清华论文提出 AAArena：智能体靠回放可登顶，复杂规则游戏仍难突破
New Tsinghua paper finds that AI agents improving game bots from match replays can top human leaderboards, but mostly stall on games with complex rules.
清华大学一篇新论文提出 AAArena 基准，取自清华年度机器人竞赛的 12 款游戏，并以 1920 个存档的人类程序作为对手。编码智能体在不改变模型权重的情况下读取规则、挑选对手、研究回放并在对局预算内重写自己的机器人。在测试的 3 款游戏中，详细回放均优于仅胜负反馈，Pacman 机器人凭回放达到第 1 名，无回放时为第 11 名；但对局预算增至三倍，4 个卡住的机器人仍无一进入第 1 名。论文地址为 arxiv.org/abs/2610.12341。
正文
当前显示原文，中文全文译文尚未就绪。
New Tsinghua paper finds that AI agents improving game bots from match replays can top human leaderboards, but mostly stall on games with complex rules.
Getting AI to learn a winning game strategy from a limited number of matches is still hard, especially against changing rivals.
They built AAArena from 12 games in Tsinghua's yearly bot-building contest, with 1,920 archived human programs as rivals. A coding agent, with its model weights unchanged, reads the rules, picks opponents, studies replays, and rewrites its bot within a match budget.
Detailed replays beat win/loss-only feedback in all 3 games tested. With replays, a Pacman bot reached rank 1, versus rank 11 without them.
Tripling the match budget did not push any of 4 stuck bots to rank 1.
– arxiv. org/abs/2610.12341
Title: "Can AI Agents Learn Their Way to the Top? Evaluating Heuristic Learning in a Long-Running Game Agent Competition"