[KongchangAI]
Unverified50% confidenceFactExact time

2017年的AlphaZero完全抛弃人类棋谱，仅通过自我对弈在数小时内超越前辈

1
Sources
50%
Confidence
Long-term
Relevance
7/12/2026
First Seen

Sources

AI如何攻克不完全信息博弈？Tactico实战全解析
redditr/reinforcementlearning7/11/2026

Related Claims

VerifiedAlphaGo Zero完全通过自我对弈超越了所有人类棋手68% similarUnverifiedOpenManus在执行马里奥风格跳跃游戏生成任务时最终显示为失败，生成的游戏角色只能上下移动，不能跳跃，功能完成度很低51% similar

Cite This Claim

Stable URI
https://kongchang.com/claim/491873
API
curl https://kongchang.com/api/v1/knowledge/claims/491873
MCP
get_claim(id=491873)