About Zijian

Hi, welcome to my website! I am Zijian Li (李子健). I received my Ph.D. degree from the Department of Electronic and Computer Engineering at HKUST, under the supervision of Prof. Jun Zhang.

I have been fortunate to conduct research internships at MSRA, Tongyi Lab, and Tencent. My current research interests lie in long-horizon data scalability and agentic reinforcement learning.

If you are interested in my research, please feel free to contact me at zijian dot li at connect dot ust dot hk.

News

• 2026/4: One paper is accepted by ICML 2026!
• 2026/4: One paper is accepted by ACL 2026 Main!
• 2026/2: One paper is accepted by CVPR 2026!
• 2026/1: Three papers are accepted by ICLR 2026!
• 2026/1: Our work "Evidence-Augmented Policy Optimization with Reward Co-Evolution for Long-Context Reasoning" releases to Arxiv! [paper]
• 2025/12: Our work "SIT-GRAPH: State Integrated Tool Graph For Multi-Turn Agents" releases to Arxiv! [paper]
• 2025/10: Our paper Chain of Functions: A Programmatic Pipeline for Fine-Grained Chart Reasoning Data is accepted by AACL 2025 Main.
• 2025/09: We are grad to release our new powerful deep research model: Tongyi DeepResearch (30B) [blob] [github] [model].
  
• 2025/09: We are grad to release our new work WebWeaver: Structuring Web-Scale Evidence with Dynamic Outlines for Open-Ended Deep Research [Hugging Face] [github]
• 2025/09: We are grad to release our new work PixelCraft: A Multi-Agent System for High-Fidelity Visual Reasoning on Structured Images [Hugging Face] [github]
  

Experiences

	Research Intern, Tencent Inc. Jan. 2026 - July 2026 Interests: Agentic RL
	Research Intern, Tongyi Lab, Alibaba Inc. July 2025 - Jan 2026 Advisor: Yong Jiang, Shen Huang Interests: Deep Research, Agent Training
	Research Intern, Machine Learning Group, Microsoft Research Asia Jan. 2024 - July 2025 Advisor: Rui Wang, Jiang Bian Interests: RAG, Agent, Chart Understanding

Selected Publications

• [ICLR 2026] PixelCraft: A Multi-Agent System for High-Fidelity Visual Reasoning on Structured Images [paper] [Code]
  Shuoshuo Zhang* Zijian Li*, Yizhen Zhang, Jingjing Fu, Lei Song, Jiang Bian, Jun Zhang, Yujiu Yang, Rui Wang
  

• [ICLR 2026] WebWeaver: Structuring Web-Scale Evidence with Dynamic Outlines for Open-Ended Deep Research [paper] [Code]
  Zijian Li, Xin Guan, Bo Zhang, Shen Huang, Houquan Zhou, Shaopeng Lai, Ming Yan, Yong Jiang, Pengjun Xie, Fei Huang, Jun Zhang, Jingren Zhou
  

• [ICLR 2026] Token-level Data Selection for Safe LLM Fine-tuning [paper]
  Yanping Li, Zhening Liu, Zijian Li, Zehong Lin, Jun Zhang
  

• [ICML 2026] Experience-Evolving Multi-Turn Tool-Use Agent with Hybrid Episodic-Procedural Memory [paper]
  Sijia Li, Yuchen Huang, Zifan Liu, Zijian Li, Jingjing fu, Lei Song, Jiang Bian, Jun Zhang, Rui Wang
  

• [CVPR 2026] RemedyGS: Defend 3D Gaussian Splatting against Computation Cost Attacks [paper]
  Yanping Li, Zhening Liu, Zijian Li, Zehong Lin, Jun Zhang
  

• [AACL 2025 Main] Chain of Functions: A Programmatic Pipeline for Fine-Grained Chart Reasoning Data [paper] [Code]
  Zijian Li, Jingjing Fu, Lei Song, Jiang Bian, Jun Zhang, Rui Wang
  

• [NAACL 2025 Main] Graph Neural Network Enhanced Retrieval for Question Answering of Large Language Models [paper] [code]
  Zijian Li, Qingyan Guo, Jiawei Shao, Lei Song, Jiang Bian, Jun Zhang, Rui Wang
  

• [IEEE Trans. Mob. Comput. (CCF A)] Federated Client-Invariant Representation Learning [paper]
  Zijian Li, Zehong Lin, Jiawei Shao, Yuyi Mao, Jun Zhang
  

• [IEEE Trans. Mob. Comput. (CCF A)] Feature Matching Data Synthesis for Federated Non-IID Data [paper]
  Zijian Li, Yuchang Sun, Jiawei Shao, Yuyi Mao, Jessie Hui Wang, Jun Zhang
  

• [IJCAI 2022 workshop] Federated learning with GAN-based data synthesis for non-IID clients [paper]
  Zijian Li, Jiawei Shao, Yuyi Mao, Jessie Hui Wang, Jun Zhang
  

• [IEEE Trans. Mob. Comput. (CCF A)] Delayed Local-SGD for Distributed Training for Linear Speedup [paper]
  XiaoLu Wang, Zijian Li, Shi Jin, Jun Zhang
  

• [Preprint] A survey of what to share in federated learning: Perspectives on model utility, privacy leakage, and communication efficiency [paper]
  Jiawei Shao*, Zijian Li*, Wenqiang Sun*, Tailin Zhou, Yuchang Sun, Lumin Liu, Zehong Lin, Yuyi Mao, Jun Zhang