PhD Researcher · Machine Learning & Human-AI Interaction
I'm a third-year PhD student at Hasselt University's Digital Future Lab, where I work at the intersection of machine learning and human-AI interaction under the supervision of Kris Luyten and Gustavo Rovelo Ruiz.
My research focuses on the Rashomon Effect: the fact that multiple machine learning models can perform similarly while behaving differently. I'm interested both in how we can efficiently discover these alternative models and in how their disagreements can be used to support human decision-making.
More broadly, I study how people make decisions with AI, and how interactive systems can help them reason about uncertainty, disagreement, and alternative model perspectives.
ICLR 2026
Gilles Eerlings · Brent Zoomers · Jori Liesenborgs · Gustavo Rovelo Ruiz · Kris Luyten
A gradient-free approach for efficiently exploring alternative, similarly performing neural networks and uncovering meaningful disagreement between them.
EICS 2024 Workshops
Gilles Eerlings · Sebe Vanbrabant · Jori Liesenborgs · Gustavo Rovelo Ruiz · Davy Vanacken · Kris Luyten
A visual dashboard for exploring agreement and disagreement across multiple similarly performing AI models, designed to help users compare model predictions without overwhelming them with information.
EICS 2025
Sebe Vanbrabant · Gilles Eerlings · Gustavo Rovelo Ruiz · Davy Vanacken
A conversational framework that combines large language models with specialized tools to let users explore and interrogate the behavior of AI models through natural language.
I presented DIVERSE as a poster in Rio de Janeiro and had many interesting discussions around model multiplicity and alternative model perspectives.
DIVERSE introduces a gradient-free approach for efficiently exploring similarly performing neural networks that disagree in meaningful ways.
I presented “AI Needs More Opinions,” a poster on the challenges and opportunities of model multiplicity in human-AI decision-making.
I presented the direction of my PhD research on model multiplicity and its role in human-AI decision-making.
I attended IUI 2025 in Cagliari, where I spent the week meeting researchers, discussing ideas, and getting to know the community.
I presented AI-Spectra, a visual dashboard for exploring agreement and disagreement across similarly performing AI models.