Teaching as Training: Incremental and Iterative AI Skill Development

We presented our contribution “Teaching as Training: Iterative and Incremental AI Skill Development” () at the EURECA-PRO Education & Research Days in Hasselt, held under the theme Glocalising Universities: A Shifting Horizon. This is joint work with Jolien Notermans (Department of Educational Development, Policy and Quality Assurance) and Sarah Doumen (Faculty of Sciences) at Hasselt University. More details on the publication page. The visual story is generated using StoryBookly.

Read more →

DIVERSE: Finding the Many Faces of AI Decision-Making

Our paper “DIVERSE: Disagreement-Inducing Vector Evolution for Rashomon Set Exploration” () has been accepted at ICLR 2026, one of the top venues for machine learning research. This is joint work with my PhD student Gilles Eerlings, Brent Zoomers, Jori Liesenborgs, and Gustavo Rovelo Ruiz at the Digital Future Lab. More details on the publication page.

Read more →

Industrial assembly is shifting toward human-robot collaboration (HRC) to leverage the complementary strengths of both agents. However, traditional task allocation referred to as the Robotic Assembly Line Balancing Problem (RALBP) remains labor-intensive and often lacks transparency. We introduce DELEGACT, a framework designed to produce workable, intelligible human-cobot task allocations. The framework uses a Vision-Language Model (VLM) to extract atomic operations from expert demonstration videos, then employs a Large Language Model (LLM) to delegate these tasks based on robot specifications, operator competencies, and material definitions. We provide a proof-of-concept prototype and preliminary testing on illustrative cases. Results demonstrate the system's ability to reason about complex constraints such as precision, weight, and ergonomics. This paper illustrates how off-the-shelf foundation models can automate HRC decision-making via a human-in-the-loop paradigm while preserving operator agency and understanding.

Read more →

We propose DIVERSE, a framework for systematically exploring the Rashomon set of deep neural networks, the collection of models that match a reference model's accuracy while differing in their predictive behavior. DIVERSE augments a pretrained model with Feature-wise Linear Modulation (FiLM) layers and uses Covariance Matrix Adaptation Evolution Strategy (CMA-ES) to search a latent modulation space, generating diverse model variants without retraining or gradient access. Across MNIST, PneumoniaMNIST, and CIFAR-10, DIVERSE uncovers multiple high-performing yet functionally distinct models. Our experiments show that DIVERSE offers a competitive and efficient exploration of the Rashomon set, making it feasible to construct diverse sets that maintain robustness and performance while supporting well-balanced model multiplicity. While retraining remains the baseline for generating Rashomon sets, DIVERSE achieves comparable diversity at reduced computational cost.

Read more →

The SAI Update magazine (Nov 2025 , sia.be) selected two projects from our Human–AI Interaction (HAII) course for its Next Technology Generation special. Proud of our students Linsey Helsen and Xander Vervaecke who turned their Human-AI Interaction project ideas into concrete, useful systems.

1) A Multi-Agent Approach to Fact-Checking (, ) — Xander Vervaecke (UHasselt) Xander’s LieSpy.ai coordinates multiple LLMs (e.g., GPT, Gemini, Mistral) to verify claims, compare reasoning, and aggregate evidence into a transparent verdict. The interface exposes sources, trust scores, and model rationales, moving fact-checking beyond a single-model answer. Key ideas: multi-agent collaboration, cross-validation, explainability.

Read more →

Read more →

I am pleased to share the results of individual student projects from our Human-AI Interaction (HAII) course at Hasselt University. The YouTube playlist showcases some of the work our students have done throughout the course.

Read more →

I wanted to show my students appropriate ways of using LLMs for and during coding, so I started building (with some LLM help) a Slidev component, LLMQuery.vue, that adds LLM interactions to slides. It feels important to actively show students how these tools can amplify human knowledge and skill building rather than replace it altogether, even if I’m far from an expert. So with a bit of LLM help , I put together a sli.dev component in Vue that integrates LLMQuery right into my Slidev presentation. Maybe it’s useful for others too, so I’m sharing it here for download and further tinkering—people who are much better at web dev (there are many!) can probably turn it into something truly polished.

Read more →

The official launch of the Digital Future Lab (DFL) marks an exciting step forward for Hasselt University and for the ecosystem of digital innovation in Flanders. With over 80 researchers across various interdisciplinary groups, DFL focuses on creating well-designed, human-centered, trustworthy, and useful digital systems that address both industrial and societal challenges. We did an interview (in Dutch, with Ann T’Syen) that can be found here.

Read more →

Effective operation in direct-control telerobotics relies heavily on real-time communication between the operator and the robot, as the operator retains full control over the robot's actions. However, in scenarios involving long distances, communication delays disrupt this feedback loop, creating significant challenges for precise control. To investigate these challenges, we conducted a user study where participants operated a TurtleBot3 Waffle Pi under varying delay conditions. Post-experiment brainstorming and analysis revealed recurring challenges, including over-correction, unpredictable robot behavior, and reduced situational awareness. Potential solutions identified include improving robot behavior predictability, integrating feedforward mechanisms, and enhancing visual feedback. These findings underscore the importance of designing intelligent interfaces to mitigate the impact of delays on telerobotic performance.

Read more →