publication.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
  "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
<head>
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta http-equiv="Content-Type" content="text/html;charset=utf-8" />
<link rel="icon" type="image/svg+xml" href="./images/SJTU-logo.svg">
<link rel="stylesheet" href="style.css" type="text/css" />
<title>Publications | SAIL Lab @ SJTU</title>
<!-- MathJax -->
<script src='https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/latest.js?config=TeX-MML-AM_CHTML' async>
</script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
	  TeX: { equationNumbers: { autoNumber: "AMS" } }
});
</script>
<!-- End MathJax -->
</head>
<body>

<header class="site-header">
  <div class="brand-bar">
    <div class="brand-inner">
      <a class="brand-seal" href="https://en.sjtu.edu.cn/" target="_blank"><img src="./images/SJTU-logo-white.svg" alt="Shanghai Jiao Tong University" /></a>
      <div class="brand-title">
        <a href="index.html">SAIL Lab</a>
        <span class="brand-sub">Safe AI and Robot Learning Lab &middot; Shanghai Jiao Tong University</span>
      </div>
      <div class="brand-links">
        <a href="https://github.com/SAIL-Research-Lab" target="_blank" title="SAIL Lab GitHub"><svg viewBox="0 0 16 16" aria-hidden="true"><path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.01 8.01 0 0 0 16 8c0-4.42-3.58-8-8-8z"/></svg></a>
        <a href="https://www.youtube.com/@ai-agent-research" target="_blank" title="SAIL Lab YouTube"><svg viewBox="0 0 24 24" aria-hidden="true"><path d="M23.5 6.19a3.02 3.02 0 0 0-2.12-2.14C19.5 3.55 12 3.55 12 3.55s-7.5 0-9.38.5A3.02 3.02 0 0 0 .5 6.19C0 8.07 0 12 0 12s0 3.93.5 5.81a3.02 3.02 0 0 0 2.12 2.14c1.88.5 9.38.5 9.38.5s7.5 0 9.38-.5a3.02 3.02 0 0 0 2.12-2.14C24 15.93 24 12 24 12s0-3.93-.5-5.81zM9.55 15.57V8.43L15.82 12l-6.27 3.57z"/></svg></a>
        <a href="mailto:shangding.gu@sjtu.edu.cn">shangding.gu@sjtu.edu.cn</a>
      </div>
      <button id="nav-toggle" aria-label="Menu" onclick="document.body.classList.toggle('nav-open')">&#9776;</button>
    </div>
  </div>
  <nav class="site-nav">
    <div class="nav-inner">
      <ul class="nav-list">
<li class="nav-item"><a href="index.html">Home</a></li>
<li class="nav-item"><a href="group.html">People</a></li>
<li class="nav-item current"><a href="publication.html">Publications</a></li>
<li class="nav-item"><a href="projects.html">Research &amp; Projects</a></li>
<li class="nav-item"><a href="softwares.html">Products &amp; Software</a></li>
<li class="nav-item"><a href="tutorial_talk.html">Tutorials &amp; Talks</a></li>
<li class="nav-item"><a href="academic_service.html">Academic Service</a></li>
<li class="nav-item"><a href="resources.html">Resources</a></li>
<li class="nav-item"><a href="join.html">Join Us</a></li>
      </ul>
    </div>
  </nav>
</header>

<div class="page">
<h1>Publications</h1>
<div class="muted" style="margin-top:-6px;">Selected recent publications and preprints by Prof. Shangding Gu and lab members.</div>
<h2>Selected Recent Publications</h2>
  <div class="pub-row">
    <div class="paper-tag">ICLR 2025</div>
    <div class="pub-main">
      <p><strong>Robust Gymnasium: A Unified Modular Benchmark for Robust Reinforcement Learning.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Laixi Shi, Muning Wen, Ming Jin, Eric Mazumdar, Yuejie Chi, Adam Wierman, Costas Spanos.</p>
      <p><em>International Conference on Learning Representations.</em></p>
      <p><a href="https://arxiv.org/pdf/2502.19652">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/Robust-Gymnasium">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">IEEE TPAMI</div>
    <div class="pub-main">
      <p><strong>Safe and Balanced: A Framework for Constrained Multi-Objective Reinforcement Learning.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Bilgehan Sel, Yuhao Ding, Lu Wang, Qingwei Lin, Alois Knoll, Ming Jin.</p>
      <p><em>IEEE Transactions on Pattern Analysis and Machine Intelligence.</em></p>
      <p><a href="https://arxiv.org/pdf/2405.16390">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/CMORL">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">TMLR</div>
    <div class="pub-main">
      <p><strong>TeaMs-RL: Teaching LLMs to Generate Better Instruction Datasets via Reinforcement Learning.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Alois Knoll, Ming Jin.</p>
      <p><em>Transactions on Machine Learning Research.</em></p>
      <p><a href="https://arxiv.org/pdf/2403.08694">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/TeaMs-RL">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">NeurIPS 2024</div>
    <div class="pub-main">
      <p><strong>Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Laixi Shi, Yuhao Ding, Alois Knoll, Costas Spanos, Adam Wierman, Ming Jin.</p>
      <p><em>Advances in Neural Information Processing Systems.</em></p>
      <p><a href="https://arxiv.org/pdf/2405.20860">[Paper]</a>, <a href="https://github.com/SafeRL-Lab">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">IEEE TPAMI</div>
    <div class="pub-main">
      <p><strong>A Review of Safe Reinforcement Learning: Methods, Theory and Applications.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Long Yang, Yali Du, Guang Chen, Florian Walter, Jun Wang, Alois Knoll.</p>
      <p><em>IEEE Transactions on Pattern Analysis and Machine Intelligence.</em></p>
      <p><a href="https://arxiv.org/pdf/2205.10330">[Paper]</a>, <a href="https://github.com/chauncygu/Safe-Reinforcement-Learning-Baselines">[Code]</a>, <a href="https://mp.weixin.qq.com/s/Ks2EdGstTPngPrnZFcAdSw">[机器之心]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">AAAI 2024</div>
    <div class="pub-main">
      <p><strong>Balancing Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Bilgehan Sel, Yuhao Ding, Lu Wang, Qingwei Lin, Ming Jin, Alois Knoll.</p>
      <p><em>Association for the Advancement of Artificial Intelligence.</em></p>
      <p><a href="https://arxiv.org/pdf/2405.01677">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/Safety-MuJoCo">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">IEEE TII</div>
    <div class="pub-main">
      <p><strong>Safe Multiagent Learning With Soft-Constrained Policy Optimization in Real Robot Control.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Dianye Huang, Muning Wen, Guang Chen, Alois Knoll.</p>
      <p><em>IEEE Transactions on Industrial Informatics.</em></p>
      <p><a href="#">[Paper]</a>, <a href="https://github.com/SAIL-Research-Lab">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">IEEE TASE</div>
    <div class="pub-main">
      <p><strong>ROSCOM: Robust Safe Reinforcement Learning on Stochastic Constraint Manifolds.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Puze Liu, Alap Kshirsagar, Guang Chen, Jan Peters, Alois Knoll.</p>
      <p><em>IEEE Transactions on Automation Science and Engineering.</em></p>
      <p><a href="https://ieeexplore.ieee.org/abstract/document/10616119">[Paper]</a>, <a href="https://github.com/SafeRL-Lab">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">AIJ</div>
    <div class="pub-main">
      <p><strong>Safe Multi-Agent Reinforcement Learning for Multi-Robot Control.</strong></p>
      <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Jakub Grudzien Kuba, Yuanpei Chen, Yali Du, Long Yang, Alois Knoll, Yaodong Yang.</p>
      <p><em>The Journal of Artificial Intelligence.</em></p>
      <p><a href="https://www.sciencedirect.com/science/article/abs/pii/S0004370223000516">[Paper]</a>, <a href="https://github.com/chauncygu/Multi-Agent-Constrained-Policy-Optimisation">[Code]</a></p>
    </div>
  </div>
  <div class="pub-row">
    <div class="paper-tag">IEEE TAI</div>
    <div class="pub-main">
      <p><strong>Safe Multi-Agent Reinforcement Learning with Bilevel Optimization in Autonomous Driving.</strong></p>
      <p>Zheng Zhi, <b><span style="text-decoration: underline">Shangding Gu</span></b>.</p>
      <p><em>IEEE Transactions on Artificial Intelligence.</em></p>
      <p><a href="https://arxiv.org/pdf/2405.18209">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/Safe-MARL-in-Autonomous-Driving">[Code]</a></p>
    </div>
  </div>

  <!-- ================================ Preprints ================================ -->
<h2>Selected Recent Preprints</h2>

<ul>
<li>
  <p><strong>From Model Scaling to System Scaling: Scaling the Harness in Agentic AI.</strong></p>
  <p><b><span style="text-decoration: underline">Shangding Gu</span></b></p>
  <p><a href="https://arxiv.org/pdf/2605.26112">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/cheetahclaws">[Code]</a>, <a href="https://x.com/dair_ai/status/2059294269698199929">[DAIR.AI]</a></p>
</li>
<li>
  <p><strong>Long Context, Less Focus: A Scaling Gap in LLMs Revealed through Privacy and Personalization.</strong></p>
  <p><b><span style="text-decoration: underline">Shangding Gu</span></b></p>
  <p><a href="https://arxiv.org/pdf/2602.15028">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/PAPerBench">[Code]</a></p>
</li>
<li>
  <p><strong>Understanding Agent Scaling in LLM-Based Multi-Agent Systems via Diversity.</strong></p>
  <p>Yingxuan Yang, Chengrui Qu, Muning Wen, Laixi Shi, Ying Wen, Weinan Zhang, Adam Wierman, <b><span style="text-decoration: underline">Shangding Gu</span></b></p>
  <p><a href="https://arxiv.org/pdf/2602.03794">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/agentic-web">[Code]</a>, <a href="https://mp.weixin.qq.com/s/aFg6PpQcPXMjd22hLWsG8g">[机器之心]</a></p>
</li>
<li>
  <p><strong>Agentic web: Weaving the next web with AI agents.</strong></p>
  <p>Yingxuan Yang, Mulei Ma, Yuxuan Huang, Huacan Chai, Chenyu Gong, Haoran Geng, Yuanjian Zhou, Ying Wen, Meng Fang, Muhao Chen, <b><span style="text-decoration: underline">Shangding Gu</span></b>, Ming Jin, Costas Spanos, Yang Yang, Pieter Abbeel, Dawn Song, Weinan Zhang, Jun Wang</p>
  <p><a href="https://arxiv.org/pdf/2507.21206?">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/Agent-Scaling">[Code]</a>, <a href="https://spectrum.ieee.org/agentic-web">[IEEE Spectrum]</a>, <a href="https://www.economist.com/interactive/science-and-technology/2025/12/10/the-next-version-of-the-web-will-be-built-for-machines-not-humans">[The Economist]</a>, <a href="https://mp.weixin.qq.com/s/Co1lBdo-nhErdeAFdewyCg">[机器之心]</a></p>
</li>
<li>
  <p><strong>AccidentBench: Benchmarking Multimodal Understanding and Reasoning in Vehicle Accidents and Beyond.</strong></p>
  <p><b><span style="text-decoration: underline">Shangding Gu</span></b>, Xiaohan Wang, Donghao Ying, Haoyu Zhao, Runing Yang, Boyi Li, Ming Jin, Marco Pavone, Serena Yeung-Levy, Jun Wang, Dawn Song, Costas Spanos</p>
  <p><a href="https://accident-bench.github.io/">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/AccidentBench">[Code]</a></p>
</li>
<li>
  <p><strong>Data Uniformity Improves Training Efficiency and More, with a Convergence Framework Beyond the NTK Regime.</strong></p>
  <p>Yuqing Wang, <b><span style="text-decoration: underline">Shangding Gu</span></b></p>
  <p><a href="https://arxiv.org/pdf/2506.24120">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/data-uniformity">[Code]</a></p>
</li>
<li>
  <p><strong>RLBenchNet: The Right Network for the Right Reinforcement Learning Task.</strong></p>
  <p>Ivan Smirnov, <b><span style="text-decoration: underline">Shangding Gu</span></b></p>
  <p><a href="https://arxiv.org/pdf/2505.15040">[Paper]</a>, <a href="https://github.com/SafeRL-Lab/BenchNetRL">[Code]</a></p>
</li>
</ul>

<p class="muted">For more details, please see <a href="https://scholar.google.com/citations?user=E1GCDXUAAAAJ&hl=en" target="_blank">Google Scholar</a>.</p>

<!-- ====================================================================
   Archived: previously displayed full list (hidden on request; preserved for reference)
====================================================================
<ul>
<li><p><b><span style="text-decoration: underline">Gu, S*.,</span></b> Shi, L., Wen, M., Jin, M., Mazumdar, E., Chi, Y., Wierman, A., Spanos, C.. (2025). Robust Gymnasium: A Unified Modular Benchmark for Robust Reinforcement Learning.    <i>ICLR 2025.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Sel, B., Ding, Y., Wang, L., Lin, Q., Knoll, A., & Jin, M. (2025). Safe and Balanced: A Framework for Constrained Multi-Objective Reinforcement Learning.   <i>IEEE Transactions on Pattern Analysis and Machine Intelligence.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Knoll, A., & Jin, M. (2024). TeaMs-RL: Teaching LLMs to Generate Better Instruction Datasets via Reinforcement Learning. <i>Transactions on Machine Learning Research.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu*, S.,</span></b> Shi, L., Ding, Y., Knoll, A., Spanos, C., Wierman, A., & Jin, M. (2024). Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation. <i>NeurIPS.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Yang, L., Du, Y., Chen, G., Walter, F., Wang, J., & Knoll, A. (2024). A review of safe reinforcement learning: Methods, theory and applications. <i>IEEE Transactions on Pattern Analysis and Machine Intelligence.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Zheng, Z.,</span></b> & <b><span style="text-decoration: underline">Gu*, S.</span></b> (2024). Safe Multi-Agent Reinforcement Learning with Bilevel Optimization in Autonomous Driving. <i>IEEE Transactions on Artificial Intelligence.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Liu, P., Kshirsagar, A., Chen, G., Peters, J., Knoll, A. (2024). ROSCOM: Robust Safe Reinforcement Learning on a Stochastic Constraint Manifolds. <i>IEEE Transactions on Automation Science and Engineering.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Huang, D., Wen, M., Chen, G., Knoll, A. (2024). Safe Multi-Agent Learning with Soft Constrained Policy Optimization in Real Robot Control. <i>IEEE Transactions on Industrial Informatics.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Bilgehan S., Ding, Y., Wang, L., Lin, Q., Jin, M., Knoll, A. (2024). Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation. <i>AAAI 2024 (Oral paper).</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Kuba, J. G., Chen, Y., Du, Y., Yang, L., Knoll, A., & Yang, Y. (2023). Safe multi-agent reinforcement learning for multi-robot control. <i>Artificial Intelligence, 319, 103905.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Kshirsagar, A., Du, Y., Chen, G., Peters, J., & Knoll, A. (2023). A human-centered safe robot reinforcement learning framework with interactive behaviors. <i>Frontiers in Neurorobotics, 17.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Chen, G., Zhang, L., Hou, J., Hu, Y., & Knoll, A. (2022). Constrained Reinforcement Learning for Vehicle Motion Planning with Topological Reachability Analysis. <i>Robotics, 11(4), 81. (Editor selected paper).</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Zhu, M., Chen, G., Wen, Y., & Knoll, A. (2022). Computing position error margin for a USV due to wind and current with a trajectory model. <i>Ocean Engineering, 262, 111950. (Top Journal in this area).</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Zhou, C., Wen, Y., Xiao, C., & Knoll, A. (2022). Motion Planning for an Unmanned Surface Vehicle with Wind and Current Effects. <i>Journal of Marine Science and Engineering, 10(3), 420.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Zhou, C.,</span></b> <b><span style="text-decoration: underline">Gu*, S.,</span></b> Wen, Y*., Du, Z., Xiao, C., Huang, L., & Zhu, M. (2020). The review unmanned surface vehicle path planning: Based on multi-modality constraint. <i>Ocean Engineering, 200, 107043. (Corresponding author, Top Journal in this area).</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Zhou, C.,</span></b> <b><span style="text-decoration: underline">Gu*, S.,</span></b> Wen, Y*., Du, Z., Xiao, C., Huang, L., & Zhu, M. (2020). Motion planning for an unmanned surface vehicle based on topological position maps. <i>Ocean Engineering, 198, 106798. (Corresponding author, Top Journal in this area).</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Zhou, C., Wen, Y., Zhong, X., Zhu, M., Xiao, C., & Du, Z. (2020). A motion planning method for unmanned surface vehicle in restricted waters. <i>Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment, 234(2), 332-345.</i></p>
</li>
<li><p><b><span style="text-decoration: underline">Gu, S.,</span></b> Zhou, C., Wen, Y., Xiao, C., Du, Z., & Huang, L. (2019). Path Search of Unmanned Surface Vehicle Based on Topological Location. <i>Navigation of China, 42(02), 52-58.</i></p>
</li>
</ul>
-->
</div>
<footer class="site-footer">
  <div class="footer-inner">
    &copy; 2026 SAIL Lab (Safe AI and Robot Learning Lab) &mdash; School of Computer Science, Shanghai Jiao Tong University<br/>
    Principal Investigator: Prof. Shangding Gu (顾尚定) &middot;
    <a href="mailto:shangding.gu@sjtu.edu.cn">shangding.gu@sjtu.edu.cn</a> &middot;
    <a href="https://github.com/SAIL-Research-Lab" target="_blank">GitHub</a> &middot;
    <a href="https://www.youtube.com/@ai-agent-research" target="_blank">YouTube</a> &middot;
    <a href="https://scholar.google.com/citations?user=E1GCDXUAAAAJ&hl=en" target="_blank">Google Scholar</a>
  </div>
</footer>

<script>
document.querySelectorAll('.nav-item a').forEach(function(a){
  a.addEventListener('click', function(){ document.body.classList.remove('nav-open'); });
});
</script>
</body>
</html>