9-12.15: Marc Mezard Statistical physics of inference
14-17.15: Andrea Montanari Lecture notes on two-layer neural networks
18.30: Welcome drink @ the institute
9-12.15: Alexander Tkatchenko: Bringing Atomistic Modeling in Chemistry and Physics and Machine Learning Together: Part 1 and Part 2.
14-14.45: Guilhem Semerjian Phase transitions in inference problems on sparse random graphs
14.45-15.30: Federico Ricci-Tersenghi Belief Propagation and Monte Carlo based algorithms to solve inference problems on sparse random graphs
15.45-16.30: Afonso Bandeira Statistical estimation under group actions: the sample complexity of multi-reference alignment
16.30-17.15: Leo Miolane Phase transitions in Generalized Linear Models
9-12.15: Dmitry Panchenko: Introductory lectures by Dmitry can be found online Lecture 1 and Lecture 2.
14-14.45: Will Perkins: Bethe states of random factor graphs
14.45-15.30: Alice Guillonnet
15.45-16.30: Christina Lee Yu: Iterative Collaborative Filtering for Sparse Matrix Estimation
16.30-17.15: Quentin Berthet Computational aspects in Statistics:Sparse PCA & Ising blockmodel
18.30: Boat trip in Cargese Harbor
9-12.15: Nicolas Brunel Learning and memory in recurrent neural networks
14-14.45: Remi Monasson
14.45-15.30: David Schwab
15.30-16.15: Surya Ganguli Theories of deep learning: generalization, expressivity, and training
18.00 - 20.00: EVENING POSTER SESSION I
9-12.15: Gérard Ben Arous
14-14.45: Jean Barbier The adaptive interpolation method for the Wigner spike model
14.45-15.30: Ahmed El Alaoui Detection limits in the spiked Wigner model
15.45-16.30: Aukosh Jagganath
16.30-17.15: Marc Lelarge Unsupervised learning:symmetric low-rank matrix estimation,community detection and triplet loss.
9-12.15: Giulio Biroli Glassy Dynamics in Physics & Beyond
14-17.15: Yann Lecun Deep Learning: Past, Present and Future
9-12.15: Sundeep Rangan Approximate Message Passing Tutorial
14-14.45: Cynthia Rush Finite Sample Analysis of AMP
14.45-15.30: PierFrancesco Urbani
15.45-16.30: Galeen Reeves
16.30-17.15: Yoshiyuki Kabashima A statistical mechanics approach to de-biasing and uncertainty estimation in LASSO
9-12.15: Naftali Tishby The Information Theory of Deep Learning: What do the layers represent?
14-14.45: Marylou Gabrie Entropy and mutual information in models of deep neural networks
14.45-15.30: Chiara Cammarota
15.30-16.15: Matthieu Wyart Loss landscape in deep learning: Role of a “Jamming” transition
18.00 - 20.00: EVENING POSTER SESSION II
9-12.15: Stephane Mallat
14-14.45: Soledad VilarOptimization and learning techniques for clustering problems
14.45-15.30: Samuel SchoenholzPRIORS FOR DEEP INFINITE NETWORKS
15.30-16.15: Francesco ZamponiRandom Close Packing vs SAT-UNSAT: a short note
9-12.15: Riccardo Zecchina
14-14.45: Levent SagunAn empirical look at the loss landscape
14.45-15.30: Jean-Philippe BouchaudEigenvector Overlaps
15.30-16.15: Giorgio Parisi
Samy Jelassi: Smoothed analysis of the low-rank approach for smooth semidefinite programs
Chris Metzler: Unsupervised Learning with Stein's Unbiased Risk Estimator
Marino Raffaele: Revisiting the challenges of MaxClique
Gabriele Sicuro: The fractional matching problem
Dmitriy (Tim) Kunisky: Tight frames, quantum information, and degree 4 sum-of-squares over the hypercube
Benjamin Aubin: Storage capacity in symmetric binary perceptrons
Sebastian Goldt: Stochastic Thermodynamics of Learning
Christian Schmidt: Estimating symmetric matrices with extensive rank
Adrian Kosowski: Ergodic Effects in Token Circulation
Chan Chun Lam: Adaptive interpolation scheme for inference problems with sparse underlying factor graph
Inbar Seroussi: Phase Transitions in Stochastic Diffusion on a General Network
Jonathan Dong: Optical realization of Echo-State Networks with light-scattering materials
Mihai Nica: Universality of log-normal distribution for randomly initialized neural nets
Andrey Lokhov: Understanding the nature of quantum annealers with statistical learning.
Eric De Giuli: Random language model -- a path to structured complexity
Grant Rotskoff: Neural networks as interacting particle systems
Endre Csóka: Local algorithms on random graphs and graph limits
Clément Luneau: Entropy of Multilayer Generalized Linear Models: proof of the replica formula with the adaptive interpolation method
Pan Zhang: Unsupervised Generative Modeling Using Matrix Product States
Andre Manoel: Approximate Message-Passing for Convex Optimization with Non-Separable Penalties
Joris Guerin: Improving Image Clustering With Multiple Pretrained CNN Feature Extractors
Ada Altieri: Constraint satisfaction mechanisms for marginal stability in large ecosystems
Federica Gerace: From statistical inference to a differential learning rule for stochastic neural networks.
Neha Wadia: In Search of Critical Points on Deep Net Optimization Landscapes
Luca Saglietti: Role of synaptic stochasticity in training low-precision neural networks
Carlo Lucibello: Limits of the MAP estimator in the phase retrieval problem
Satoshi Takabe: Trainable ISTA for Sparse Signal Recovery
Antoine Maillard: The committee machine: Computational to statistical gaps in learning a two-layers neural network
Stefano Sarao: Performance of Langevin dynamics in high dimensional inference
Aurélien Decelle: Thermodynamics properties of restricted boltzmann machines
Beatriz Seoane Bartolomé: Can a neural network learn a gauge symmetry?
Alia Abbara: Universal transitions in noiseless compressed sensing and phase retrieval
Tomoyuki Obuchi: Accelerating Cross-Validation in Multinomial Logistic Regression with L1-Regularization
You enjoy the school ? We organizers (Florent Krzakala and Lenka Zdeborova) are looking for postdocs on these topics. Come talk to us during the conference!
Tweets sur #cargese2018 <script>!function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src=p+"://platform.twitter.com/widgets.js";fjs.parentNode.insertBefore(js,fjs);}}(document,"script","twitter-wjs");</script>
Tweet #cargese2018<script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script> More information on the institute webpage
