Publications related to On the atypical solutions of the symmetric binary perceptron

Related publications (11)

Phase Diagram of Compressed Sensing with l0-norm Regularization

Lenka Zdeborová, Florent Gérard Krzakala, Damien Yves Valery Barbier

Noiseless compressive sensing is a two-step setting that allows for undersampling a sparse signal and then reconstructing it without loss of information. The LASSO algorithm, based on l(1) regularization, provides an efficient and robust way to address thi ...

IOP Publishing Ltd2025

Microscopic Origin of Reduced Magnetic Order in a Frustrated Metal

Andreas Martin Läuchli Herzig

Although magnetic frustration in metals provides a promising avenue for novel quantum phenomena, their microscopic interpretation is often challenging. Here, we use the face-centered cubic intermetallic HoInCu_{4} as model material to show that Hamiltonian ...

American Physical Society (APS)2025

Statistical mechanics of deep learning

Freya Behrens

2024

Neural Redshift: Random Networks are not Random Functions

Valentin Hartmann

Our understanding of the generalization capabilities of neural networks (NNs) is still incomplete. Prevailing explanations are based on implicit biases of gradient descent (GD) but they cannot account for the capabilities of models from gradient-free metho ...

IEEE2024

Theory and applications of the sum-of-squares technique

Luca Pesce, Giovanni Piccioli, Elisabetta Cornacchia

The sum-of-squares (SOS) approximation method is a technique used in optimization problems to derive lower bounds on the optimal value of an objective function. By representing the objective function as a sum of squares in a feature space, the SOS method t ...

2024

Predicting nonlinear optical scattering with physics-driven neural networks

Demetri Psaltis, Carlo Gigli

Deep neural networks trained on physical losses are emerging as promising surrogates for nonlinear numerical solvers. These tools can predict solutions to Maxwell's equations and compute gradients of output fields with respect to the material and geometric ...

AIP Publishing2023

Characterization of two electronic subsystems in cuprates through optical conductivity

Ana Akrap, Edoardo Martino, Neven Barisic

Understanding the physical properties of unconventional superconductors as well as of other correlated materials presents a formidable challenge. Their unusual evolution with doping, frequency, and temperature has frequently led to non-Fermi-liquid (non-FL ...

2023

Essays in macro-finance and deep learning

Goutham Gopalakrishna

This thesis studies the origins and consequences of financial crises, and computational techniques to solve continuous-time economic models that explain such crises. The first chapter shows that financial recessions are typically characterised by a large ...

EPFL2023

Reconstruction of decays to merged photons using end-to-end deep learning with domain continuation in the CMS detector

Jian Wang, Matthias Finger, Qian Wang, Yiming Li, João Miguel das Neves Duarte, Varun Sharma, Yi Zhang, Lei Zhang, Tian Cheng, Yixing Chen, Alexis Kalogeropoulos, Ioannis Papadopoulos, Hua Zhang, Siyuan Wang, Jessica Prisciandaro, Xin Chen, Michele Bianco, Sebastiana Gianì, Sun Hee Kim, Davide Di Croce, Guido Andreassi, Arvind Shah, Jian Zhao, Rakesh Chawla, Jan Steggemann, Konstantin Androsov, Federica Legger, Gabriele Grosso

A novel technique based on machine learning is introduced to reconstruct the decays of highly Lorentz boosted particles. Using an end-to-end deep learning strategy, the technique bypasses existing rule-based particle reconstruction methods typically used i ...

American Physical Society2023

Bottlenecks, Modularity, and the Neural Control of Behavior

Riccardo Ravasio

In almost all animals, the transfer of information from the brain to the motor circuitry is facilitated by a relatively small number of neurons, leading to a constraint on the amount of information that can be transmitted. Our knowledge of how animals enco ...

FRONTIERS MEDIA SA2022

Self-Supervised Generative Style Transfer for One-Shot Medical Image Segmentation

Jean-Philippe Thiran, Seyedbehzad Bozorgtabar, Mohammad Saeed Rad, Guillaume Marc Georges Vray, Devavrat Tomar

In medical image segmentation, supervised deep networks' success comes at the cost of requiring abundant labeled data. While asking domain experts to annotate only one or a few of the cohort's images is feasible, annotating all available images is impracti ...

IEEE COMPUTER SOC2022