VeriX: Towards Verified Explainability of Deep Neural Networks

TL;DR

VeriX introduces a system for generating robust, optimal explanations and counterfactuals for deep neural networks, enhancing interpretability and decision boundary understanding through constraint solving and sensitivity heuristics.

Contribution

It presents a novel approach combining constraint solving and sensitivity ranking to produce verified explanations and counterfactuals for neural networks.

Findings

Effective on image recognition benchmarks

Applicable to autonomous aircraft taxiing scenarios

Produces robust, optimal explanations

Abstract

We present VeriX (Verified eXplainability), a system for producing optimal robust explanations and generating counterfactuals along decision boundaries of machine learning models. We build such explanations and counterfactuals iteratively using constraint solving techniques and a heuristic based on feature-level sensitivity ranking. We evaluate our method on image recognition benchmarks and a real-world scenario of autonomous aircraft taxiing.