The expanding scale of machine learning models and datasets, along with methodological innovations, is rapidly extending the capabilities of artificial intelligence. Therefore, addressing fundamental questions and thoroughly understanding the basic building blocks of models and training algorithms is of utmost importance.

Given the high-dimensional nature of real-world data, tools from high-dimensional inference and statistical physics are particularly apt for tackling these theoretical challenges. Since the seminal works on the perceptron by Gardner and Derrida in the 1980s, there has been a rich interplay across disciplines, yielding a plethora of innovative ideas and insights into neural network functioning.

This multidisciplinary conference seeks to unite leading researchers from the fields of statistical physics, mathematical physics, and machine learning. Our goal is to offer diverse perspectives on crucial topics in contemporary machine learning, including:

• Associative memories, diffusion models, energy-based models;

• Representation learning, structured data modeling;

• Language modeling, self-supervised learning, alternative learning paradigms;

• Mathematical physics approaches to high-dimensional probability, spin glasses, Boltzmann machines;

• Learning dynamics and information propagation in neural networks.

The event is scheduled for September 2nd to 6th, 2024, in Roccella Jonica, Calabria, Italy, an ideal setting for scientific exchange and close to a marvelous beach. Alongside the scientific agenda, we will host a social dinner and organize hiking groups for the free afternoon in Aspromonte, offering stunning views of the Ionian Sea.

Stay tuned for more details!

Deadline for applications: May 15th. Registration is free.

Location: Convento dei Minimi, Largo Molinero, 89047 Roccella Ionica RC. You will see the banner of the conference hanging outside.

Invited Speakers

Elena Agliari - Sapienza University of Rome

Luca Ambrogioni - Radboud University

Giulio Biroli - École Normale Supérieure 

Eric de Giuli - Toronto Metropolitan University

Sebastian Goldt - International School for Advanced Studies (SISSA)

Alessandro Ingrosso - Radboud University

Dmitry Krotov - IBM Research

Florent Krzakala - École Polytechnique Fédérale de Lausanne (EPFL)

Alessandro Laio - International School for Advanced Studies (SISSA)

Bruno Loureiro - École Normale Supérieure (ENS)

Antoine Maillard - ETH Zurich

Emanuele Mingione - Alma Mater Studiorum, Università di Bologna

Marco Mondelli - Institute for Science and Technology Austria (ISTA)

Federico Ricci Tersenghi - Sapienza University of Rome

Pietro Rotondo - Università degli studi di Parma

Gabriele Sicuro - Alma Mater Studiorum, Università di Bologna

Hidenori Tanaka - Harvard University

Lenka Zdeborová - École Polytechnique Fédérale de Lausanne (EPFL)

Riccardo Zecchina - Bocconi University

Schedule

The schedule of the event is now available at Scientific programme.