To recognize the quality of PhD graduates in Artificial Intelligence from a Belgian, Dutch, and Luxembourgish universities, the BNVKI has set up the BNVKI PhD Dissertation Award. This award highlights groundbreaking methodological advancements and major impact in AI resulting from recent PhD graduates. Our jury members rank candidates’ dissertation based on originality, significance, potential impact on theory and/or practice and quality of presentation.
We are pleased to announce the results of this year’s BNVKI Artificial Intelligence PhD Dissertation Award.
Winner: Jaap Jumelet (University of Amsterdam)
Jaap’s thesis addresses a fundamental question in AI research: do large language models possess a deep understanding of grammatical structure comparable to that of humans? Situated at the intersection of natural language processing, linguistics, and interpretability, his work introduces novel interpretability techniques that shed new light on the inner workings of large-scale language models.
The jury was unanimous in its excellence across all four evaluation criteria: originality, significance, potential, and presentation. Jury members described the work as very impressive, noted Jaap as a strong EurAI PhD Award candidate, and highlighted the combination of intellectual ambition and rigorous execution. The research was praised for its high impact and for being both highly interesting and very well conducted. Moreover, the jury members complimented the candidate on the strong methodological contributions made in this thesis.
Runner-up: Kaizheng Wang (KU Leuven)
Kaizheng’s dissertation tackles a critical challenge in modern machine learning: how to robustly represent and quantify uncertainty in the predictions of deep neural networks. His work proposes novel deep credal neural networks that produce credal sets defined by upper and lower probability bounds, enabling a principled and scalable approach to epistemic uncertainty quantification in classification tasks.
The jury praised the excellence of his work, with particular recognition for its significance and potential impact on the field of reliable and robust machine learning.
Congratulations to both Jaap and Kaizheng on this well-deserved recognition! We would like to extend our sincere congratulations to all candidates who submitted their dissertation this year. The jury noted that the overall quality of this year’s batch was exceptionally high, making the selection process both challenging and rewarding. We see this as a strong reflection of the vibrant PhD research being conducted across the BeNeLux AI community.
The jury consists of the following members: Cédrick Fairon (UC Louvain), Lynn Houthuys (VUB), Annette ten Teije (Vrije Univ Amsterdam), Sicco Verwer (TU Delft), Shihan Wang (Utrecht University)
