Research
Research Interests / 研究方向
My research interests include Reinforcement Learning, Multi-Agent Reinforcement Learning, and Game AI. I am especially interested in robust evaluation, out-of-distribution opponent generalization, and agent behavior in complex games.
我的研究方向包括强化学习、多智能体强化学习和游戏 AI， 重点关注鲁棒评估、分布外对手泛化以及复杂游戏中的智能体行为。
Current Work
Current Work / 当前工作
-
Opponent Generalization in SMAC-Hard
Studying out-of-distribution opponent generalization in reinforcement learning using SMAC-Hard as the evaluation environment.
基于 SMAC-Hard 环境研究强化学习中的分布外对手泛化问题。
-
LLM Agents in Real Games
Building on STS2CLI to study long-context decision making and experience transfer for large language model agents in real games.
基于 STS2CLI 研究大模型智能体在真实游戏中的超长上下文决策和经验迁移问题。
CyberBarbarian/sts2-cli Headless Slay the Spire 2 CLI — play the full game from a terminal.
