EURO 2025 Leeds
Abstract Submission

EURO-Online login

2030. Explaining Machine Learning Models: A Multi-Objective Optimization Perspective

Invited abstract in session TB-24: Explaining Machine Learning Models: A Multi-Objective Optimization Perspective, stream Tutorials (and Workshops).

Tuesday, 10:30-12:00
Room: Esther Simpson 3.02

Authors (first author is the speaker)

1. Dolores Romero Morales
Copenhagen Business School

Abstract

State-of-the-art Artificial Intelligence (AI) and Machine Learning (ML) algorithms are ubiquitous. While they can achieve high accuracies, they are also criticized for not being transparent about how they arrive at their decisions, preventing their adoption in Data-Driven Decision-Making. Even when in place, they can be unfair to the citizen, and there are well-documented examples of discriminatory outcomes in high-stakes algorithmic decision-making.

In this tutorial, we aim to train ML models that are more transparent and less unfair, by striking a balance between accuracy, explainability, and unfairness. We will first navigate through Multi-Objective Optimization (MOO) models that train ML models with enhanced explainability and fairness. Then, we will move to models MOO models that derive local and global explanations to the predictions of an opaque model.

Keywords

Status: accepted

Back to the list of papers