Scholar
Thomas Demeester
Google Scholar ID: x_0MCH4AAAAJ
Associate professor, Ghent University - imec
Citations & Impact
Citations
3,971
H-index
27
i10-index
57
Publications
20
Co-authors
52
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Published multiple papers, including works on Hopfield Networks and Deep Equilibrium Models, clinical reasoning, causal inference, and the inferential utility of synthetic medical data. Developed BioLORD 2023 (biomedical sentence encoder) and transtokenization method.
Research Experience
- Post-doc involved in Information Retrieval research, collaborating with the Database Group at the University of Twente, Netherlands. Gradually shifted to Natural Language Processing and machine learning. Conducted a research stay at University College London in 2016, working with Prof. Sebastian Riedel. Now an associate professor at Ghent University-imec, leading an AI research team focused on health applications.
Education
- Received M.Sc. in Electrical Engineering from Ghent University in 2005, after completing his final year and master thesis at ETH Zurich, Switzerland. Obtained Ph.D. from the Department of Information Technology at Ghent University in 2009, under the supervision of Prof. Daniel De Zutter, focusing on computational electromagnetics.
Background
- Research interests: Natural Language Processing, Deep Learning, and AI. Brief bio: Currently an associate professor at the Internet Technology and Data Science Lab (IDLab), Ghent University - imec, Belgium. Main research areas include energy-based models, neuro-symbolic AI, and drug design.
Miscellany
- Collaborated with Luca Ambrogioni from Radboud University on negative guidance strategies for diffusion models; worked with Stijn Vansteelandt and the Syndara team at UZ Ghent hospital on the application of synthetic medical data.
