Machine Learning playing GO

A picture of two persons playing GO — ©TUM-DI-LAB. M.Sc. Bernhard Werner (right) playing GO. Everyone interested is invited to play every Thursday at 2 PM in the building of Mathematics.

Sponsored by: Lehrstuhl für Geometrie und Visualisierung
Faculty Lead: M.Sc. Bernhard Werner
Project Lead: Dr. Ricardo Acevedo Cabra
Semester: Winter 2017/2018

Overview

The game Go is perhaps one of the most challenging of classical games for artificial intelligence due to its huge search space and the difficulty of evaluating moves and board positions. The goal of this project is to develop from scratch a learning machine, e.g., based on a neural network architecture, that plays at best Go on a reduced size board. The network would be trained with the results of gaming against minimax algorithms of different moves tree depths. For detailed info contact M.Sc. Bernhard Werner

Results

The results of this project were summarized in a final presentation and explained in detail in the final documentations of GO group 1 and GO group 2