Timezone: »

 
Poster
Self-explaining deep models with logic rule reasoning
Seungeon Lee · Xiting Wang · Sungwon Han · Xiaoyuan Yi · Xing Xie · Meeyoung Cha

Tue Nov 29 09:00 AM -- 11:00 AM (PST) @ Hall J #108
in Poster Session 1 »

We present SELOR, a framework for integrating self-explaining capabilities into a given deep model to achieve both high prediction performance and human precision. By “human precision”, we refer to the degree to which humans agree with the reasons models provide for their predictions. Human precision affects user trust and allows users to collaborate closely with the model. We demonstrate that logic rule explanations naturally satisfy them with the expressive power required for good predictive performance. We then illustrate how to enable a deep model to predict and explain with logic rules. Our method does not require predefined logic rule sets or human annotations and can be learned efficiently and easily with widely-used deep learning modules in a differentiable way. Extensive experiments show that our method gives explanations closer to human decision logic than other methods while maintaining the performance of the deep learning model.

Keywords: Deep Learning · Reasoning · interpretability · Self-explaining
[ Paper]  [ Slides]  [ Poster]  [ OpenReview] 

Author Information

Seungeon Lee (Korea Advanced Institute of Science & Technology)
Xiting Wang (Microsoft Research Asia)
Sungwon Han (KAIST)
Xiaoyuan Yi (Microsoft)
Xing Xie (Microsoft Research Asia)
Meeyoung Cha (KAIST)

More from the Same Authors