Overview
The lab pursues research at the intersection of machine learning and interpretability, focusing on understanding the inner-workings of AI systems and their learning processes. sqIRL publishes research on topics including interpretable classifiers, capsule networks, and spiking neural networks.
In the news
- We are happy to welcome Achraf Yandouzi to the group. As you may see from the post below, Achraf profile speaks for itself. In the coming years, he will investigating how to design efficient models and learning processes. Lots of success on your doctoral trajectory, looking forward to celebrate your achievements and support you when needed. join us celebrating the beginning of this journey!
- Glad to have contributed to this year's #AIMLAI. If you are at #ECMLPKDD 2026, this Monday afternoon Peter Kirby will be presenting his work on HDC-based anomaly detectors. On Wednesday Arian Sabaghi will be presenting his work on weakly supervised object localization. Looking forward to seeing you there. #interpretability #AI #ML #XAI IDLab (UAntwerp - imec) #UAntwerp #imec
- Next week #sqIRL/IDLab (UAntwerp - imec) will be at #ECMLPKDD 2026. Some of the work we will presenting include: An evaluation pipeline for visual explanations for autonomous vessels. 📄 Structure-Aware Explanation Evaluation in Autonomous Maritime Navigation Arian Sabaghi, José Oramas M. #XKDD Workshop How to train anomaly detectors while keeping the computations low via HDC classifiers. 📄 Pruning Hyperdimensional Anomaly Detectors Peter Kirby, Werner Van Leekwijck, José Oramas M. #DLMNH Workshop 📄 Efficient Object Localization
- Check out our recent work on lightweight learning-based encoders for hyperdimensional computing classifiers. We show improvement over standard HDC methods across established benchmarks with no increase in model size or inference complexity. Moreover, we achieve comparable predictive performance w.r.t. existing encoder training approaches while following a demonstrably simpler algorithm. Congratulations Peter for this strong first step on your doctoral trajectory. Keep up the good work. #HDC #RepresentationLearning #VSA #AI #ML
- Another #sqIRL defending her PhD this year. Congratulations Saja on the successful defense of your PhD and for your contributions around the interpretability of Capsule Networks. Lots of success on what will follow. We are proud of you. #CapsNets #Interpretability #Explainabilty #CapsuleNetworks #DeepLearning #XAI IDLab (UAntwerp - imec) University of Antwerp - Department of Computer Science
Related profiles
Something wrong or missing? Send an update. Fixed within 24 hours.