Machine learning for theories and theories of machine learning

From Sunday 29th September to Thursday 3rd October 2024

Machine learning and AI are revolutionizing the sciences, reshaping everything from how we ask questions to how we build theories. At the same time, the complexity of modern deep neural networks, datasets, and computational paradigms urges us to look for new perspectives to unravel their inner machinations.
This symposium will explore the interplay between deep learning, AI, statistical mechanics, neuroscience, scientific machine learning, and applied harmonic analysis. We will discuss how AI may crack outstanding scientific problems like earthquake dynamics or the multi-scale structure of images and language, but also how drawing on analytical and experimental wisdom from neuroscience, physics, biology, ...—fields that have long grappled with extraordinary complexities—can help us better understand AI itself.


Stéphane Mallat — Collège de France/ENS

Maarten de Hoop — Rice University

Friedemann Zenke — FMI Basel

Afonso Bandeira — ETH Zurich

Misha Belkin — UCSD

Rava Azeredo da Silveira — CNRS / ENS / IOB

Mark Transtrum — Brigham Young University

Qin Li — University of Wisconsin

Julia Kempe — New York University

Chris Marone — La Sapienza / Penn State University

Marylou Gabrié — École Polytechnique

Eric Vanden-Eijnden — New York University

Matti Lassas — University of Helsinki

Steven Brunton — University of Washington

Paul Johnson — Los Alamos National Laboratory

Nicolas Flammarion — EPFL

Ryan Cotterell — ETH Zurich

Michael Mahoney — UC Berkeley

Mitya Chklovskii — Flatiron Institute/NYU Medical Center

Cengiz Pehlevan — Harvard University

Participant information

Participant information will be announced soon.

Organizers and sponsors

Ivan Dokmanić

(University of Basel)

Stéphane Mallat

(Collège de France)

Maarten de Hoop

(Rice university)


Grand Park Hotel Rovinj by Maistra Collection

Smareglijeva ulica 1A, 52210 Rovinj, Croatia