BriefGPT.xyz
Sep, 2018
SAI，一款玩围棋的合理人工智能
SAI, a Sensible Artificial Intelligence that plays Go
HTML
PDF
Francesco Morandin, Gianluca Amato, Rosa Gini, Carlo Metta, Maurizio Parton...
TL;DR
提出了一种基于多贴子调整的AlphaGo Zero/Leela Zero改进方法，采用双参数S型函数建立了贴子与胜率之间的模型，并通过自我博弈训练获得了在7x7围棋中表现出色的智能体。通过该模型的S型参数可以估计棋盘上的得分差异和游戏结果。
Abstract
We propose a
multiple-komi modification
of the
alphago zero
/
leela zero
paradigm. The winrate as a function of the komi is modeled with a t
→