Self-explaining deep models with logic rule reasoning

TL;DR

The paper introduces SELOR, a framework that integrates logic rule explanations into deep models, achieving high predictive accuracy and explanations aligned with human reasoning without needing predefined rules or annotations.

Contribution

SELOR enables deep models to predict and explain using logic rules in a differentiable manner, improving human-aligned explanations without sacrificing performance.

Findings

Provides explanations closer to human logic than existing methods

Maintains high prediction accuracy comparable to standard deep models

Does not require predefined rules or human annotations

Abstract

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 human precision 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 deep learning models.