Unverified50% confidenceFactExact time

DeepMind的AlphaZero于2017年发布，仅凭自我对弈就在国际象棋、将棋和围棋上达到超越人类顶尖水平

1

Sources

50%

Confidence

Long-term

Relevance

8/7/2026

First Seen

Sources

nanoAlphaZero：单文件JAX实现24小时训练2700分棋力

Related Claims

Unverified2016年DeepMind的AlphaGo通过与人类棋谱学习加自我对弈击败了围棋世界冠军78% similar VerifiedDeepMind的AlphaGo通过自我对弈数百万局来习得围棋策略76% similar VerifiedDeepMind的AlphaStar使用人类对局数据进行行为克隆初始化，再通过自我对弈强化学习达到职业选手水平70% similar UnverifiedGoogle DeepMind的AlphaCode系统在编程竞赛中达到了中等水平70% similar UnverifiedDeepMind的AlphaGo击败了李世石70% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/700726

API

curl https://kongchang.com/api/v1/knowledge/claims/700726

MCP

get_claim(id=700726)