Data Visualization for gaming agent of Go using deep reinforcement learning

Zhongmou Cai zcai34@ucsc.edu

Fall 2020, CSE 261 Advanced Data Visualization
University of California, Santa Cruz

Abstract

In this project, I will implement an artificial intelligent agent to play the Go game, using the deep reinforcement learning algorithm. There will be a corresponding search tree(game tree)you, where every node has an expected utility. Also I will visualize the structure graph of the deep learning model.

Resources

View Paper

Project Images

visualization of searching tree

visualization of model structure

the game board

Project Goal:
visualize the searching process of a game agent
visualize a basic deep neural network using tensorflow
help people understand more about reinforcement learning and gaming AI