This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| David Mezard | |
|---|---|
| Name | David Mezard |
| Nationality | French |
| Fields | Statistical physics; Statistical learning; Machine learning; Computer science |
| Workplaces | École Normale Supérieure; École Polytechnique; CNRS; University of Paris; University of Cambridge |
| Alma mater | École Normale Supérieure; Université Paris-Sud |
| Doctoral advisor | Marc Mézard |
| Known for | Spin glass theory; Message-passing algorithms; Compressed sensing; Inference |
David Mezard is a French theoretical physicist and computer scientist known for contributions at the interface of statistical physics, information theory, and computer science. He has worked on spin glass models, message-passing algorithms, and the statistical foundations of inference and learning, influencing research in machine learning, signal processing, and combinatorial optimization. Mezard's work bridges traditions from the École Normale Supérieure, collaborations with researchers at institutions such as CNRS, École Polytechnique, and interactions with communities around the University of Cambridge, Princeton University, and Bell Labs.
Mezard completed his formative studies at the École Normale Supérieure and pursued doctoral research at Université Paris-Sud under supervision aligned with senior researchers associated with Institut des Hautes Études Scientifiques and CNRS. During his education he engaged with research groups connected to centers such as Laboratoire de Physique Théorique and worked alongside scholars linked to Yves Laszlo, Giorgio Parisi, and figures from the lineage of Philippe Nozières and Pierre-Gilles de Gennes. Early influences included developments originating from the Ising model, the Sherrington–Kirkpatrick model, and foundational results in spin glass theory by researchers at Université Pierre et Marie Curie and Sapienza University of Rome.
Mezard has held appointments at major French institutions including École Normale Supérieure, École Polytechnique, and research positions within CNRS and the broader Centre national de la recherche scientifique network. He has been affiliated with international centers like the Institut des Hautes Études Scientifiques, visited groups at Massachusetts Institute of Technology, Stanford University, and collaborated with faculty from Université Paris-Saclay and University of Cambridge. His academic roles have combined teaching responsibilities alongside leadership in laboratories connected to Laboratoire de Physique Statistique, contributions to doctoral supervision tied to Thèse de Doctorat programs, and participation in program committees for conferences organized by groups such as NeurIPS, COLT, ICML, and IEEE symposia.
Mezard's research spans theoretical and algorithmic themes linking statistical mechanics models to computational tasks in information theory and computer science. He contributed to understanding the statistical physics of the Sherrington–Kirkpatrick model and the development of the replica method popularized in work by Giorgio Parisi and applied in analyses influenced by Marc Mézard and collaborators from CEA Saclay. Mezard was instrumental in formulating message-passing frameworks such as belief propagation and survey propagation that connect to algorithms used in error-correcting codes, low-density parity-check codes, and problems studied by researchers at Bell Labs and University of Illinois Urbana-Champaign. His work on compressed sensing intersected with advances by teams at Caltech, ETH Zurich, and University of Minnesota, providing rigorous and heuristic analyses of reconstruction thresholds and phase transitions paralleling phenomena in the random graph theory community associated with Erdős–Rényi ensembles.
He advanced the statistical treatment of inference problems, linking phase diagrams familiar from spin glass theory to algorithmic hardness results relevant for satisfiability problem instances scrutinized in research at Princeton University and University of California, Berkeley. Mezard's collaborations with scholars from INRIA, CentraleSupélec, Columbia University, and Tel Aviv University produced cross-disciplinary contributions to Bayesian inference, sparse estimation, and the theoretical underpinnings of modern machine learning techniques. His perspectives influenced algorithmic design in domains touching computer vision, bioinformatics, and neuroscience laboratories at Max Planck Society and CNRS institutes.
Mezard's work has been recognized by awards and distinctions from French and international scientific bodies including honors associated with CNRS, elective roles in academies such as Académie des sciences, invitations to deliver keynote lectures at meetings organized by American Physical Society, and fellowships or visiting positions connected to institutions like Institut des Hautes Études Scientifiques and Kavli Institute for Theoretical Physics. He has been cited in relation to prize committees and program leadership for conferences including NeurIPS, ICML, and The Royal Society colloquia.
- Mezard, M., Parisi, G., Virasoro, M. A., "Spin Glass Theory and Beyond", works associated with Cambridge University Press and researchers from Sapienza University of Rome and École Normale Supérieure. - Mezard, M., Montanari, A., "Information, Physics, and Computation", influential in communities around Princeton University Press and Université Paris-Sud. - Mezard, M., Parisi, G., Zecchina, R., papers on survey propagation, cited alongside studies at Université de Rome La Sapienza and groups linked to Bell Labs. - Mezard, M., Krzakala, F., Zdeborová, L., works on compressed sensing and phase transitions, referenced by teams at Caltech and ETH Zurich. - Selected articles in journals such as Physical Review Letters, Journal of Statistical Mechanics: Theory and Experiment, and publications affiliated with IEEE Transactions on Information Theory.
Category:French physicists Category:Statistical physicists Category:Machine learning researchers