Regular biography
Clément Hongler is an Associate Professor at the École Polytechnique Fédérale de Lausanne (EPFL), affiliated with the Department of Mathematics (MATH). His research interests include statistical mechanics, quantum field theory, and deep learning theory. Hongler's work explores connections between lattice models and conformal field theories, the dynamics of learning in neural networks, artificial life through continuous cellular automata, and decentralized systems. He has held academic positions at Columbia University and EPFL, where he has been involved in teaching and mentoring PhD students. His teaching includes the course 'Probabilistic models of modern AI' (MATH-415).
Scholar-generated biography
Clément Hongler is a researcher at the Institute of Mathematics, EPFL, focusing on Probability, Complex Analysis, Statistical Mechanics, Conformal Field Theory, and Neural Networks. His work explores the interplay between mathematical physics and machine learning, particularly in understanding the behavior of deep neural networks and their connections to statistical mechanics. Hongler's research includes the analysis of neural tangent kernels, generalization in deep learning, and the convergence of Ising interfaces to Schramm's SLE curves. He also investigates conformal invariance in spin correlations and crossing probabilities in the Ising model, contributing to both theoretical and applied aspects of complex systems.