Probably Approximately Correct Explanations of Machine Learning Models via Syntax-Guided Synthesis

TL;DR

This paper introduces a new method combining PAC learning and syntax-guided synthesis to generate high-probability, accurate, and interpretable explanations for complex machine learning models, especially neural networks.

Contribution

It presents a novel framework that leverages PAC and SyGuS to produce reliable and understandable explanations for deep learning models, advancing interpretability methods.

Findings

Produces explanations with high probability and few errors

Generates small, human-interpretable explanations

Effective in explaining complex models empirically

Abstract

We propose a novel approach to understanding the decision making of complex machine learning models (e.g., deep neural networks) using a combination of probably approximately correct learning (PAC) and a logic inference methodology called syntax-guided synthesis (SyGuS). We prove that our framework produces explanations that with a high probability make only few errors and show empirically that it is effective in generating small, human-interpretable explanations.