Unverified60% confidenceFactExact time
DeepMind employed League Training in AlphaStar, maintaining an opponent pool of historical versions at different skill levels and styles to avoid overfitting to a single strategy.
2
Sources
60%
Confidence
Long-term
Relevance
8/7/2026
First Seen
Sources
M.A.R.A：用强化学习从零训练会战斗的AI坦克
Related Claims
UnverifiedDeepMind used a large-scale distributed training cluster for StarCraft II because a single machine could not complete training within an acceptable timeframe.75% similarVerifiedDeepMind's AlphaStar uses human match data for behavior cloning initialization, then achieves professional player level through self-play reinforcement learning.74% similarUnverifiedSome of DeepSeek's models restrict using their outputs to train competing models.71% similarUnverifiedAlphaGo是DeepMind强化学习路径的标志性成果，但其泛化能力高度受限于训练环境70% similarUnverifiedTeams including DeepMind, Tesla Optimus, and Unitree have all adopted reinforcement learning-driven motion control solutions.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/699283API
curl https://kongchang.com/api/v1/knowledge/claims/699283MCP
get_claim(id=699283)