoRoPE features \textit{low-frequency temporal allocation} to mitigate periodic oscillations, a \textit{diagonal layout} to maintain spatial symmetry, and \textit{adjustable temporal spacing} to decouple temporal and spatial indexing. VideoRoPE consistently surpasses previous RoPE variants, across diverse downstream tasks such as long video retrieval, video understanding, and video hallucination. Our code will be available at \href{https://github.com/Wiselnn570/VideoRoPE}{https://github.com/Wiselnn570/VideoRoPE}. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.05173v1-abstract-full').style.display = 'none'; document.getElementById('2502.05173v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.04358">arXiv:2502.04358</a> <span> [<a href="https://arxiv.org/pdf/2502.04358">pdf</a>, <a href="https://arxiv.org/format/2502.04358">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="Computational Complexity">cs.CC</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="Neural and Evolutionary Computing">cs.NE</span> </div> </div> <p class="title is-5 mathjax"> Position: Scaling LLM Agents Requires Asymptotic Analysis with LLM Primitives </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Meyerson%2C+E">Elliot Meyerson</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xin Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.04358v1-abstract-short" style="display: inline;"> Decomposing hard problems into subproblems often makes them easier and more efficient to solve. With large language models (LLMs) crossing critical reliability thresholds for a growing slate of capabilities, there is an increasing effort to decompose systems into sets of LLM-based agents, each of whom can be delegated sub-tasks. However, this decomposition (even when automated) is often intuitive,… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.04358v1-abstract-full').style.display = 'inline'; document.getElementById('2502.04358v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.04358v1-abstract-full" style="display: none;"> Decomposing hard problems into subproblems often makes them easier and more efficient to solve. With large language models (LLMs) crossing critical reliability thresholds for a growing slate of capabilities, there is an increasing effort to decompose systems into sets of LLM-based agents, each of whom can be delegated sub-tasks. However, this decomposition (even when automated) is often intuitive, e.g., based on how a human might assign roles to members of a human team. How close are these role decompositions to optimal? This position paper argues that asymptotic analysis with LLM primitives is needed to reason about the efficiency of such decomposed systems, and that insights from such analysis will unlock opportunities for scaling them. By treating the LLM forward pass as the atomic unit of computational cost, one can separate out the (often opaque) inner workings of a particular LLM from the inherent efficiency of how a set of LLMs are orchestrated to solve hard problems. In other words, if we want to scale the deployment of LLMs to the limit, instead of anthropomorphizing LLMs, asymptotic analysis with LLM primitives should be used to reason about and develop more powerful decompositions of large problems into LLM agents. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.04358v1-abstract-full').style.display = 'none'; document.getElementById('2502.04358v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">12 pages including references</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2502.02590">arXiv:2502.02590</a> <span> [<a href="https://arxiv.org/pdf/2502.02590">pdf</a>, <a href="https://arxiv.org/format/2502.02590">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> Articulate AnyMesh: Open-Vocabulary 3D Articulated Objects Modeling </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaowen Qiu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+J">Jincheng Yang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yian Wang</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Z">Zhehuan Chen</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yufei Wang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+T">Tsun-Hsuan Wang</a>, <a href="/search/cs?searchtype=author&query=Xian%2C+Z">Zhou Xian</a>, <a href="/search/cs?searchtype=author&query=Gan%2C+C">Chuang Gan</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2502.02590v1-abstract-short" style="display: inline;"> 3D articulated objects modeling has long been a challenging problem, since it requires to capture both accurate surface geometries and semantically meaningful and spatially precise structures, parts, and joints. Existing methods heavily depend on training data from a limited set of handcrafted articulated object categories (e.g., cabinets and drawers), which restricts their ability to model a wide… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.02590v1-abstract-full').style.display = 'inline'; document.getElementById('2502.02590v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2502.02590v1-abstract-full" style="display: none;"> 3D articulated objects modeling has long been a challenging problem, since it requires to capture both accurate surface geometries and semantically meaningful and spatially precise structures, parts, and joints. Existing methods heavily depend on training data from a limited set of handcrafted articulated object categories (e.g., cabinets and drawers), which restricts their ability to model a wide range of articulated objects in an open-vocabulary context. To address these limitations, we propose Articulate Anymesh, an automated framework that is able to convert any rigid 3D mesh into its articulated counterpart in an open-vocabulary manner. Given a 3D mesh, our framework utilizes advanced Vision-Language Models and visual prompting techniques to extract semantic information, allowing for both the segmentation of object parts and the construction of functional joints. Our experiments show that Articulate Anymesh can generate large-scale, high-quality 3D articulated objects, including tools, toys, mechanical devices, and vehicles, significantly expanding the coverage of existing 3D articulated object datasets. Additionally, we show that these generated assets can facilitate the acquisition of new articulated object manipulation skills in simulation, which can then be transferred to a real robotic system. Our Github website is https://articulate-anymesh.github.io. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2502.02590v1-abstract-full').style.display = 'none'; document.getElementById('2502.02590v1-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> February 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.16629">arXiv:2501.16629</a> <span> [<a href="https://arxiv.org/pdf/2501.16629">pdf</a>, <a href="https://arxiv.org/format/2501.16629">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="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> CHiP: Cross-modal Hierarchical Direct Preference Optimization for Multimodal LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Fu%2C+J">Jinlan Fu</a>, <a href="/search/cs?searchtype=author&query=Huangfu%2C+S">Shenzhen Huangfu</a>, <a href="/search/cs?searchtype=author&query=Fei%2C+H">Hao Fei</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+X">Xiaoyu Shen</a>, <a href="/search/cs?searchtype=author&query=Hooi%2C+B">Bryan Hooi</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Ng%2C+S">See-Kiong Ng</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2501.16629v1-abstract-short" style="display: inline;"> Multimodal Large Language Models (MLLMs) still struggle with hallucinations despite their impressive capabilities. Recent studies have attempted to mitigate this by applying Direct Preference Optimization (DPO) to multimodal scenarios using preference pairs from text-based responses. However, our analysis of representation distributions reveals that multimodal DPO struggles to align image and text… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.16629v1-abstract-full').style.display = 'inline'; document.getElementById('2501.16629v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.16629v1-abstract-full" style="display: none;"> Multimodal Large Language Models (MLLMs) still struggle with hallucinations despite their impressive capabilities. Recent studies have attempted to mitigate this by applying Direct Preference Optimization (DPO) to multimodal scenarios using preference pairs from text-based responses. However, our analysis of representation distributions reveals that multimodal DPO struggles to align image and text representations and to distinguish between hallucinated and non-hallucinated descriptions. To address these challenges, in this work, we propose a Cross-modal Hierarchical Direct Preference Optimization (CHiP) to address these limitations. We introduce a visual preference optimization module within the DPO framework, enabling MLLMs to learn from both textual and visual preferences simultaneously. Furthermore, we propose a hierarchical textual preference optimization module that allows the model to capture preferences at multiple granular levels, including response, segment, and token levels. We evaluate CHiP through both quantitative and qualitative analyses, with results across multiple benchmarks demonstrating its effectiveness in reducing hallucinations. On the Object HalBench dataset, CHiP outperforms DPO in hallucination reduction, achieving improvements of 52.7% and 55.5% relative points based on the base model Muffin and LLaVA models, respectively. We make all our datasets and code publicly available: https://github.com/LVUGAI/CHiP. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.16629v1-abstract-full').style.display = 'none'; document.getElementById('2501.16629v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by ICLR 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.15581">arXiv:2501.15581</a> <span> [<a href="https://arxiv.org/pdf/2501.15581">pdf</a>, <a href="https://arxiv.org/format/2501.15581">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"> Error Classification of Large Language Models on Math Word Problems: A Dynamically Adaptive Framework </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sun%2C+Y">Yuhong Sun</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+Z">Zhangyue Yin</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+H">Hui Zhao</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2501.15581v1-abstract-short" style="display: inline;"> Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. Math Word Problems (MWPs) serve as a crucial benchmark for evaluating LLMs' reasoning abilities. While most research primarily focuses on improving accuracy, it often neglects understanding and addressing the underlying patterns of errors. Current error classification methods rely on static and predefine… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.15581v1-abstract-full').style.display = 'inline'; document.getElementById('2501.15581v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.15581v1-abstract-full" style="display: none;"> Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. Math Word Problems (MWPs) serve as a crucial benchmark for evaluating LLMs' reasoning abilities. While most research primarily focuses on improving accuracy, it often neglects understanding and addressing the underlying patterns of errors. Current error classification methods rely on static and predefined categories, which limit their ability to capture the full spectrum of error patterns in mathematical reasoning. To enable systematic error analysis, we collect error samples from 15 different LLMs of varying sizes across four distinct MWP datasets using multiple sampling strategies. Based on this extensive collection, we introduce MWPES-300K, a comprehensive dataset containing 304,865 error samples that cover diverse error patterns and reasoning paths. To reduce human bias and enable fine-grained analysis of error patterns, we propose a novel framework for automated dynamic error classification in mathematical reasoning. Experimental results demonstrate that dataset characteristics significantly shape error patterns, which evolve from basic to complex manifestations as model capabilities increase. With deeper insights into error patterns, we propose error-aware prompting that incorporates common error patterns as explicit guidance, leading to significant improvements in mathematical reasoning performance. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.15581v1-abstract-full').style.display = 'none'; document.getElementById('2501.15581v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">22 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/2501.15383">arXiv:2501.15383</a> <span> [<a href="https://arxiv.org/pdf/2501.15383">pdf</a>, <a href="https://arxiv.org/format/2501.15383">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"> Qwen2.5-1M Technical Report </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Yang%2C+A">An Yang</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+B">Bowen Yu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+C">Chengyuan Li</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+D">Dayiheng Liu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+F">Fei Huang</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+H">Haoyan Huang</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+J">Jiandong Jiang</a>, <a href="/search/cs?searchtype=author&query=Tu%2C+J">Jianhong Tu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jianwei Zhang</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+J">Jingren Zhou</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+J">Junyang Lin</a>, <a href="/search/cs?searchtype=author&query=Dang%2C+K">Kai Dang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+K">Kexin Yang</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+L">Le Yu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+M">Mei Li</a>, <a href="/search/cs?searchtype=author&query=Sun%2C+M">Minmin Sun</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+Q">Qin Zhu</a>, <a href="/search/cs?searchtype=author&query=Men%2C+R">Rui Men</a>, <a href="/search/cs?searchtype=author&query=He%2C+T">Tao He</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+W">Weijia Xu</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+W">Wenbiao Yin</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+W">Wenyuan Yu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiafei Qiu</a>, <a href="/search/cs?searchtype=author&query=Ren%2C+X">Xingzhang Ren</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+X">Xinlong Yang</a> , et al. (3 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="2501.15383v1-abstract-short" style="display: inline;"> We introduce Qwen2.5-1M, a series of models that extend the context length to 1 million tokens. Compared to the previous 128K version, the Qwen2.5-1M series have significantly enhanced long-context capabilities through long-context pre-training and post-training. Key techniques such as long data synthesis, progressive pre-training, and multi-stage supervised fine-tuning are employed to effectively… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.15383v1-abstract-full').style.display = 'inline'; document.getElementById('2501.15383v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.15383v1-abstract-full" style="display: none;"> We introduce Qwen2.5-1M, a series of models that extend the context length to 1 million tokens. Compared to the previous 128K version, the Qwen2.5-1M series have significantly enhanced long-context capabilities through long-context pre-training and post-training. Key techniques such as long data synthesis, progressive pre-training, and multi-stage supervised fine-tuning are employed to effectively enhance long-context performance while reducing training costs. To promote the use of long-context models among a broader user base, we present and open-source our inference framework. This framework includes a length extrapolation method that can expand the model context lengths by at least four times, or even more, without additional training. To reduce inference costs, we implement a sparse attention method along with chunked prefill optimization for deployment scenarios and a sparsity refinement method to improve precision. Additionally, we detail our optimizations in the inference engine, including kernel optimization, pipeline parallelism, and scheduling optimization, which significantly enhance overall inference performance. By leveraging our inference framework, the Qwen2.5-1M models achieve a remarkable 3x to 7x prefill speedup in scenarios with 1 million tokens of context. This framework provides an efficient and powerful solution for developing applications that require long-context processing using open-source models. The Qwen2.5-1M series currently includes the open-source models Qwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1M, as well as the API-accessed model Qwen2.5-Turbo. Evaluations show that Qwen2.5-1M models have been greatly improved in long-context tasks without compromising performance in short-context scenarios. Specifically, the Qwen2.5-14B-Instruct-1M model significantly outperforms GPT-4o-mini in long-context tasks and supports contexts eight times longer. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.15383v1-abstract-full').style.display = 'none'; document.getElementById('2501.15383v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.13492">arXiv:2501.13492</a> <span> [<a href="https://arxiv.org/pdf/2501.13492">pdf</a>, <a href="https://arxiv.org/format/2501.13492">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"> Quantized Spike-driven Transformer </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xuerui Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+M">Malu Zhang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jieyuan Zhang</a>, <a href="/search/cs?searchtype=author&query=Wei%2C+W">Wenjie Wei</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+H">Honglin Cao</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+J">Junsheng Guo</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+R">Rui-Jie Zhu</a>, <a href="/search/cs?searchtype=author&query=Shan%2C+Y">Yimeng Shan</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+Y">Yang Yang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+H">Haizhou 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="2501.13492v3-abstract-short" style="display: inline;"> Spiking neural networks are emerging as a promising energy-efficient alternative to traditional artificial neural networks due to their spike-driven paradigm. However, recent research in the SNN domain has mainly focused on enhancing accuracy by designing large-scale Transformer structures, which typically rely on substantial computational resources, limiting their deployment on resource-constrain… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.13492v3-abstract-full').style.display = 'inline'; document.getElementById('2501.13492v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.13492v3-abstract-full" style="display: none;"> Spiking neural networks are emerging as a promising energy-efficient alternative to traditional artificial neural networks due to their spike-driven paradigm. However, recent research in the SNN domain has mainly focused on enhancing accuracy by designing large-scale Transformer structures, which typically rely on substantial computational resources, limiting their deployment on resource-constrained devices. To overcome this challenge, we propose a quantized spike-driven Transformer baseline (QSD-Transformer), which achieves reduced resource demands by utilizing a low bit-width parameter. Regrettably, the QSD-Transformer often suffers from severe performance degradation. In this paper, we first conduct empirical analysis and find that the bimodal distribution of quantized spike-driven self-attention (Q-SDSA) leads to spike information distortion (SID) during quantization, causing significant performance degradation. To mitigate this issue, we take inspiration from mutual information entropy and propose a bi-level optimization strategy to rectify the information distribution in Q-SDSA. Specifically, at the lower level, we introduce an information-enhanced LIF to rectify the information distribution in Q-SDSA. At the upper level, we propose a fine-grained distillation scheme for the QSD-Transformer to align the distribution in Q-SDSA with that in the counterpart ANN. By integrating the bi-level optimization strategy, the QSD-Transformer can attain enhanced energy efficiency without sacrificing its high-performance advantage.For instance, when compared to the prior SNN benchmark on ImageNet, the QSD-Transformer achieves 80.3% top-1 accuracy, accompanied by significant reductions of 6.0$\times$ and 8.1$\times$ in power consumption and model size, respectively. Code is available at https://github.com/bollossom/QSD-Transformer. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.13492v3-abstract-full').style.display = 'none'; document.getElementById('2501.13492v3-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 23 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by ICLR 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.12547">arXiv:2501.12547</a> <span> [<a href="https://arxiv.org/pdf/2501.12547">pdf</a>, <a href="https://arxiv.org/format/2501.12547">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"> Human-like conceptual representations emerge from language prediction </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xu%2C+N">Ningyu Xu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Q">Qi Zhang</a>, <a href="/search/cs?searchtype=author&query=Du%2C+C">Chao Du</a>, <a href="/search/cs?searchtype=author&query=Luo%2C+Q">Qiang Luo</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+M">Menghan Zhang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2501.12547v1-abstract-short" style="display: inline;"> Recent advances in large language models (LLMs) provide a new opportunity to address the long-standing question of how concepts are represented and organized in the mind, which is central to unravelling the nature of human cognition. Here, we reframed the classic reverse dictionary task to simulate human concept inference in context and investigated the emergence of human-like conceptual represent… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.12547v1-abstract-full').style.display = 'inline'; document.getElementById('2501.12547v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.12547v1-abstract-full" style="display: none;"> Recent advances in large language models (LLMs) provide a new opportunity to address the long-standing question of how concepts are represented and organized in the mind, which is central to unravelling the nature of human cognition. Here, we reframed the classic reverse dictionary task to simulate human concept inference in context and investigated the emergence of human-like conceptual representations within LLMs. We found that LLMs were able to infer concepts from definitional descriptions and construct representation spaces that converge towards a shared, context-independent structure. These representations effectively predicted human behavioural judgments and aligned well with neural activity patterns in the human brain, offering evidence for biological plausibility. These findings demonstrate that human-like conceptual representations and organization can naturally emerge from language prediction, even without real-world grounding. Our work supports the view that LLMs serve as valuable tools for understanding complex human cognition and paves the way for better alignment between artificial and human intelligence. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.12547v1-abstract-full').style.display = 'none'; document.getElementById('2501.12547v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 21 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2501.09026">arXiv:2501.09026</a> <span> [<a href="https://arxiv.org/pdf/2501.09026">pdf</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Social and Information Networks">cs.SI</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="Computers and Society">cs.CY</span> </div> </div> <p class="title is-5 mathjax"> Intelligent Anti-Money Laundering Solution Based upon Novel Community Detection in Massive Transaction Networks on Spark </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Li%2C+X">Xurui Li</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+X">Xiang Cao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xuetao Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+J">Jintao Zhao</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+J">Jianbin Zheng</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2501.09026v1-abstract-short" style="display: inline;"> Criminals are using every means available to launder the profits from their illegal activities into ostensibly legitimate assets. Meanwhile, most commercial anti-money laundering systems are still rule-based, which cannot adapt to the ever-changing tricks. Although some machine learning methods have been proposed, they are mainly focused on the perspective of abnormal behavior for single accounts.… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.09026v1-abstract-full').style.display = 'inline'; document.getElementById('2501.09026v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2501.09026v1-abstract-full" style="display: none;"> Criminals are using every means available to launder the profits from their illegal activities into ostensibly legitimate assets. Meanwhile, most commercial anti-money laundering systems are still rule-based, which cannot adapt to the ever-changing tricks. Although some machine learning methods have been proposed, they are mainly focused on the perspective of abnormal behavior for single accounts. Considering money laundering activities are often involved in gang criminals, these methods are still not intelligent enough to crack down on criminal gangs all-sidedly. In this paper, a systematic solution is presented to find suspicious money laundering gangs. A temporal-directed Louvain algorithm has been proposed to detect communities according to relevant anti-money laundering patterns. All processes are implemented and optimized on Spark platform. This solution can greatly improve the efficiency of anti-money laundering work for financial regulation agencies. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2501.09026v1-abstract-full').style.display = 'none'; document.getElementById('2501.09026v1-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> January 2025. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.20439">arXiv:2412.20439</a> <span> [<a href="https://arxiv.org/pdf/2412.20439">pdf</a>, <a href="https://arxiv.org/format/2412.20439">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"> Image Augmentation Agent for Weakly Supervised Semantic Segmentation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wu%2C+W">Wangyu Wu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xianglin Qiu</a>, <a href="/search/cs?searchtype=author&query=Song%2C+S">Siqi Song</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Z">Zhenhong Chen</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xiaowei Huang</a>, <a href="/search/cs?searchtype=author&query=Ma%2C+F">Fei Ma</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+J">Jimin Xiao</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.20439v1-abstract-short" style="display: inline;"> Weakly-supervised semantic segmentation (WSSS) has achieved remarkable progress using only image-level labels. However, most existing WSSS methods focus on designing new network structures and loss functions to generate more accurate dense labels, overlooking the limitations imposed by fixed datasets, which can constrain performance improvements. We argue that more diverse trainable images provide… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.20439v1-abstract-full').style.display = 'inline'; document.getElementById('2412.20439v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.20439v1-abstract-full" style="display: none;"> Weakly-supervised semantic segmentation (WSSS) has achieved remarkable progress using only image-level labels. However, most existing WSSS methods focus on designing new network structures and loss functions to generate more accurate dense labels, overlooking the limitations imposed by fixed datasets, which can constrain performance improvements. We argue that more diverse trainable images provides WSSS richer information and help model understand more comprehensive semantic pattern. Therefore in this paper, we introduce a novel approach called Image Augmentation Agent (IAA) which shows that it is possible to enhance WSSS from data generation perspective. IAA mainly design an augmentation agent that leverages large language models (LLMs) and diffusion models to automatically generate additional images for WSSS. In practice, to address the instability in prompt generation by LLMs, we develop a prompt self-refinement mechanism. It allow LLMs to re-evaluate the rationality of generated prompts to produce more coherent prompts. Additionally, we insert an online filter into diffusion generation process to dynamically ensure the quality and balance of generated images. Experimental results show that our method significantly surpasses state-of-the-art WSSS approaches on the PASCAL VOC 2012 and MS COCO 2014 datasets. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.20439v1-abstract-full').style.display = 'none'; document.getElementById('2412.20439v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.18919">arXiv:2412.18919</a> <span> [<a href="https://arxiv.org/pdf/2412.18919">pdf</a>, <a href="https://arxiv.org/format/2412.18919">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> <p class="title is-5 mathjax"> An Attentive Dual-Encoder Framework Leveraging Multimodal Visual and Semantic Information for Automatic OSAHS Diagnosis </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wei%2C+Y">Yingchen Wei</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xihe Qiu</a>, <a href="/search/cs?searchtype=author&query=Tan%2C+X">Xiaoyu Tan</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+J">Jingjing Huang</a>, <a href="/search/cs?searchtype=author&query=Chu%2C+W">Wei Chu</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+Y">Yinghui Xu</a>, <a href="/search/cs?searchtype=author&query=Qi%2C+Y">Yuan Qi</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.18919v1-abstract-short" style="display: inline;"> Obstructive sleep apnea-hypopnea syndrome (OSAHS) is a common sleep disorder caused by upper airway blockage, leading to oxygen deprivation and disrupted sleep. Traditional diagnosis using polysomnography (PSG) is expensive, time-consuming, and uncomfortable. Existing deep learning methods using facial image analysis lack accuracy due to poor facial feature capture and limited sample sizes. To add… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.18919v1-abstract-full').style.display = 'inline'; document.getElementById('2412.18919v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.18919v1-abstract-full" style="display: none;"> Obstructive sleep apnea-hypopnea syndrome (OSAHS) is a common sleep disorder caused by upper airway blockage, leading to oxygen deprivation and disrupted sleep. Traditional diagnosis using polysomnography (PSG) is expensive, time-consuming, and uncomfortable. Existing deep learning methods using facial image analysis lack accuracy due to poor facial feature capture and limited sample sizes. To address this, we propose a multimodal dual encoder model that integrates visual and language inputs for automated OSAHS diagnosis. The model balances data using randomOverSampler, extracts key facial features with attention grids, and converts physiological data into meaningful text. Cross-attention combines image and text data for better feature extraction, and ordered regression loss ensures stable learning. Our approach improves diagnostic efficiency and accuracy, achieving 91.3% top-1 accuracy in a four-class severity classification task, demonstrating state-of-the-art performance. Code will be released upon acceptance. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.18919v1-abstract-full').style.display = 'none'; document.getElementById('2412.18919v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">5 pages, 2 figures, Published as a conference paper at ICASSP 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.18194">arXiv:2412.18194</a> <span> [<a href="https://arxiv.org/pdf/2412.18194">pdf</a>, <a href="https://arxiv.org/format/2412.18194">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computation and Language">cs.CL</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Computer Vision and Pattern Recognition">cs.CV</span> </div> </div> <p class="title is-5 mathjax"> VLABench: A Large-Scale Benchmark for Language-Conditioned Robotics Manipulation with Long-Horizon Reasoning Tasks </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+S">Shiduo Zhang</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+Z">Zhe Xu</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+P">Peiju Liu</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+X">Xiaopeng Yu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yuan Li</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+Q">Qinghui Gao</a>, <a href="/search/cs?searchtype=author&query=Fei%2C+Z">Zhaoye Fei</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+Z">Zhangyue Yin</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.18194v1-abstract-short" style="display: inline;"> General-purposed embodied agents are designed to understand the users' natural instructions or intentions and act precisely to complete universal tasks. Recently, methods based on foundation models especially Vision-Language-Action models (VLAs) have shown a substantial potential to solve language-conditioned manipulation (LCM) tasks well. However, existing benchmarks do not adequately meet the ne… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.18194v1-abstract-full').style.display = 'inline'; document.getElementById('2412.18194v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.18194v1-abstract-full" style="display: none;"> General-purposed embodied agents are designed to understand the users' natural instructions or intentions and act precisely to complete universal tasks. Recently, methods based on foundation models especially Vision-Language-Action models (VLAs) have shown a substantial potential to solve language-conditioned manipulation (LCM) tasks well. However, existing benchmarks do not adequately meet the needs of VLAs and relative algorithms. To better define such general-purpose tasks in the context of LLMs and advance the research in VLAs, we present VLABench, an open-source benchmark for evaluating universal LCM task learning. VLABench provides 100 carefully designed categories of tasks, with strong randomization in each category of task and a total of 2000+ objects. VLABench stands out from previous benchmarks in four key aspects: 1) tasks requiring world knowledge and common sense transfer, 2) natural language instructions with implicit human intentions rather than templates, 3) long-horizon tasks demanding multi-step reasoning, and 4) evaluation of both action policies and language model capabilities. The benchmark assesses multiple competencies including understanding of mesh\&texture, spatial relationship, semantic instruction, physical laws, knowledge transfer and reasoning, etc. To support the downstream finetuning, we provide high-quality training data collected via an automated framework incorporating heuristic skills and prior information. The experimental results indicate that both the current state-of-the-art pretrained VLAs and the workflow based on VLMs face challenges in our tasks. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.18194v1-abstract-full').style.display = 'none'; document.getElementById('2412.18194v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.17603">arXiv:2412.17603</a> <span> [<a href="https://arxiv.org/pdf/2412.17603">pdf</a>, <a href="https://arxiv.org/format/2412.17603">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="Machine Learning">stat.ML</span> </div> </div> <p class="title is-5 mathjax"> EasyTime: Time Series Forecasting Made Easy </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiangfei Qiu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+X">Xiuwen Li</a>, <a href="/search/cs?searchtype=author&query=Pang%2C+R">Ruiyang Pang</a>, <a href="/search/cs?searchtype=author&query=Pan%2C+Z">Zhicheng Pan</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+X">Xingjian Wu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+L">Liu Yang</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jilin Hu</a>, <a href="/search/cs?searchtype=author&query=Shu%2C+Y">Yang Shu</a>, <a href="/search/cs?searchtype=author&query=Lu%2C+X">Xuesong Lu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+C">Chengcheng Yang</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+C">Chenjuan Guo</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+A">Aoying Zhou</a>, <a href="/search/cs?searchtype=author&query=Jensen%2C+C+S">Christian S. Jensen</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+B">Bin Yang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.17603v1-abstract-short" style="display: inline;"> Time series forecasting has important applications across diverse domains. EasyTime, the system we demonstrate, facilitates easy use of time-series forecasting methods by researchers and practitioners alike. First, EasyTime enables one-click evaluation, enabling researchers to evaluate new forecasting methods using the suite of diverse time series datasets collected in the preexisting time series… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.17603v1-abstract-full').style.display = 'inline'; document.getElementById('2412.17603v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.17603v1-abstract-full" style="display: none;"> Time series forecasting has important applications across diverse domains. EasyTime, the system we demonstrate, facilitates easy use of time-series forecasting methods by researchers and practitioners alike. First, EasyTime enables one-click evaluation, enabling researchers to evaluate new forecasting methods using the suite of diverse time series datasets collected in the preexisting time series forecasting benchmark (TFB). This is achieved by leveraging TFB's flexible and consistent evaluation pipeline. Second, when practitioners must perform forecasting on a new dataset, a nontrivial first step is often to find an appropriate forecasting method. EasyTime provides an Automated Ensemble module that combines the promising forecasting methods to yield superior forecasting accuracy compared to individual methods. Third, EasyTime offers a natural language Q&A module leveraging large language models. Given a question like "Which method is best for long term forecasting on time series with strong seasonality?", EasyTime converts the question into SQL queries on the database of results obtained by TFB and then returns an answer in natural language and charts. By demonstrating EasyTime, we intend to show how it is possible to simplify the use of time series forecasting and to offer better support for the development of new generations of time series forecasting methods. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.17603v1-abstract-full').style.display = 'none'; document.getElementById('2412.17603v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by ICDE2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.16985">arXiv:2412.16985</a> <span> [<a href="https://arxiv.org/pdf/2412.16985">pdf</a>, <a href="https://arxiv.org/format/2412.16985">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Distributed, Parallel, and Cluster Computing">cs.DC</span> </div> </div> <p class="title is-5 mathjax"> BladeDISC++: Memory Optimizations Based On Symbolic Shape </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Yuan%2C+X">Xiulong Yuan</a>, <a href="/search/cs?searchtype=author&query=Yan%2C+X">Xu Yan</a>, <a href="/search/cs?searchtype=author&query=Shen%2C+W">Wenting Shen</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiafei Qiu</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+A">Ang Wang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jie Zhang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Y">Yong Li</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+W">Wei Lin</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.16985v1-abstract-short" style="display: inline;"> Recent deep learning workloads exhibit dynamic characteristics, leading to the rising adoption of dynamic shape compilers. These compilers can generate efficient kernels for dynamic shape graphs characterized by a fixed graph topology and uncertain tensor shapes. However, memory optimization, although particularly crucial in this large model era, remains relatively underexplored for dynamic shape… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16985v1-abstract-full').style.display = 'inline'; document.getElementById('2412.16985v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.16985v1-abstract-full" style="display: none;"> Recent deep learning workloads exhibit dynamic characteristics, leading to the rising adoption of dynamic shape compilers. These compilers can generate efficient kernels for dynamic shape graphs characterized by a fixed graph topology and uncertain tensor shapes. However, memory optimization, although particularly crucial in this large model era, remains relatively underexplored for dynamic shape graphs. The fundamental challenge lies in the lack of precise tensor shapes which are essential in conventional methods such as operation scheduling(op scheduling) and rematerialization. To address this challenge, we propose op scheduling and rematerialization approaches based on symbolic shapes and developed BladeDISC++. Besides, since rematerialization decisions cannot be made solely at compile time when tensor shapes are unknown, BladeDISC++ employs a compilation-runtime combined strategy to optimally address shape dynamics. Evaluations indicate that BladeDISC++ effectively reduces memory usage for dynamic shape graphs, achieving memory consumption comparable to optimizations using precise shapes, thereby promoting the broader adoption of dynamic shape compilers. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16985v1-abstract-full').style.display = 'none'; document.getElementById('2412.16985v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Journal ref:</span> [1]"NeurIPS BladeDISC++: Memory Optimizations Based On Symbolic Shape" Neurips.cc, 2024. https://neurips.cc/virtual/2024/103601 (accessed Dec. 22, 2024) </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.16677">arXiv:2412.16677</a> <span> [<a href="https://arxiv.org/pdf/2412.16677">pdf</a>, <a href="https://arxiv.org/format/2412.16677">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"> VAST 1.0: A Unified Framework for Controllable and Consistent Video Generation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zhang%2C+C">Chi Zhang</a>, <a href="/search/cs?searchtype=author&query=Liang%2C+Y">Yuanzhi Liang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xi Qiu</a>, <a href="/search/cs?searchtype=author&query=Yi%2C+F">Fangqiu Yi</a>, <a href="/search/cs?searchtype=author&query=Li%2C+X">Xuelong 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="2412.16677v1-abstract-short" style="display: inline;"> Generating high-quality videos from textual descriptions poses challenges in maintaining temporal coherence and control over subject motion. We propose VAST (Video As Storyboard from Text), a two-stage framework to address these challenges and enable high-quality video generation. In the first stage, StoryForge transforms textual descriptions into detailed storyboards, capturing human poses and ob… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16677v1-abstract-full').style.display = 'inline'; document.getElementById('2412.16677v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.16677v1-abstract-full" style="display: none;"> Generating high-quality videos from textual descriptions poses challenges in maintaining temporal coherence and control over subject motion. We propose VAST (Video As Storyboard from Text), a two-stage framework to address these challenges and enable high-quality video generation. In the first stage, StoryForge transforms textual descriptions into detailed storyboards, capturing human poses and object layouts to represent the structural essence of the scene. In the second stage, VisionForge generates videos from these storyboards, producing high-quality videos with smooth motion, temporal consistency, and spatial coherence. By decoupling text understanding from video generation, VAST enables precise control over subject dynamics and scene composition. Experiments on the VBench benchmark demonstrate that VAST outperforms existing methods in both visual quality and semantic expression, setting a new standard for dynamic and coherent video generation. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.16677v1-abstract-full').style.display = 'none'; document.getElementById('2412.16677v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 21 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.14135">arXiv:2412.14135</a> <span> [<a href="https://arxiv.org/pdf/2412.14135">pdf</a>, <a href="https://arxiv.org/format/2412.14135">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Artificial Intelligence">cs.AI</span> <span class="tag is-small is-grey tooltip is-tooltip-top" data-tooltip="Machine Learning">cs.LG</span> </div> </div> <p class="title is-5 mathjax"> Scaling of Search and Learning: A Roadmap to Reproduce o1 from Reinforcement Learning Perspective </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Zeng%2C+Z">Zhiyuan Zeng</a>, <a href="/search/cs?searchtype=author&query=Cheng%2C+Q">Qinyuan Cheng</a>, <a href="/search/cs?searchtype=author&query=Yin%2C+Z">Zhangyue Yin</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+B">Bo Wang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+S">Shimin Li</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Y">Yunhua Zhou</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.14135v1-abstract-short" style="display: inline;"> OpenAI o1 represents a significant milestone in Artificial Inteiligence, which achieves expert-level performances on many challanging tasks that require strong reasoning ability.OpenAI has claimed that the main techinique behinds o1 is the reinforcement learining. Recent works use alternative approaches like knowledge distillation to imitate o1's reasoning style, but their effectiveness is limited… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.14135v1-abstract-full').style.display = 'inline'; document.getElementById('2412.14135v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.14135v1-abstract-full" style="display: none;"> OpenAI o1 represents a significant milestone in Artificial Inteiligence, which achieves expert-level performances on many challanging tasks that require strong reasoning ability.OpenAI has claimed that the main techinique behinds o1 is the reinforcement learining. Recent works use alternative approaches like knowledge distillation to imitate o1's reasoning style, but their effectiveness is limited by the capability ceiling of the teacher model. Therefore, this paper analyzes the roadmap to achieving o1 from the perspective of reinforcement learning, focusing on four key components: policy initialization, reward design, search, and learning. Policy initialization enables models to develop human-like reasoning behaviors, equipping them with the ability to effectively explore solution spaces for complex problems. Reward design provides dense and effective signals via reward shaping or reward modeling, which is the guidance for both search and learning. Search plays a crucial role in generating high-quality solutions during both training and testing phases, which can produce better solutions with more computation. Learning utilizes the data generated by search for improving policy, which can achieve the better performance with more parameters and more searched data. Existing open-source projects that attempt to reproduce o1 can be seem as a part or a variant of our roadmap. Collectively, these components underscore how learning and search drive o1's advancement, making meaningful contributions to the development of LLM. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.14135v1-abstract-full').style.display = 'none'; document.getElementById('2412.14135v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.13823">arXiv:2412.13823</a> <span> [<a href="https://arxiv.org/pdf/2412.13823">pdf</a>, <a href="https://arxiv.org/format/2412.13823">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"> Prompt Categories Cluster for Weakly Supervised Semantic Segmentation </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wu%2C+W">Wangyu Wu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xianglin Qiu</a>, <a href="/search/cs?searchtype=author&query=Song%2C+S">Siqi Song</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xiaowei Huang</a>, <a href="/search/cs?searchtype=author&query=Ma%2C+F">Fei Ma</a>, <a href="/search/cs?searchtype=author&query=Xiao%2C+J">Jimin Xiao</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.13823v1-abstract-short" style="display: inline;"> Weakly Supervised Semantic Segmentation (WSSS), which leverages image-level labels, has garnered significant attention due to its cost-effectiveness. The previous methods mainly strengthen the inter-class differences to avoid class semantic ambiguity which may lead to erroneous activation. However, they overlook the positive function of some shared information between similar classes. Categories w… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.13823v1-abstract-full').style.display = 'inline'; document.getElementById('2412.13823v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.13823v1-abstract-full" style="display: none;"> Weakly Supervised Semantic Segmentation (WSSS), which leverages image-level labels, has garnered significant attention due to its cost-effectiveness. The previous methods mainly strengthen the inter-class differences to avoid class semantic ambiguity which may lead to erroneous activation. However, they overlook the positive function of some shared information between similar classes. Categories within the same cluster share some similar features. Allowing the model to recognize these features can further relieve the semantic ambiguity between these classes. To effectively identify and utilize this shared information, in this paper, we introduce a novel WSSS framework called Prompt Categories Clustering (PCC). Specifically, we explore the ability of Large Language Models (LLMs) to derive category clusters through prompts. These clusters effectively represent the intrinsic relationships between categories. By integrating this relational information into the training network, our model is able to better learn the hidden connections between categories. Experimental results demonstrate the effectiveness of our approach, showing its ability to enhance performance on the PASCAL VOC 2012 dataset and surpass existing state-of-the-art methods in WSSS. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.13823v1-abstract-full').style.display = 'none'; document.getElementById('2412.13823v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.12737">arXiv:2412.12737</a> <span> [<a href="https://arxiv.org/pdf/2412.12737">pdf</a>, <a href="https://arxiv.org/format/2412.12737">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"> PolSAM: Polarimetric Scattering Mechanism Informed Segment Anything Model </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yuqing Wang</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+Z">Zhongling Huang</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+S">Shuxin Yang</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+H">Hao Tang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaolan Qiu</a>, <a href="/search/cs?searchtype=author&query=Han%2C+J">Junwei Han</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+D">Dingwen Zhang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.12737v1-abstract-short" style="display: inline;"> PolSAR data presents unique challenges due to its rich and complex characteristics. Existing data representations, such as complex-valued data, polarimetric features, and amplitude images, are widely used. However, these formats often face issues related to usability, interpretability, and data integrity. Most feature extraction networks for PolSAR are small, limiting their ability to capture feat… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.12737v1-abstract-full').style.display = 'inline'; document.getElementById('2412.12737v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.12737v1-abstract-full" style="display: none;"> PolSAR data presents unique challenges due to its rich and complex characteristics. Existing data representations, such as complex-valued data, polarimetric features, and amplitude images, are widely used. However, these formats often face issues related to usability, interpretability, and data integrity. Most feature extraction networks for PolSAR are small, limiting their ability to capture features effectively. To address these issues, We propose the Polarimetric Scattering Mechanism-Informed SAM (PolSAM), an enhanced Segment Anything Model (SAM) that integrates domain-specific scattering characteristics and a novel prompt generation strategy. PolSAM introduces Microwave Vision Data (MVD), a lightweight and interpretable data representation derived from polarimetric decomposition and semantic correlations. We propose two key components: the Feature-Level Fusion Prompt (FFP), which fuses visual tokens from pseudo-colored SAR images and MVD to address modality incompatibility in the frozen SAM encoder, and the Semantic-Level Fusion Prompt (SFP), which refines sparse and dense segmentation prompts using semantic information. Experimental results on the PhySAR-Seg datasets demonstrate that PolSAM significantly outperforms existing SAM-based and multimodal fusion models, improving segmentation accuracy, reducing data storage, and accelerating inference time. The source code and datasets will be made publicly available at \url{https://github.com/XAI4SAR/PolSAM}. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.12737v1-abstract-full').style.display = 'none'; document.getElementById('2412.12737v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 17 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <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 manuscript is 15 pages long, includes 14 figures and 5 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/2412.10859">arXiv:2412.10859</a> <span> [<a href="https://arxiv.org/pdf/2412.10859">pdf</a>, <a href="https://arxiv.org/format/2412.10859">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="Machine Learning">stat.ML</span> </div> </div> <p class="title is-5 mathjax"> DUET: Dual Clustering Enhanced Multivariate Time Series Forecasting </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiangfei Qiu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+X">Xingjian Wu</a>, <a href="/search/cs?searchtype=author&query=Lin%2C+Y">Yan Lin</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+C">Chenjuan Guo</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jilin Hu</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+B">Bin Yang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.10859v3-abstract-short" style="display: inline;"> Multivariate time series forecasting is crucial for various applications, such as financial investment, energy management, weather forecasting, and traffic optimization. However, accurate forecasting is challenging due to two main factors. First, real-world time series often show heterogeneous temporal patterns caused by distribution shifts over time. Second, correlations among channels are comple… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.10859v3-abstract-full').style.display = 'inline'; document.getElementById('2412.10859v3-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.10859v3-abstract-full" style="display: none;"> Multivariate time series forecasting is crucial for various applications, such as financial investment, energy management, weather forecasting, and traffic optimization. However, accurate forecasting is challenging due to two main factors. First, real-world time series often show heterogeneous temporal patterns caused by distribution shifts over time. Second, correlations among channels are complex and intertwined, making it hard to model the interactions among channels precisely and flexibly. In this study, we address these challenges by proposing a general framework called DUET, which introduces dual clustering on the temporal and channel dimensions to enhance multivariate time series forecasting. First, we design a Temporal Clustering Module (TCM) that clusters time series into fine-grained distributions to handle heterogeneous temporal patterns. For different distribution clusters, we design various pattern extractors to capture their intrinsic temporal patterns, thus modeling the heterogeneity. Second, we introduce a novel Channel-Soft-Clustering strategy and design a Channel Clustering Module (CCM), which captures the relationships among channels in the frequency domain through metric learning and applies sparsification to mitigate the adverse effects of noisy channels. Finally, DUET combines TCM and CCM to incorporate both the temporal and channel dimensions. Extensive experiments on 25 real-world datasets from 10 application domains, demonstrate the state-of-the-art performance of DUET. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.10859v3-abstract-full').style.display = 'none'; document.getElementById('2412.10859v3-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 January, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 14 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by KDD 2025 research track</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.10087">arXiv:2412.10087</a> <span> [<a href="https://arxiv.org/pdf/2412.10087">pdf</a>, <a href="https://arxiv.org/format/2412.10087">other</a>] </span> </p> <div class="tags is-inline-block"> <span class="tag is-small is-link tooltip is-tooltip-top" data-tooltip="Robotics">cs.RO</span> </div> </div> <p class="title is-5 mathjax"> Consensus-Based Dynamic Task Allocation for Multi-Robot System Considering Payloads Consumption </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xuekai Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhu%2C+P">Pengming Zhu</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+Y">Yiming Hu</a>, <a href="/search/cs?searchtype=author&query=Zeng%2C+Z">Zhiwen Zeng</a>, <a href="/search/cs?searchtype=author&query=Lu%2C+H">Huimin Lu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.10087v1-abstract-short" style="display: inline;"> This paper presents a consensus-based payload algorithm (CBPA) to deal with the condition of robots' capability decrease for multi-robot task allocation. During the execution of complex tasks, robots' capabilities could decrease with the consumption of payloads, which causes a problem that the robot coalition would not meet the tasks' requirements in real time. The proposed CBPA is an enhanced ver… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.10087v1-abstract-full').style.display = 'inline'; document.getElementById('2412.10087v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.10087v1-abstract-full" style="display: none;"> This paper presents a consensus-based payload algorithm (CBPA) to deal with the condition of robots' capability decrease for multi-robot task allocation. During the execution of complex tasks, robots' capabilities could decrease with the consumption of payloads, which causes a problem that the robot coalition would not meet the tasks' requirements in real time. The proposed CBPA is an enhanced version of the consensus-based bundle algorithm (CBBA) and comprises two primary core phases: the payload bundle construction and consensus phases. In the payload bundle construction phase, CBPA introduces a payload assignment matrix to track the payloads carried by the robots and the demands of multi-robot tasks in real time. Then, robots share their respective payload assignment matrix in the consensus phase. These two phases are iterated to dynamically adjust the number of robots performing multi-robot tasks and the number of tasks each robot performs and obtain conflict-free results to ensure that the robot coalition meets the demand and completes all tasks as quickly as possible. Physical experiment shows that CBPA is appropriate in complex and dynamic scenarios where robots need to collaborate and task requirements are tightly coupled to the robots' payloads. Numerical experiments show that CBPA has higher total task gains than CBBA. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.10087v1-abstract-full').style.display = 'none'; document.getElementById('2412.10087v1-abstract-short').style.display = 'inline';">△ Less</a> </span> </p> <p class="is-size-7"><span class="has-text-black-bis has-text-weight-semibold">Submitted</span> 13 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.07360">arXiv:2412.07360</a> <span> [<a href="https://arxiv.org/pdf/2412.07360">pdf</a>, <a href="https://arxiv.org/format/2412.07360">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"> Efficient 3D Recognition with Event-driven Spike Sparse Convolution </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xuerui Qiu</a>, <a href="/search/cs?searchtype=author&query=Yao%2C+M">Man Yao</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+J">Jieyuan Zhang</a>, <a href="/search/cs?searchtype=author&query=Chou%2C+Y">Yuhong Chou</a>, <a href="/search/cs?searchtype=author&query=Qiao%2C+N">Ning Qiao</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+S">Shibo Zhou</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+B">Bo Xu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+G">Guoqi 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="2412.07360v2-abstract-short" style="display: inline;"> Spiking Neural Networks (SNNs) provide an energy-efficient way to extract 3D spatio-temporal features. Point clouds are sparse 3D spatial data, which suggests that SNNs should be well-suited for processing them. However, when applying SNNs to point clouds, they often exhibit limited performance and fewer application scenarios. We attribute this to inappropriate preprocessing and feature extraction… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.07360v2-abstract-full').style.display = 'inline'; document.getElementById('2412.07360v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.07360v2-abstract-full" style="display: none;"> Spiking Neural Networks (SNNs) provide an energy-efficient way to extract 3D spatio-temporal features. Point clouds are sparse 3D spatial data, which suggests that SNNs should be well-suited for processing them. However, when applying SNNs to point clouds, they often exhibit limited performance and fewer application scenarios. We attribute this to inappropriate preprocessing and feature extraction methods. To address this issue, we first introduce the Spike Voxel Coding (SVC) scheme, which encodes the 3D point clouds into a sparse spike train space, reducing the storage requirements and saving time on point cloud preprocessing. Then, we propose a Spike Sparse Convolution (SSC) model for efficiently extracting 3D sparse point cloud features. Combining SVC and SSC, we design an efficient 3D SNN backbone (E-3DSNN), which is friendly with neuromorphic hardware. For instance, SSC can be implemented on neuromorphic chips with only minor modifications to the addressing function of vanilla spike convolution. Experiments on ModelNet40, KITTI, and Semantic KITTI datasets demonstrate that E-3DSNN achieves state-of-the-art (SOTA) results with remarkable efficiency. Notably, our E-3DSNN (1.87M) obtained 91.7\% top-1 accuracy on ModelNet40, surpassing the current best SNN baselines (14.3M) by 3.0\%. To our best knowledge, it is the first direct training 3D SNN backbone that can simultaneously handle various 3D computer vision tasks (e.g., classification, detection, and segmentation) with an event-driven nature. Code is available: https://github.com/bollossom/E-3DSNN/. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.07360v2-abstract-full').style.display = 'none'; document.getElementById('2412.07360v2-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 February, 2025; <span class="has-text-black-bis has-text-weight-semibold">v1</span> submitted 10 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by AAAI 2025</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.06444">arXiv:2412.06444</a> <span> [<a href="https://arxiv.org/pdf/2412.06444">pdf</a>, <a href="https://arxiv.org/format/2412.06444">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"> The Complexity of Tullock Contests </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=He%2C+Y">Yu He</a>, <a href="/search/cs?searchtype=author&query=Yao%2C+F">Fan Yao</a>, <a href="/search/cs?searchtype=author&query=Yu%2C+Y">Yang Yu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaoyun Qiu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+M">Minming Li</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+H">Haifeng Xu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.06444v1-abstract-short" style="display: inline;"> This paper investigates the algorithmic complexity of computing the pure Nash Equilibrium (PNE) in Tullock contests. A key aspect of this analysis lies in the elasticity parameter $r_i$, which dictates whether a contestant $i$'s cost function is convex, concave, or neither. Our primary contribution is the identification of how the domains of $r_i$ govern the computational complexity of solving Tul… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.06444v1-abstract-full').style.display = 'inline'; document.getElementById('2412.06444v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.06444v1-abstract-full" style="display: none;"> This paper investigates the algorithmic complexity of computing the pure Nash Equilibrium (PNE) in Tullock contests. A key aspect of this analysis lies in the elasticity parameter $r_i$, which dictates whether a contestant $i$'s cost function is convex, concave, or neither. Our primary contribution is the identification of how the domains of $r_i$ govern the computational complexity of solving Tullock contests. Specifically, we establish the following results: - Efficient Algorithms for $r_i \notin (1,2]$: When no contestant's elasticity parameter $r_i$ lies within the interval $(1,2]$, we show that an efficient algorithm can be designed to compute the pure Nash Equilibrium. - Hardness Results for $r_i \in (1,2]$: When many $r_i$ values fall within the interval $(1,2]$, we prove that determining the existence of a pure Nash Equilibrium cannot be achieved in polynomial time, assuming the Exponential Time Hypothesis (ETH). - Approximation Algorithms for $r_i \in (1,2]$: In cases where many $r_i$ values fall within the interval $(1,2]$, we propose a Fully Polynomial-Time Approximation Scheme (FPTAS) to compute an $蔚$-approximate PNE, provided an exact PNE exists. All our algorithms are implemented efficiently to handle large-scale instances. Computational experiments validate their effectiveness, even under challenging scenarios with complex elasticity distributions. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.06444v1-abstract-full').style.display = 'none'; document.getElementById('2412.06444v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.03565">arXiv:2412.03565</a> <span> [<a href="https://arxiv.org/pdf/2412.03565">pdf</a>, <a href="https://arxiv.org/format/2412.03565">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"> Inst-IT: Boosting Multimodal Instance Understanding via Explicit Visual Prompt Instruction Tuning </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Peng%2C+W">Wujian Peng</a>, <a href="/search/cs?searchtype=author&query=Meng%2C+L">Lingchen Meng</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Y">Yitong Chen</a>, <a href="/search/cs?searchtype=author&query=Xie%2C+Y">Yiweng Xie</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Y">Yang Liu</a>, <a href="/search/cs?searchtype=author&query=Gui%2C+T">Tao Gui</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+H">Hang Xu</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.03565v1-abstract-short" style="display: inline;"> Large Multimodal Models (LMMs) have made significant breakthroughs with the advancement of instruction tuning. However, while existing models can understand images and videos at a holistic level, they still struggle with instance-level understanding that requires a more nuanced comprehension and alignment. Instance-level understanding is crucial, as it focuses on the specific elements that we are… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.03565v1-abstract-full').style.display = 'inline'; document.getElementById('2412.03565v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.03565v1-abstract-full" style="display: none;"> Large Multimodal Models (LMMs) have made significant breakthroughs with the advancement of instruction tuning. However, while existing models can understand images and videos at a holistic level, they still struggle with instance-level understanding that requires a more nuanced comprehension and alignment. Instance-level understanding is crucial, as it focuses on the specific elements that we are most interested in. Excitingly, existing works find that the state-of-the-art LMMs exhibit strong instance understanding capabilities when provided with explicit visual cues. Motivated by this, we introduce an automated annotation pipeline assisted by GPT-4o to extract instance-level information from images and videos through explicit visual prompting for instance guidance. Building upon this pipeline, we proposed Inst-IT, a solution to enhance LMMs in Instance understanding via explicit visual prompt Instruction Tuning. Inst-IT consists of a benchmark to diagnose multimodal instance-level understanding, a large-scale instruction-tuning dataset, and a continuous instruction-tuning training paradigm to effectively enhance spatial-temporal instance understanding capabilities of existing LMMs. Experimental results show that, with the boost of Inst-IT, our models not only achieve outstanding performance on Inst-IT Bench but also demonstrate significant improvements across various generic image and video understanding benchmarks. This highlights that our dataset not only boosts instance-level understanding but also strengthens the overall capabilities of generic image and video comprehension. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.03565v1-abstract-full').style.display = 'none'; document.getElementById('2412.03565v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Project page at https://inst-it.github.io</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2412.03105">arXiv:2412.03105</a> <span> [<a href="https://arxiv.org/pdf/2412.03105">pdf</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> <p class="title is-5 mathjax"> Few-Shot Learning with Adaptive Weight Masking in Conditional GANs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jiacheng Hu</a>, <a href="/search/cs?searchtype=author&query=Qi%2C+Z">Zhen Qi</a>, <a href="/search/cs?searchtype=author&query=Wei%2C+J">Jianjun Wei</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+J">Jiajing Chen</a>, <a href="/search/cs?searchtype=author&query=Bao%2C+R">Runyuan Bao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinyu Qiu</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2412.03105v1-abstract-short" style="display: inline;"> Deep learning has revolutionized various fields, yet its efficacy is hindered by overfitting and the requirement of extensive annotated data, particularly in few-shot learning scenarios where limited samples are available. This paper introduces a novel approach to few-shot learning by employing a Residual Weight Masking Conditional Generative Adversarial Network (RWM-CGAN) for data augmentation. T… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.03105v1-abstract-full').style.display = 'inline'; document.getElementById('2412.03105v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2412.03105v1-abstract-full" style="display: none;"> Deep learning has revolutionized various fields, yet its efficacy is hindered by overfitting and the requirement of extensive annotated data, particularly in few-shot learning scenarios where limited samples are available. This paper introduces a novel approach to few-shot learning by employing a Residual Weight Masking Conditional Generative Adversarial Network (RWM-CGAN) for data augmentation. The proposed model integrates residual units within the generator to enhance network depth and sample quality, coupled with a weight mask regularization technique in the discriminator to improve feature learning from small-sample categories. This method addresses the core issues of robustness and generalization in few-shot learning by providing a controlled and clear augmentation of the sample space. Extensive experiments demonstrate that RWM-CGAN not only expands the sample space effectively but also enriches the diversity and quality of generated samples, leading to significant improvements in detection and classification accuracy on public datasets. The paper contributes to the advancement of few-shot learning by offering a practical solution to the challenges posed by data scarcity and the need for rapid generalization to new tasks or categories. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2412.03105v1-abstract-full').style.display = 'none'; document.getElementById('2412.03105v1-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 December, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> December 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.19466">arXiv:2411.19466</a> <span> [<a href="https://arxiv.org/pdf/2411.19466">pdf</a>, <a href="https://arxiv.org/format/2411.19466">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> <p class="title is-5 mathjax"> ForgerySleuth: Empowering Multimodal Large Language Models for Image Manipulation Detection </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sun%2C+Z">Zhihao Sun</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+H">Haoran Jiang</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+H">Haoran Chen</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+Y">Yixin Cao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.19466v1-abstract-short" style="display: inline;"> Multimodal large language models have unlocked new possibilities for various multimodal tasks. However, their potential in image manipulation detection remains unexplored. When directly applied to the IMD task, M-LLMs often produce reasoning texts that suffer from hallucinations and overthinking. To address this, in this work, we propose ForgerySleuth, which leverages M-LLMs to perform comprehensi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.19466v1-abstract-full').style.display = 'inline'; document.getElementById('2411.19466v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.19466v1-abstract-full" style="display: none;"> Multimodal large language models have unlocked new possibilities for various multimodal tasks. However, their potential in image manipulation detection remains unexplored. When directly applied to the IMD task, M-LLMs often produce reasoning texts that suffer from hallucinations and overthinking. To address this, in this work, we propose ForgerySleuth, which leverages M-LLMs to perform comprehensive clue fusion and generate segmentation outputs indicating specific regions that are tampered with. Moreover, we construct the ForgeryAnalysis dataset through the Chain-of-Clues prompt, which includes analysis and reasoning text to upgrade the image manipulation detection task. A data engine is also introduced to build a larger-scale dataset for the pre-training phase. Our extensive experiments demonstrate the effectiveness of ForgeryAnalysis and show that ForgerySleuth significantly outperforms existing methods in generalization, robustness, and explainability. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.19466v1-abstract-full').style.display = 'none'; document.getElementById('2411.19466v1-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, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.16579">arXiv:2411.16579</a> <span> [<a href="https://arxiv.org/pdf/2411.16579">pdf</a>, <a href="https://arxiv.org/format/2411.16579">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"> Enhancing LLM Reasoning via Critique Models with Test-Time and Training-Time Supervision </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Xi%2C+Z">Zhiheng Xi</a>, <a href="/search/cs?searchtype=author&query=Yang%2C+D">Dingwen Yang</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+J">Jixuan Huang</a>, <a href="/search/cs?searchtype=author&query=Tang%2C+J">Jiafu Tang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+G">Guanyu Li</a>, <a href="/search/cs?searchtype=author&query=Ding%2C+Y">Yiwen Ding</a>, <a href="/search/cs?searchtype=author&query=He%2C+W">Wei He</a>, <a href="/search/cs?searchtype=author&query=Hong%2C+B">Boyang Hong</a>, <a href="/search/cs?searchtype=author&query=Do%2C+S">Shihan Do</a>, <a href="/search/cs?searchtype=author&query=Zhan%2C+W">Wenyu Zhan</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+X">Xiao Wang</a>, <a href="/search/cs?searchtype=author&query=Zheng%2C+R">Rui Zheng</a>, <a href="/search/cs?searchtype=author&query=Ji%2C+T">Tao Ji</a>, <a href="/search/cs?searchtype=author&query=Shi%2C+X">Xiaowei Shi</a>, <a href="/search/cs?searchtype=author&query=Zhai%2C+Y">Yitao Zhai</a>, <a href="/search/cs?searchtype=author&query=Weng%2C+R">Rongxiang Weng</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+J">Jingang Wang</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+X">Xunliang Cai</a>, <a href="/search/cs?searchtype=author&query=Gui%2C+T">Tao Gui</a>, <a href="/search/cs?searchtype=author&query=Wu%2C+Z">Zuxuan Wu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Q">Qi Zhang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a>, <a href="/search/cs?searchtype=author&query=Jiang%2C+Y">Yu-Gang Jiang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.16579v1-abstract-short" style="display: inline;"> Training large language models (LLMs) to spend more time thinking and reflection before responding is crucial for effectively solving complex reasoning tasks in fields such as science, coding, and mathematics. However, the effectiveness of mechanisms like self-reflection and self-correction depends on the model's capacity to accurately assess its own performance, which can be limited by factors su… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.16579v1-abstract-full').style.display = 'inline'; document.getElementById('2411.16579v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.16579v1-abstract-full" style="display: none;"> Training large language models (LLMs) to spend more time thinking and reflection before responding is crucial for effectively solving complex reasoning tasks in fields such as science, coding, and mathematics. However, the effectiveness of mechanisms like self-reflection and self-correction depends on the model's capacity to accurately assess its own performance, which can be limited by factors such as initial accuracy, question difficulty, and the lack of external feedback. In this paper, we delve into a two-player paradigm that separates the roles of reasoning and critique models, where the critique model provides step-level feedback to supervise the reasoning (actor) model during both test-time and train-time. We first propose AutoMathCritique, an automated and scalable framework for collecting critique data, resulting in a dataset of $76,321$ responses paired with step-level feedback. Fine-tuning language models with this dataset enables them to generate natural language feedback for mathematical reasoning. We demonstrate that the critique models consistently improve the actor's performance on difficult queries at test-time, especially when scaling up inference-time computation. Motivated by these findings, we introduce the critique-based supervision to the actor's self-training process, and propose a critique-in-the-loop self-improvement method. Experiments show that the method improves the actor's exploration efficiency and solution diversity, especially on challenging queries, leading to a stronger reasoning model. Lastly, we take the preliminary step to explore training self-talk reasoning models via critique supervision and showcase its potential. Our code and datasets are at \href{https://mathcritique.github.io/}{https://mathcritique.github.io/}. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.16579v1-abstract-full').style.display = 'none'; document.getElementById('2411.16579v1-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Preprint</span> </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.16061">arXiv:2411.16061</a> <span> [<a href="https://arxiv.org/pdf/2411.16061">pdf</a>, <a href="https://arxiv.org/format/2411.16061">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 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/10.1109/TPAMI.2025.3530246">10.1109/TPAMI.2025.3530246 <i class="fa fa-external-link" aria-hidden="true"></i></a></span> </div> </div> </div> <p class="title is-5 mathjax"> Scaling Spike-driven Transformer with Efficient Spike Firing Approximation Training </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Yao%2C+M">Man Yao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xuerui Qiu</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+T">Tianxiang Hu</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+J">Jiakui Hu</a>, <a href="/search/cs?searchtype=author&query=Chou%2C+Y">Yuhong Chou</a>, <a href="/search/cs?searchtype=author&query=Tian%2C+K">Keyu Tian</a>, <a href="/search/cs?searchtype=author&query=Liao%2C+J">Jianxing Liao</a>, <a href="/search/cs?searchtype=author&query=Leng%2C+L">Luziwei Leng</a>, <a href="/search/cs?searchtype=author&query=Xu%2C+B">Bo Xu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+G">Guoqi 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="2411.16061v1-abstract-short" style="display: inline;"> The ambition of brain-inspired Spiking Neural Networks (SNNs) is to become a low-power alternative to traditional Artificial Neural Networks (ANNs). This work addresses two major challenges in realizing this vision: the performance gap between SNNs and ANNs, and the high training costs of SNNs. We identify intrinsic flaws in spiking neurons caused by binary firing mechanisms and propose a Spike Fi… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.16061v1-abstract-full').style.display = 'inline'; document.getElementById('2411.16061v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.16061v1-abstract-full" style="display: none;"> The ambition of brain-inspired Spiking Neural Networks (SNNs) is to become a low-power alternative to traditional Artificial Neural Networks (ANNs). This work addresses two major challenges in realizing this vision: the performance gap between SNNs and ANNs, and the high training costs of SNNs. We identify intrinsic flaws in spiking neurons caused by binary firing mechanisms and propose a Spike Firing Approximation (SFA) method using integer training and spike-driven inference. This optimizes the spike firing pattern of spiking neurons, enhancing efficient training, reducing power consumption, improving performance, enabling easier scaling, and better utilizing neuromorphic chips. We also develop an efficient spike-driven Transformer architecture and a spike-masked autoencoder to prevent performance degradation during SNN scaling. On ImageNet-1k, we achieve state-of-the-art top-1 accuracy of 78.5\%, 79.8\%, 84.0\%, and 86.2\% with models containing 10M, 19M, 83M, and 173M parameters, respectively. For instance, the 10M model outperforms the best existing SNN by 7.2\% on ImageNet, with training time acceleration and inference energy efficiency improved by 4.5$\times$ and 3.9$\times$, respectively. We validate the effectiveness and efficiency of the proposed method across various tasks, including object detection, semantic segmentation, and neuromorphic vision tasks. This work enables SNNs to match ANN performance while maintaining the low-power advantage, marking a significant step towards SNNs as a general visual backbone. Code is available at https://github.com/BICLab/Spike-Driven-Transformer-V3. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.16061v1-abstract-full').style.display = 'none'; document.getElementById('2411.16061v1-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, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.10508">arXiv:2411.10508</a> <span> [<a href="https://arxiv.org/pdf/2411.10508">pdf</a>, <a href="https://arxiv.org/format/2411.10508">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"> DR-BFR: Degradation Representation with Diffusion Models for Blind Face Restoration </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinmin Qiu</a>, <a href="/search/cs?searchtype=author&query=Li%2C+B">Bonan Li</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Z">Zicheng Zhang</a>, <a href="/search/cs?searchtype=author&query=Han%2C+C">Congying Han</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+T">Tiande Guo</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.10508v1-abstract-short" style="display: inline;"> Blind face restoration (BFR) is fundamentally challenged by the extensive range of degradation types and degrees that impact model generalization. Recent advancements in diffusion models have made considerable progress in this field. Nevertheless, a critical limitation is their lack of awareness of specific degradation, leading to potential issues such as unnatural details and inaccurate textures.… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.10508v1-abstract-full').style.display = 'inline'; document.getElementById('2411.10508v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.10508v1-abstract-full" style="display: none;"> Blind face restoration (BFR) is fundamentally challenged by the extensive range of degradation types and degrees that impact model generalization. Recent advancements in diffusion models have made considerable progress in this field. Nevertheless, a critical limitation is their lack of awareness of specific degradation, leading to potential issues such as unnatural details and inaccurate textures. In this paper, we equip diffusion models with the capability to decouple various degradation as a degradation prompt from low-quality (LQ) face images via unsupervised contrastive learning with reconstruction loss, and demonstrate that this capability significantly improves performance, particularly in terms of the naturalness of the restored images. Our novel restoration scheme, named DR-BFR, guides the denoising of Latent Diffusion Models (LDM) by incorporating Degradation Representation (DR) and content features from LQ images. DR-BFR comprises two modules: 1) Degradation Representation Module (DRM): This module extracts degradation representation with content-irrelevant features from LQ faces and estimates a reasonable distribution in the degradation space through contrastive learning and a specially designed LQ reconstruction. 2) Latent Diffusion Restoration Module (LDRM): This module perceives both degradation features and content features in the latent space, enabling the restoration of high-quality images from LQ inputs. Our experiments demonstrate that the proposed DR-BFR significantly outperforms state-of-the-art methods quantitatively and qualitatively across various datasets. The DR effectively distinguishes between various degradations in blind face inverse problems and provides a reasonably powerful prompt to LDM. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.10508v1-abstract-full').style.display = 'none'; document.getElementById('2411.10508v1-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.09823">arXiv:2411.09823</a> <span> [<a href="https://arxiv.org/pdf/2411.09823">pdf</a>, <a href="https://arxiv.org/format/2411.09823">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"> Architect: Generating Vivid and Interactive 3D Scenes with Hierarchical 2D Inpainting </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yian Wang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xiaowen Qiu</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+J">Jiageng Liu</a>, <a href="/search/cs?searchtype=author&query=Chen%2C+Z">Zhehuan Chen</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+J">Jiting Cai</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+Y">Yufei Wang</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+T">Tsun-Hsuan Wang</a>, <a href="/search/cs?searchtype=author&query=Xian%2C+Z">Zhou Xian</a>, <a href="/search/cs?searchtype=author&query=Gan%2C+C">Chuang Gan</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="2411.09823v1-abstract-short" style="display: inline;"> Creating large-scale interactive 3D environments is essential for the development of Robotics and Embodied AI research. Current methods, including manual design, procedural generation, diffusion-based scene generation, and large language model (LLM) guided scene design, are hindered by limitations such as excessive human effort, reliance on predefined rules or training datasets, and limited 3D spa… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.09823v1-abstract-full').style.display = 'inline'; document.getElementById('2411.09823v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.09823v1-abstract-full" style="display: none;"> Creating large-scale interactive 3D environments is essential for the development of Robotics and Embodied AI research. Current methods, including manual design, procedural generation, diffusion-based scene generation, and large language model (LLM) guided scene design, are hindered by limitations such as excessive human effort, reliance on predefined rules or training datasets, and limited 3D spatial reasoning ability. Since pre-trained 2D image generative models better capture scene and object configuration than LLMs, we address these challenges by introducing Architect, a generative framework that creates complex and realistic 3D embodied environments leveraging diffusion-based 2D image inpainting. In detail, we utilize foundation visual perception models to obtain each generated object from the image and leverage pre-trained depth estimation models to lift the generated 2D image to 3D space. Our pipeline is further extended to a hierarchical and iterative inpainting process to continuously generate placement of large furniture and small objects to enrich the scene. This iterative structure brings the flexibility for our method to generate or refine scenes from various starting points, such as text, floor plans, or pre-arranged environments. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.09823v1-abstract-full').style.display = 'none'; document.getElementById('2411.09823v1-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 November, 2024; <span class="has-text-black-bis has-text-weight-semibold">originally announced</span> November 2024. </p> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.06899">arXiv:2411.06899</a> <span> [<a href="https://arxiv.org/pdf/2411.06899">pdf</a>, <a href="https://arxiv.org/format/2411.06899">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"> LongSafety: Enhance Safety for Long-Context LLMs </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Huang%2C+M">Mianqiu Huang</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+X">Xiaoran Liu</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+S">Shaojun Zhou</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+M">Mozhi Zhang</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Li%2C+L">Linyang Li</a>, <a href="/search/cs?searchtype=author&query=Tan%2C+C">Chenkun Tan</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+Y">Yang Gao</a>, <a href="/search/cs?searchtype=author&query=Wang%2C+P">Pengyu Wang</a>, <a href="/search/cs?searchtype=author&query=Li%2C+L">Linlin Li</a>, <a href="/search/cs?searchtype=author&query=Liu%2C+Q">Qun Liu</a>, <a href="/search/cs?searchtype=author&query=Zhou%2C+Y">Yaqian Zhou</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Huang%2C+X">Xuanjing Huang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.06899v2-abstract-short" style="display: inline;"> Recent advancements in model architectures and length extrapolation techniques have significantly extended the context length of large language models (LLMs), paving the way for their application in increasingly complex tasks. However, despite the growing capabilities of long-context LLMs, the safety issues in long-context scenarios remain underexplored. While safety alignment in short context has… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06899v2-abstract-full').style.display = 'inline'; document.getElementById('2411.06899v2-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.06899v2-abstract-full" style="display: none;"> Recent advancements in model architectures and length extrapolation techniques have significantly extended the context length of large language models (LLMs), paving the way for their application in increasingly complex tasks. However, despite the growing capabilities of long-context LLMs, the safety issues in long-context scenarios remain underexplored. While safety alignment in short context has been widely studied, the safety concerns of long-context LLMs have not been adequately addressed. In this work, we introduce \textbf{LongSafety}, a comprehensive safety alignment dataset for long-context LLMs, containing 10 tasks and 17k samples, with an average length of 40.9k tokens. Our experiments demonstrate that training with LongSafety can enhance long-context safety performance while enhancing short-context safety and preserving general capabilities. Furthermore, we demonstrate that long-context safety does not equal long-context alignment with short-context safety data and LongSafety has generalizing capabilities in context length and long-context safety scenarios. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.06899v2-abstract-full').style.display = 'none'; document.getElementById('2411.06899v2-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 February, 2025; <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> </li> <li class="arxiv-result"> <div class="is-marginless"> <p class="list-title is-inline-block"><a href="https://arxiv.org/abs/2411.02908">arXiv:2411.02908</a> <span> [<a href="https://arxiv.org/pdf/2411.02908">pdf</a>, <a href="https://arxiv.org/format/2411.02908">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="Distributed, Parallel, and Cluster Computing">cs.DC</span> </div> </div> <p class="title is-5 mathjax"> Photon: Federated LLM Pre-Training </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Sani%2C+L">Lorenzo Sani</a>, <a href="/search/cs?searchtype=author&query=Iacob%2C+A">Alex Iacob</a>, <a href="/search/cs?searchtype=author&query=Cao%2C+Z">Zeyu Cao</a>, <a href="/search/cs?searchtype=author&query=Lee%2C+R">Royson Lee</a>, <a href="/search/cs?searchtype=author&query=Marino%2C+B">Bill Marino</a>, <a href="/search/cs?searchtype=author&query=Gao%2C+Y">Yan Gao</a>, <a href="/search/cs?searchtype=author&query=Cai%2C+D">Dongqi Cai</a>, <a href="/search/cs?searchtype=author&query=Li%2C+Z">Zexi Li</a>, <a href="/search/cs?searchtype=author&query=Zhao%2C+W">Wanru Zhao</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xinchi Qiu</a>, <a href="/search/cs?searchtype=author&query=Lane%2C+N+D">Nicholas D. Lane</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="2411.02908v1-abstract-short" style="display: inline;"> Scaling large language models (LLMs) demands extensive data and computing resources, which are traditionally constrained to data centers by the high-bandwidth requirements of distributed training. Low-bandwidth methods like federated learning (FL) could enable collaborative training of larger models across weakly-connected GPUs if they can effectively be used for pre-training. To achieve this, we… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.02908v1-abstract-full').style.display = 'inline'; document.getElementById('2411.02908v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.02908v1-abstract-full" style="display: none;"> Scaling large language models (LLMs) demands extensive data and computing resources, which are traditionally constrained to data centers by the high-bandwidth requirements of distributed training. Low-bandwidth methods like federated learning (FL) could enable collaborative training of larger models across weakly-connected GPUs if they can effectively be used for pre-training. To achieve this, we introduce Photon, the first complete system for federated end-to-end LLM training, leveraging cross-silo FL for global-scale training with minimal communication overheads. Using Photon, we train the first federated family of decoder-only LLMs from scratch. We show that: (1) Photon can train model sizes up to 7B in a federated fashion while reaching an even better perplexity than centralized pre-training; (2) Photon model training time decreases with available compute, achieving a similar compute-time trade-off to centralized; and (3) Photon outperforms the wall-time of baseline distributed training methods by 35% via communicating 64x-512xless. Our proposal is robust to data heterogeneity and converges twice as fast as previous methods like DiLoCo. This surprising data efficiency stems from a unique approach combining small client batch sizes with extremely high learning rates, enabled by federated averaging's robustness to hyperparameters. Photon thus represents the first economical system for global internet-wide LLM pre-training. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.02908v1-abstract-full').style.display = 'none'; document.getElementById('2411.02908v1-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 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">13 pages, 9 appendix pages, 10 figures, 3 algorithms, 8 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/2411.01855">arXiv:2411.01855</a> <span> [<a href="https://arxiv.org/pdf/2411.01855">pdf</a>, <a href="https://arxiv.org/format/2411.01855">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"> Can Language Models Learn to Skip Steps? </p> <p class="authors"> <span class="search-hit">Authors:</span> <a href="/search/cs?searchtype=author&query=Liu%2C+T">Tengxiao Liu</a>, <a href="/search/cs?searchtype=author&query=Guo%2C+Q">Qipeng Guo</a>, <a href="/search/cs?searchtype=author&query=Hu%2C+X">Xiangkun Hu</a>, <a href="/search/cs?searchtype=author&query=Jiayang%2C+C">Cheng Jiayang</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Y">Yue Zhang</a>, <a href="/search/cs?searchtype=author&query=Qiu%2C+X">Xipeng Qiu</a>, <a href="/search/cs?searchtype=author&query=Zhang%2C+Z">Zheng Zhang</a> </p> <p class="abstract mathjax"> <span class="has-text-black-bis has-text-weight-semibold">Abstract</span>: <span class="abstract-short has-text-grey-dark mathjax" id="2411.01855v1-abstract-short" style="display: inline;"> Trained on vast corpora of human language, language models demonstrate emergent human-like reasoning abilities. Yet they are still far from true intelligence, which opens up intriguing opportunities to explore the parallels of humans and model behaviors. In this work, we study the ability to skip steps in reasoning - a hallmark of human expertise developed through practice. Unlike humans, who may… <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.01855v1-abstract-full').style.display = 'inline'; document.getElementById('2411.01855v1-abstract-short').style.display = 'none';">▽ More</a> </span> <span class="abstract-full has-text-grey-dark mathjax" id="2411.01855v1-abstract-full" style="display: none;"> Trained on vast corpora of human language, language models demonstrate emergent human-like reasoning abilities. Yet they are still far from true intelligence, which opens up intriguing opportunities to explore the parallels of humans and model behaviors. In this work, we study the ability to skip steps in reasoning - a hallmark of human expertise developed through practice. Unlike humans, who may skip steps to enhance efficiency or to reduce cognitive load, models do not inherently possess such motivations to minimize reasoning steps. To address this, we introduce a controlled framework that stimulates step-skipping behavior by iteratively refining models to generate shorter and accurate reasoning paths. Empirical results indicate that models can develop the step skipping ability under our guidance. Moreover, after fine-tuning on expanded datasets that include both complete and skipped reasoning sequences, the models can not only resolve tasks with increased efficiency without sacrificing accuracy, but also exhibit comparable and even enhanced generalization capabilities in out-of-domain scenarios. Our work presents the first exploration into human-like step-skipping ability and provides fresh perspectives on how such cognitive abilities can benefit AI models. <a class="is-size-7" style="white-space: nowrap;" onclick="document.getElementById('2411.01855v1-abstract-full').style.display = 'none'; document.getElementById('2411.01855v1-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">originally announced</span> November 2024. </p> <p class="comments is-size-7"> <span class="has-text-black-bis has-text-weight-semibold">Comments:</span> <span class="has-text-grey-dark mathjax">Accepted by NeurIPS 2024</span> </p> </li> </ol> <nav class="pagination is-small is-centered breathe-horizontal" role="navigation" aria-label="pagination"> <a href="" class="pagination-previous is-invisible">Previous </a> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=50" class="pagination-next" >Next </a> <ul class="pagination-list"> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=0" class="pagination-link is-current" aria-label="Goto page 1">1 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=50" class="pagination-link " aria-label="Page 2" aria-current="page">2 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=100" class="pagination-link " aria-label="Page 3" aria-current="page">3 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=150" class="pagination-link " aria-label="Page 4" aria-current="page">4 </a> </li> <li> <a href="/search/?searchtype=author&query=Qiu%2C+X&start=200" class="pagination-link " aria-label="Page 5" aria-current="page">5 </a> </li> <li><span class="pagination-ellipsis">…</span></li> </ul> </nav> <div class="is-hidden-tablet">  <span class="help" style="display: inline-block;"><a href="https://github.com/arXiv/arxiv-search/releases">Search v0.5.6 released 2020-02-24</a>  </span> </div> </div> </main> <footer> <div class="columns is-desktop" role="navigation" aria-label="Secondary">  <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/about">About</a></li> <li><a href="https://info.arxiv.org/help">Help</a></li> </ul> </div> <div class="column"> <ul class="nav-spaced"> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>contact arXiv</title><desc>Click here to contact arXiv</desc><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg> <a href="https://info.arxiv.org/help/contact.html"> Contact</a> </li> <li> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><title>subscribe to arXiv mailings</title><desc>Click here to subscribe</desc><path d="M476 3.2L12.5 270.6c-18.1 10.4-15.8 35.6 2.2 43.2L121 358.4l287.3-253.2c5.5-4.9 13.3 2.6 8.6 8.3L176 407v80.5c0 23.6 28.5 32.9 42.5 15.8L282 426l124.6 52.2c14.2 6 30.4-2.9 33-18.2l72-432C515 7.8 493.3-6.8 476 3.2z"/></svg> <a href="https://info.arxiv.org/help/subscribe"> Subscribe</a> </li> </ul> </div> </div> </div>   <div class="column"> <div class="columns"> <div class="column"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/license/index.html">Copyright</a></li> <li><a href="https://info.arxiv.org/help/policies/privacy_policy.html">Privacy Policy</a></li> </ul> </div> <div class="column sorry-app-links"> <ul class="nav-spaced"> <li><a href="https://info.arxiv.org/help/web_accessibility.html">Web Accessibility Assistance</a></li> <li> <p class="help"> <a class="a11y-main-link" href="https://status.arxiv.org" target="_blank">arXiv Operational Status <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 256 512" class="icon filter-dark_grey" role="presentation"><path d="M224.3 273l-136 136c-9.4 9.4-24.6 9.4-33.9 0l-22.6-22.6c-9.4-9.4-9.4-24.6 0-33.9l96.4-96.4-96.4-96.4c-9.4-9.4-9.4-24.6 0-33.9L54.3 103c9.4-9.4 24.6-9.4 33.9 0l136 136c9.5 9.4 9.5 24.6.1 34z"/></svg></a><br> Get status notifications via <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/email/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon filter-black" role="presentation"><path d="M502.3 190.8c3.9-3.1 9.7-.2 9.7 4.7V400c0 26.5-21.5 48-48 48H48c-26.5 0-48-21.5-48-48V195.6c0-5 5.7-7.8 9.7-4.7 22.4 17.4 52.1 39.5 154.1 113.6 21.1 15.4 56.7 47.8 92.2 47.6 35.7.3 72-32.8 92.3-47.6 102-74.1 131.6-96.3 154-113.7zM256 320c23.2.4 56.6-29.2 73.4-41.4 132.7-96.3 142.8-104.7 173.4-128.7 5.8-4.5 9.2-11.5 9.2-18.9v-19c0-26.5-21.5-48-48-48H48C21.5 64 0 85.5 0 112v19c0 7.4 3.4 14.3 9.2 18.9 30.6 23.9 40.7 32.4 173.4 128.7 16.8 12.2 50.2 41.8 73.4 41.4z"/></svg>email</a> or <a class="is-link" href="https://subscribe.sorryapp.com/24846f03/slack/new" target="_blank"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon filter-black" role="presentation"><path d="M94.12 315.1c0 25.9-21.16 47.06-47.06 47.06S0 341 0 315.1c0-25.9 21.16-47.06 47.06-47.06h47.06v47.06zm23.72 0c0-25.9 21.16-47.06 47.06-47.06s47.06 21.16 47.06 47.06v117.84c0 25.9-21.16 47.06-47.06 47.06s-47.06-21.16-47.06-47.06V315.1zm47.06-188.98c-25.9 0-47.06-21.16-47.06-47.06S139 32 164.9 32s47.06 21.16 47.06 47.06v47.06H164.9zm0 23.72c25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06H47.06C21.16 243.96 0 222.8 0 196.9s21.16-47.06 47.06-47.06H164.9zm188.98 47.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06s-21.16 47.06-47.06 47.06h-47.06V196.9zm-23.72 0c0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06V79.06c0-25.9 21.16-47.06 47.06-47.06 25.9 0 47.06 21.16 47.06 47.06V196.9zM283.1 385.88c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06-25.9 0-47.06-21.16-47.06-47.06v-47.06h47.06zm0-23.72c-25.9 0-47.06-21.16-47.06-47.06 0-25.9 21.16-47.06 47.06-47.06h117.84c25.9 0 47.06 21.16 47.06 47.06 0 25.9-21.16 47.06-47.06 47.06H283.1z"/></svg>slack</a> </p> </li> </ul> </div> </div> </div>  </div> </footer> <script src="https://static.arxiv.org/static/base/1.0.0a5/js/member_acknowledgement.js"></script> </body> </html>

CINXE.COM

Search | arXiv e-print repository