معرفی
Tom Heskes is a Full Professor of Artificial Intelligence at the Institute for Computing and Information Sciences (iCIS) within Radboud University's Faculty of Science, where he also serves as Vice Dean. His research group, part of the Nijmegen ELLIS unit, focuses on foundational machine learning and cross-disciplinary applications in healthcare, energy, and physical sciences.
His educational background includes a PhD from Radboud University (1993) and postdoctoral work at the University of Illinois' Beckman Institute.
- PhD, Radboud University Nijmegen (1989-1993)
- Postdoc, Beckman Institute, University of Illinois (1993-1994)
Heskes specializes in probabilistic machine learning, with core interests in causal discovery, uncertainty quantification, and deep learning interpretability. His group develops statistical frameworks for learning from complex data, emphasizing robustness in biological, medical, and industrial contexts. Recent work integrates graph neural networks with power systems engineering to enhance grid reliability.
His publication trends reveal a dual focus: advancing unsupervised learning methodologies (e.g., anomaly detection robustness) while deploying AI in critical infrastructure like energy grids and healthcare monitoring. This reflects his commitment to both theoretical rigor and societal impact.
Notable scientific recognition includes:
- Vici grant from NWO (2006)
- TOP grant from NWO (2014)
As an advisor, Heskes mentors PhD candidates including Luc Evers (Parkinson's disease monitoring). His extensive grant portfolio spans NWO programs (NWA OBSerVeD, CORTEX), EU initiatives (MATRICS), and industry collaborations like Alliander's AI for Energy Grids Lab. He also initiated Radboud AI and co-founded Machine2Learn (acquired by AMD in 2024).
Heskes co-directs the AI for Parkinson lab and contributes to iCIS's leadership as former Research Director (2009-2014). His group actively participates in ELLIS networks and interdisciplinary institutes like Radboudumc, driving AI innovation across scientific domains.
