A Proposal of Visualization Method for Critical Area in Computer Go

Deep Learning for the game of Go recently had a tremendous success with the victory of AlphaGo against Ke Jie in May 2017. However, there is no clear understanding of why they perform so well. In this paper, we introduce a visualization technique that performs a sensitivity analysis of the classifier output by occluding portions of the input Go board, revealing which parts of the board are important for predicting the next move. Using this tool, we start with the experiment about the accuracy of the critical area revealed. We also suppose that by showing the critical area, it will allow Go beginners to understand the board visually that they may have been confused about.