// ♞ Reinforcement Learning

AlphaGo / AlphaZero (Self-Play + MCTS)

Neural evaluation plus tree search, learned from scratch by self-play.

AlphaGo / AlphaZero（自对弈 + 蒙特卡洛树搜索） 2016 Advanced core

IN PLAIN WORDS

A policy net proposes moves, a value net evaluates, MCTS searches with both, and self-play data retrains the nets. AlphaZero surpassed humans in Go, chess and shogi without human games.

Used in: Game AIs: Leela Chess Zero, KataGo

TIMELINE · prerequisites → this → what builds on it

1992 Policy Gradie…1998 Convolutional…2016 AlphaGo / AlphaZe…2019 MuZero

01Original paper

Mastering the game of Go with deep neural networks and tree search

Silver et al. · Nature · 2016

DOI ↗ ✓ Verified · Crossref API doi:10.1038/nature16961

Note: AlphaGo Zero：DOI 10.1038/nature24270；AlphaZero：arXiv:1712.01815

02Where it sits

Learn the left column first; the right column builds on this one.

Learn first

AlphaGo / AlphaZero (Self-Play + MCTS)

Then you can learn

MuZero

Left: prerequisites · right: builds on this · dashed: related. Full map →

03Learning resources

Official docs, explainers, courses and reference code. All links checked.

Explainers

Google DeepMind：AlphaZero ↗EN

Courses

huggingface.co ↗EN

04Where the course teaches it

U33 Reinforcement Learning Basics MDPs, Q-learning, DQN, policy gradients, PPO and AlphaZero — the prerequisites for post-training and reasoning models. 14 h12 algorithms4 papers

For AI / data

Verification: verified, checked 2026-10-21.
JSON: /api/learn/algorithms.json (id: mcts-alphazero)