PDF] ELF OpenGo: An Analysis and Open Reimplementation of AlphaZero

ELF OpenGo is the first open-source Go AI to convincingly demonstrate superhuman performance with a perfect (20:0) record against global top professionals and is proposed, anopen-source reimplementation of the AlphaZero algorithm. The AlphaGo, AlphaGo Zero, and AlphaZero series of algorithms are remarkable demonstrations of deep reinforcement learning's capabilities, achieving superhuman performance in the complex game of Go with progressively increasing autonomy. However, many obstacles remain in the understanding of and usability of these promising approaches by the research community. Toward elucidating unresolved mysteries and facilitating future research, we propose ELF OpenGo, an open-source reimplementation of the AlphaZero algorithm. ELF OpenGo is the first open-source Go AI to convincingly demonstrate superhuman performance with a perfect (20:0) record against global top professionals. We apply ELF OpenGo to conduct extensive ablation studies, and to identify and analyze numerous interesting phenomena in both the model training and in the gameplay inference procedures. Our code, models, selfplay datasets, and auxiliary data are publicly available.

Whatever Happened to the Logic of Discovery? From Transparent Logic to Alien Reasoning

Electronics, Free Full-Text

summary/README.md at master · gopala-kr/summary · GitHub

Electronics, Free Full-Text

LightZero/README.zh.md at main · opendilab/LightZero · GitHub

Conclusion and Bibliography for “Understanding the diffusion of large language models” — Rethink Priorities

PDF] Mobile Networks for Computer Go

PDF] Accelerating and Improving AlphaZero Using Population Based Training

PDF) Alpha-T: Learning to Traverse over Graphs with An AlphaZero-inspired Self-Play Framework

PDF] ELF OpenGo: An Analysis and Open Reimplementation of AlphaZero

Multiplayer AlphaZero – arXiv Vanity

PDF] ELF OpenGo: An Analysis and Open Reimplementation of AlphaZero

Spatial state-action features for general games - ScienceDirect

PDF] Mobile Networks for Computer Go