Multi-Agent Systems
Coordination, communication, and collaboration among multiple intelligent agents in dynamic environments.
Mingze Kong
First-year Ph.D. Student in Computer Science
The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen)
Advised by Prof. Zhongxiang Dai
Large Language Models × Reinforcement Learning
About
I am a first-year Ph.D. student in Computer Science at The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen), advised by Prof. Zhongxiang Dai. My research lies at the intersection of Large Language Models (LLMs) and Reinforcement Learning.
Prior to my doctoral studies, I was a Research Assistant in the same group. I earned my B.S. in Computer Science from Lanzhou University. I am always open to academic collaboration — feel free to reach out.
Research Interests
LLM Agents
LLMs as autonomous decision-makers that plan, execute, and adapt in real-world scenarios.
Reinforcement Learning
Learning optimal policies through interaction, with a focus on sample efficiency and generalization.
Multi-Armed Bandits
Theory and practice of exploration–exploitation trade-offs under uncertainty.
News
2026
New preprint “Workflow-R1: Group Sub-sequence Policy Optimization for Multi-turn Workflow Construction” is on arXiv, with code released.
2025
New preprint “T-POP: Test-Time Personalization with Online Preference Feedback” is on arXiv.
2025
“Online Clustering of Dueling Bandits” is accepted at ICML 2025.
2025
Workshop paper “Meta-Prompt Optimization for LLM-Based Sequential Decision Making” at the Reasoning and Planning for LLMs Workshop @ ICLR 2025.
Publications
Workflow-R1: Group Sub-sequence Policy Optimization for Multi-turn Workflow Construction
arXiv preprint, 2026
Preprint
Meta-Prompt Optimization for LLM-Based Sequential Decision Making
Reasoning and Planning for LLMs Workshop @ ICLR 2025
Workshop