DIG/publications.html at main · Gen-Verse/DIG · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
<!DOCTYPE html>
<html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>Publications — DIG</title><meta name="description" content="Selected publications from the Discovery Intelligence Group and Ling Yang."><link rel="icon" href="assets/dig-logo.svg"><link rel="stylesheet" href="assets/site.css"></head><body>
<nav class="site-nav"><div class="nav-in"><a class="brand" href="index.html"><img src="assets/dig-logo.svg" alt="DIG"><span><span class="name">DIG</span><span class="sub">Discovery Intelligence Group</span></span></a><button class="menu-btn" aria-label="Toggle navigation">☰</button><div class="nav-links"><a href="research.html">Research</a><a href="people.html">People</a><a href="projects.html">Projects</a><a href="publications.html">Publications</a><a href="join.html" class="nav-cta">Join DIG</a></div></div></nav>
<header class="page-hero"><div class="wrap"><div class="kicker">Selected publications</div><h1>Work across models, agents, RL, and discovery.</h1><p>This page highlights representative work connected to the research trajectory behind DIG. For the complete publication record, see the PI's Google Scholar profile.</p><div class="hero-actions"><a class="btn primary" href="https://scholar.google.com.hk/citations?user=sIKujqAAAAAJ&hl=en" target="_blank" rel="noopener">Full Google Scholar ↗</a></div></div></header>
<main class="section"><div class="wrap"><div class="pub-list">
<div class="pub"><div class="venue">ICML 2026</div><div><h3>RLAnything: Forge Environment, Policy, and Reward Model in Completely Dynamic RL System</h3><p>General agentic RL · corresponding-author work</p></div><div class="pub-links"><a class="mini" href="https://www.arxiv.org/abs/2602.02488">Paper</a><a class="mini" href="https://github.com/Gen-Verse/Open-AgentRL">Code</a></div></div>
<div class="pub"><div class="venue">ICML 2026<span class="award">Spotlight · Top 3%</span></div><div><h3>Latent Collaboration in Multi-Agent Systems (LatentMAS)</h3><p>Efficient multi-agent reasoning in latent space · corresponding-author work</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2511.20639">Paper</a><a class="mini" href="https://github.com/Gen-Verse/LatentMAS">Code</a></div></div>
<div class="pub"><div class="venue">TECH REPORT</div><div><h3>OpenClaw-RL: Train Any Agent Simply by Talking</h3><p>Fully asynchronous RL for interactive agents</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2603.10165">Paper</a><a class="mini" href="https://github.com/Gen-Verse/OpenClaw-RL">Code</a></div></div>
<div class="pub"><div class="venue">ICLR 2026</div><div><h3>Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models (TraceRL)</h3><p>Trajectory-aware RL for diffusion LLMs · corresponding-author work</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2509.06949">Paper</a><a class="mini" href="https://github.com/Gen-Verse/dLLM-RL">Code</a></div></div>
<div class="pub"><div class="venue">NEURIPS 2025</div><div><h3>MMaDA: Multimodal Large Diffusion Language Models</h3><p>Unified multimodal generative modeling</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2505.15809">Paper</a><a class="mini" href="https://github.com/Gen-Verse/MMaDA">Code</a></div></div>
<div class="pub"><div class="venue">NEURIPS 2025<span class="award">Spotlight · Top 3%</span></div><div><h3>Co-Evolving LLM Coder and Unit Tester via Reinforcement Learning (ReasonFlux-Coder)</h3><p>Coder / unit-tester co-evolution via RL · corresponding-author work</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2506.03136">Paper</a><a class="mini" href="https://github.com/Gen-Verse/CURE">Code</a></div></div>
<div class="pub"><div class="venue">NEURIPS 2025</div><div><h3>ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs</h3><p>Trajectory-aware process reward models · corresponding-author work</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2506.18896">Paper</a><a class="mini" href="https://github.com/Gen-Verse/ReasonFlux">Code</a></div></div>
<div class="pub"><div class="venue">PREPRINT</div><div><h3>ReasonFlux: Hierarchical LLM Reasoning via Scaling Thought Templates</h3><p>Structured reasoning and post-training</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/abs/2502.06772">Paper</a><a class="mini" href="https://github.com/Gen-Verse/ReasonFlux">Code</a></div></div>
<div class="pub"><div class="venue">NEURIPS 2024<span class="award">Spotlight · Top 3%</span></div><div><h3>Buffer of Thoughts: Thought-Augmented Reasoning with Large Language Models</h3><p>Structured reasoning with thought templates</p></div><div class="pub-links"><a class="mini" href="https://arxiv.org/pdf/2406.04271">Paper</a><a class="mini" href="https://github.com/YangLing0818/buffer-of-thought-llm">Code</a></div></div>
<div class="pub"><div class="venue">ICML 2024</div><div><h3>Mastering Text-to-Image Diffusion: Recaptioning, Planning, and Generating with Multimodal LLMs</h3><p>Generative modeling foundation (RPG)</p></div><div class="pub-links"><a class="mini" href="https://openreview.net/forum?id=DgLFkAPwuZ">Paper</a><a class="mini" href="https://github.com/YangLing0818/RPG-DiffusionMaster">Code</a></div></div>
</div></div></main>
<section class="section"><div class="wrap"><div class="callout"><div><div class="kicker">Research record</div><h2>70+ papers across machine learning and AI.</h2><p>Representative venues include NeurIPS, ICML, ICLR, CVPR, ACL and related conferences, with multiple Spotlight, Oral, and Best Paper recognitions.</p></div><div class="callout-note"><b>Complete bibliography</b>Use Google Scholar for the full and most up-to-date publication record.</div></div></div></section>
<footer><div class="wrap footer-in"><span>© 2026 Discovery Intelligence Group · Peking University</span><span class="footer-links"><a href="https://github.com/Gen-Verse">GitHub</a><a href="https://huggingface.co/Gen-Verse">Hugging Face</a><a href="mailto:yangling0818@163.com">Contact</a></span></div></footer><script src="assets/site.js"></script></body></html>