A Model-Agnostic SAT-based Approach for Symbolic Explanation Enumeration

TL;DR

This paper introduces a model-agnostic SAT-based method to generate symbolic explanations, such as sufficient reasons and counterfactuals, for individual predictions, demonstrated on image classification tasks.

Contribution

It presents a novel, generic SAT-based framework for generating diverse symbolic explanations for any predictive model, enhancing interpretability.

Findings

Feasibility demonstrated on image classification

Effective generation of sufficient reasons

Successful creation of counterfactual explanations

Abstract

In this paper titled A Model-Agnostic SAT-based approach for Symbolic Explanation Enumeration we propose a generic agnostic approach allowing to generate different and complementary types of symbolic explanations. More precisely, we generate explanations to locally explain a single prediction by analyzing the relationship between the features and the output. Our approach uses a propositional encoding of the predictive model and a SAT-based setting to generate two types of symbolic explanations which are Sufficient Reasons and Counterfactuals. The experimental results on image classification task show the feasibility of the proposed approach and its effectiveness in providing Sufficient Reasons and Counterfactuals explanations.