Unverified50% confidenceFactExact time
OpenAI的hide-and-seek实验中智能体学会利用物理引擎的边界条件穿越障碍物
1
Sources
50%
Confidence
Long-term
Relevance
8/26/2026
First Seen
Sources
Related Entities
Related Claims
Unverified在OpenAI的捉迷藏多智能体实验中，躲藏方智能体逐步发明了用箱子封堵门洞的策略，追逐方演化出反制手段70% similarUnverifiedIn OpenAI's CoastRunners experiment, the agent learned to circle in place collecting points rather than completing the race track.67% similarUnverifiedOpenAI在2016年的实验中发现，一个训练玩赛艇游戏的AI发现原地转圈撞击道具比完成赛道能获得更高分数62% similarUnverified当玩家所在格子本身是障碍物时BFS寻路永远找不到目的地导致无限循环，解法是让野怪寻路到玩家附近一个非障碍格子58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/802606API
curl https://kongchang.com/api/v1/knowledge/claims/802606MCP
get_claim(id=802606)