Margot Herin, currently a postdoctoral researcher at the Instituto de Telecomunicações in Lisbon, has been named the winner of the 2026 Cor Baayen Early Career Researcher Award. The annual ERCIM award recognises outstanding early-career researchers in computer science and applied mathematics.
Margot Herin works at the intersection of machine learning and decision theory, with the aim of developing artificial intelligence systems that are both more transparent and more rational. Her research focuses in particular on preference learning: methods that learn how decisions or choices are made while preserving desirable rationality properties and offering interpretable models whose parameters and computations have a clear meaning.
Herin received her PhD in 2025 from Sorbonne University in Paris for a thesis entitled Learning preference models: a marriage between decision theory and machine learning, supervised by Patrice Perny and Nataliya Sokolovska. The thesis combines mathematical foundations with extensive experimental validation on real and synthetic data. It has already received significant recognition, including the Gilles Kahn 2025 PhD Award of the French Computer Science Society, under the patronage of the French Academy of Sciences, and the 2026 PhD Award of the CNRS Operations Research and Decision research group.
Her work also demonstrates the potential of preference-learning methods in applications beyond core AI research. In an interdisciplinary project involving researchers in law and computer science, the methods were applied to predicting compensatory damages in divorce cases. She has also collaborated on uncertainty quantification and sensitivity analysis, developing new indices for identifying influential variables in numerical models of physical phenomena, which have found applications, for instance, in environmental monitoring.
More recently, her research has expanded towards large language models, including methods for aligning LLMs with diverse preferences and optimisation techniques for steering their generation processes.
Herin's work has appeared at leading international AI conferences including IJCAI, AAAI and UAI, as well as in journals such as Theory and Decision, Annals of Mathematics and Artificial Intelligence and SIAM/ASA Journal on Uncertainty Quantification. She received an IJCAI 2024 Distinguished Paper Award.
The 2026 selection committee unanimously selected Margot Herin from seven finalists. She was nominated by Jean-Frédéric Gerbeau of Inria, France.
The Cor Baayen Early Career Researcher Award was established in 1995 and is named after ERCIM’s first President. It recognises outstanding scientific promise and achievement in informatics and applied mathematics and carries a prize of €5,000.
2026 Cor Baayen Early Career Researcher Award
Winner:
Margot Herin, nominated by Jean-Frédéric Gerbeau, Inria, France
Finalists:
• Adrián Segura Ortiz, nominated by University of Málaga, Spain
• Giovanni Puccetti, nominated by CNR, Italy
• Betül Bayrak Çik, nominated by NTNU, Norway
• Delia Rico Marchena, nominated by University of Málaga, Spain
• Marco De Vincenzi, nominated by CNR, Italy
• Corentin Bisot, nominated by Inria, France
Selection Committee:
The Selection Committee was composed of Gabriel David (INESC-TEC), Monica Divitini (NTNU – chair of the ERCIM Human Capital Task Group), Georgia Kapitsaki (University of Cyprus), Bruno Lévy (Inria), Theodore Patkos (ICS-FORTH), Ernesto Pimentel (University of Malaga), and Fabrizio Sebastiani (ISTI-CNR).
The decision was unanimous.
More information about the ERCIM Cor Baayen Early Career Researcher Award:
https://www.ercim.eu/people-careers/cor-baayen-award