Marc Mézard Awarded the 2026 Dirac Medal
The Abdus Salam International Centre for Theoretical Physics (ICTP) in Trieste has awarded the 2026 Dirac Medal to Marc Mézard, professor of theoretical physics in the Department of Computing Sciences at Bocconi University, along with Bernard Derrida, Deepak Dhar, and Haim Sompolinsky. The four scholars are being honored “their pioneering contributions to equilibrium statistical mechanics and for extending its concepts and methods into non-equilibrium statistical mechanics, optimization problems, theoretical neuroscience, and, finally, artificial intelligence”
Established in 1985, the Dirac Medal recognizes particularly significant contributions to theoretical physics. Among its recipients are scientists who have subsequently been awarded the Nobel Prize or the Fields Medal. The 2026 award recognizes the value of an approach capable of connecting seemingly disparate fields: understanding the collective behavior of a material’s components can also provide tools for tackling a computational problem or studying how a neural network learns.
For Mézard, the award specifically recognizes his development of the cavity method—in collaboration with Giorgio Parisi and Miguel Virasoro—and his ability to apply ideas derived from the study of disordered systems to problems in computer science, information processing, and learning in neural networks.
This is a journey in which Mézard has played a leading role. At the heart of his research are disordered systems, in which numerous interactions impose competing constraints that cannot all be satisfied simultaneously. The cavity method devised by Mézard, Parisi, and Virasoro makes it possible to study how the behavior of a single element relates to that of the whole and has provided the foundation for algorithms capable of identifying favorable configurations among a vast number of possibilities.
This led to the transition from physics to computer science. Finding a good solution among many alternatives, while respecting constraints that are often difficult to reconcile, is a challenge common to numerous optimization problems. Mézard has helped demonstrate how to tackle this challenge using ideas derived from the study of disordered materials. These tools have also found application in the reliable transmission of information and, more recently, in artificial intelligence. Today, his research focuses on learning in neural networks and generative models.
A member of the Accademia dei Lincei, Mézard joined Bocconi in 2022, after serving as director of the École normale supérieure in Paris since 2012. He has been a member of the ICTP Scientific Council since 2021 and has chaired it since 2024. The Dirac Medal joins other honors such as the Lars Onsager Prize from the American Physical Society and the Three Physicists Prize from the École normale supérieure, confirming the international significance of research that continues to bridge the fields of physics, information, and learning.