FairXCluster

Welcome to FairXCluster Project

Understanding the reasoning behind the decisions made by learned models is a central quest in AI . This is achieved by either using interpretable and transparent models, or by generating post-hoc explanations for opaque models. Furthermore, as AI-powered systems are increasingly being used to assist or even replace human decision making, important concerns regarding the trustworthiness and ethical standing of such systems are raised corroborated by several cases where such systems did misbehave. Algorithmic fairness attempts to conceptualize and operationalize notions of fairness through formal technical definitions, metrics, and mitigation methods.

We will address these challenges through a novel counterfactual approach. Counterfactual explanations are a specific category of explanations that consider what would have happened had the input to a model been changed. Most previous research has applied counterfactual explanations in prediction tasks. In simple terms, in this context, counterfactuals seek for the smallest changes in a data instance that would result in a different prediction. To the best of our knowledge, counterfactual explanations have not yet been used in clustering.