Interpretation Support System for Classification Patterns in Deep learning

Abstract

This paper describes an interpretation support system for classification patterns based on the contents of learning results in deep learning with texts, and verified its effectiveness. It is well known that classification patterns by deep learning models are often difficult to interpret the reasons derived. The proposed system extracts the contents of learning results in deep learning with texts and provides seeds for interpretations of the patterns learned. Then, the system displays learned network structures so that anyone can easily understand learning results. In verification experiments to confirm the effectiveness of the system, based on the learning result of deep learning classifying sentences, test subjects were instructed to give meanings of classification patterns peculiar to each output. The results show that test subjects who represent novice data scientists could understand the meanings of the classification patterns of deep learning with texts.