A simple bot to play go written in Kotlin inspired by the Manning book "Deep Learning and the Game of Go"
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
src
.gitignore
README.md
build.gradle
gradlew
gradlew.bat
settings.gradle
start-gtp.sh
start-human.sh

README.md

kakomu

A simple bot able to play game of go in Kotlin.

The general approach is inspired by the book "Deep Learning And The Game of Go". I introduced some optimizations and Kotlin idioms.

Currently it's using MonteCarlo Tree Search but hopefully it will also make use of DeepLearning.

The goals of Kakomu are:

  • Enjoy and having fun

  • Be able to experiment with MCTS, DeepLearning, etc

  • Keep code simple and easy to understand and modify

  • Get some hints on how to improve at Go

Not a goal:

  • Make a top level AI player

  • Create a framework for Go Game IA engine

BACKLOG (sort of)

SGF files load and save

GTF support to use a nicer GUI

DeepLearning training to guess best moves (esp. fuseki)

Multi threads support