Why Should I Choose You? AutoXAI: A Framework for Selecting and Tuning eXplainable AI Solutions

TL;DR

AutoXAI is a framework that automates the selection and tuning of the most suitable XAI solutions and hyperparameters based on user-specific needs, constraints, and evaluation metrics, simplifying the complex choice process.

Contribution

This paper introduces AutoXAI, a novel framework that integrates context-aware recommendation and AutoML strategies to automate XAI solution selection and hyperparameter tuning.

Findings

AutoXAI effectively recommends XAI solutions tailored to user needs.

AutoXAI identifies optimal hyperparameters matching user constraints.

Application to use cases demonstrates improved relevance of selected XAI solutions.

Abstract

In recent years, a large number of XAI (eXplainable Artificial Intelligence) solutions have been proposed to explain existing ML (Machine Learning) models or to create interpretable ML models. Evaluation measures have recently been proposed and it is now possible to compare these XAI solutions. However, selecting the most relevant XAI solution among all this diversity is still a tedious task, especially when meeting specific needs and constraints. In this paper, we propose AutoXAI, a framework that recommends the best XAI solution and its hyperparameters according to specific XAI evaluation metrics while considering the user's context (dataset, ML model, XAI needs and constraints). It adapts approaches from context-aware recommender systems and strategies of optimization and evaluation from AutoML (Automated Machine Learning). We apply AutoXAI to two use cases, and show that it recommends XAI solutions adapted to the user's needs with the best hyperparameters matching the user's constraints.