Exploiting auto-encoders and segmentation methods for middle-level explanations of image classification systems

TL;DR

This paper introduces an XAI framework that uses auto-encoders to extract middle-level features for providing multiple, human-understandable explanations of image classification decisions, enhancing interpretability.

Contribution

It proposes a novel method leveraging auto-encoders to generate diverse, middle-level explanations for image classifiers, addressing the need for explanations aligned with human understanding.

Findings

Encouraging results on two image datasets demonstrate the effectiveness of the approach.

Different auto-encoder types can produce varied explanations for the same model behavior.

The method shows potential for application beyond image data to other domains.

Abstract

A central issue addressed by the rapidly growing research area of eXplainable Artificial Intelligence (XAI) is to provide methods to give explanations for the behaviours of Machine Learning (ML) non-interpretable models after the training. Recently, it is becoming more and more evident that new directions to create better explanations should take into account what a good explanation is to a human user. This paper suggests taking advantage of developing an XAI framework that allows producing multiple explanations for the response of image a classification system in terms of potentially different middle-level input features. To this end, we propose an XAI framework able to construct explanations in terms of input features extracted by auto-encoders. We start from the hypothesis that some autoencoders, relying on standard data representation approaches, could extract more salient and understandable input properties, which we call here \textit{Middle-Level input Features} (MLFs), for a user with respect to raw low-level features. Furthermore, extracting different types of MLFs through different type of autoencoders, different types of explanations for the same ML system behaviour can be returned. We experimentally tested our method on two different image datasets and using three different types of MLFs. The results are encouraging. Although our novel approach was tested in the context of image classification, it can potentially be used on other data types to the extent that auto-encoders to extract humanly understandable representations can be applied.