Yingxuan Yang

PhD Student, School of Artificial Intelligence, Shanghai Jiao Tong University

Hi 👋! I am a PhD student at Shanghai Jiao Tong University, advised by Prof. Ying Wen. I received my Bachelor's and Master's degrees from Shanghai Jiao Tong University, advised by Prof. Weinan Zhang as a member of APEX Lab.

My research spans large language models, LLM agents, and multi-agent systems. What I care about most is how these layers connect: how the models we train shape the agents they become, and how those agents can be evaluated and coordinated to work reliably at scale.

I'm always happy to connect—feel free to reach out to discuss research or collaborations at zoeyyx[at]sjtu[dot]edu[dot]cn.

Yingxuan Yang
✉️zoeyyx@sjtu.edu.cn 🎓 💬WeChat: ID_yangyx

Recent News 📰

Internships

Z.ai (Zhipu AI), Post-training TeamFeb 2026 – Jun 2026
Research Intern; Contributor to GLM-5.1 and GLM-5.2.
Agentic coding and agentic RL.
Meituan, Evaluation GroupMay 2024 – Mar 2025
Research Intern; evaluation of LLM agents.
Designed a multi-round benchmark for evaluating agent workflows.

Publications

Show: Selected by Year by Topic Full publication list on Google Scholar.
AgentNet: Decentralized Evolutionary Coordination for LLM-based Multi-Agent Systems
Yingxuan Yang*, Huacan Chai*, Shuai Shao, Yuanyi Song, Siyuan Qi, Renting Rui, Weinan Zhang
NeurIPS 2025
A decentralized, evolutionary coordination framework for LLM multi-agent systems.
Code
Agentic Web: Weaving the Next Web with AI Agents
Yingxuan Yang, Mulei Ma, Yuxuan Huang, Huacan Chai, Chenyu Gong, Haoran Geng, …, Pieter Abbeel, Dawn Song, Weinan Zhang, Jun Wang
arXiv 2025
A vision for the next generation of the web woven together by autonomous AI agents.
Featured in IEEE Spectrum & The Economist
Understanding and Optimizing Agentic Workflows via Shapley Value
Yingxuan Yang, Bo Huang, Siyuan Qi, Chengyu Feng, Haoyi Hu, Yichen Zhu, …, Yong Yu, Weinan Zhang
ACL 2026 Findings
Attributing and optimizing agentic workflows through a Shapley-value lens.
Code
Understanding Agent Scaling in LLM-Based Multi-Agent Systems via Diversity
Yingxuan Yang, Chengrui Qu, Muning Wen, Laixi Shi, Ying Wen, Weinan Zhang, Adam Wierman, Shangding Gu
Under Review
How diversity governs the scaling behavior of LLM-based multi-agent systems.
Code
Diversity-Aware Self-Paced Data Selection for LLM Fine-Tuning
Yingxuan Yang, Huayi Wang, Muning Wen, Xiaoyun Mo, Qiuying Peng, Jun Wang, Weinan Zhang
ECAI 2025 (Oral)
Diversity-aware, self-paced data selection for more efficient LLM fine-tuning.
Progra: Progress-Aware Reinforcement Learning for Multi-Turn Function Calling
Huacan Chai, Zijie Cao, Maolin Ran, Yingxuan Yang, Jianghao Lin, Xin Peng, …, Fei Huang, Ying Wen
ACL 2026 Findings (SAC Highlights) 🏆
Injects progress awareness into multi-turn function-calling LLMs via task-progress summaries and a progress-guided RL objective that aligns local actions with global task completion.
TRAD: Enhancing LLM Agents with Step-Wise Thought Retrieval and Aligned Decision
Ruiwen Zhou, Yingxuan Yang, Muning Wen, Ying Wen, Wenhao Wang, Chunling Xi, Guoqiang Xu, Yong Yu, Weinan Zhang
SIGIR 2024
Step-wise thought retrieval and aligned decision-making for stronger LLM agents.
Code
Agent Exchange: An Auction Platform for AI Agent Marketplaces
Yingxuan Yang, Ying Wen, Jun Wang, Weinan Zhang
ICLR 2026 Workshop on AI for Mechanism Design and Strategic Decision Making
An auction-based marketplace platform for trading AI agent services.
DynaTree: Dynamic Agentic Retrieval Tree for Time-Sensitive News Retrieval
Siyuan Qi, Xin Wang, Yingxuan Yang, Hao Guo, Jianghao Lin, Weiwen Liu, Yong Yu, Weinan Zhang
KDD 2026
A dynamic agentic retrieval tree for freshness-aware, time-sensitive news retrieval.
Where LLM Agents Fail and How They Can Learn from Failures
Kunlun Zhu, Zhenyu Liu, Bowen Li, Minghang Tian, Yingxuan Yang, …, James Zou, Jiaxuan You
Manuscript
Diagnosing where LLM agents fail and turning those failures into learning signals.
Agents' Last Exam
Yiyou Sun*, Xinyang Han*, Weichen Zhang*, …, Yingxuan Yang, …, Dawn Song
arXiv 2026
A frontier benchmark probing the limits of today's autonomous agents.
Which LLM Multi-Agent Protocol to Choose?
Hongyi Du, Jiaqi Su, Jisen Li, Lijie Ding, Yingxuan Yang, Peixuan Han, Xiangru Tang, Kunlun Zhu, Jiaxuan You
ICML 2026
A systematic comparison to guide the choice of multi-agent communication protocols.
SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory
Huacan Chai, Y. Wang, Yingxuan Yang, D. Peng, Yuanyi Song, Z. Fu, Weiwen Liu, Jianghao Lin, Jun Wang, et al.
arXiv 2026
A benchmark for evaluating source-distributed multimodal memory in LLM agents.
Unlocking the Potential of Decentralized LLM Multi-Agent Systems: Privacy Preservation and Monetization
Yingxuan Yang, Qiuying Peng, Jun Wang, Ying Wen, Weinan Zhang
AAMAS 2025
Privacy-preserving and monetizable designs for decentralized multi-agent systems.
Featured in MIT Technology Review China
A Survey of AI Agent Protocols
Yingxuan Yang, Huacan Chai, Yuanyi Song, Siyuan Qi, Muning Wen, Ning Li, …, Ying Wen, Yong Yu, Weinan Zhang
arXiv 2025
A comprehensive survey of communication and interoperability protocols for AI agents.
Code
Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
Chenyu Zhou, Huacan Chai*, Wenteng Chen*, …, Yingxuan Yang*, …
arXiv 2026
A unified review of memory, skills, protocols, and harness engineering in LLM agents.

Professional Activities

Invited Talks
Standards
Academic Service
Teaching