Gerard Ben Arous
7 papers · 2015–2024 · 6 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+3 more ↓ Show less ↑
π£ Hot Topic Early Bird π Conference Polyglot (6) π Academic Marathon (9) π Cross-Pollinator (15) π Interdisciplinary Bridge
πΊοΈ
Taxonomy Completionist
(13)
π§
Keyword Pioneer
π
Keyword Trendsetter Combo
(3)
Conferences
COLT (2)
AISTATS (1)
ICLR (1)
ICML (1)
JMLR (1)
NIPS (1)
Top co-authors
Keywords
stochastic gradient descent
(3)
loss surface
(2)
neural network
(2)
spin-glass model
(2)
high-dimensional inference
(1)
loss landscape
(1)
markov chain monte carlo
(1)
sparse principal component analysis
(1)
high-dimensional analysis
(1)
free energy
(1)
phase retrieval
(1)
random matrix theory
(1)
generalized linear model
(1)
gradient flow
(1)
stochastic differential equation
(1)
training dynamics
(1)
multilayer network
(1)
statistical physics
(1)
saddle point
(1)
critical point
(1)
Papers
High-dimensional SGD aligns with emerging outlier eigenspaces
ICLR 2024
High-dimensional limit theorems for SGD: Effective dynamics and critical scaling
NIPS 2022
Online stochastic gradient descent on non-convex losses from high-dimensional inference
JMLR 2021
Free Energy Wells and Overlap Gap Property in Sparse PCA
COLT 2020
Comparing Dynamics: Deep Neural Networks versus Glassy Systems
ICML 2018
Open Problem: The landscape of the loss surfaces of multilayer networks
COLT 2015
The Loss Surfaces of Multilayer Networks
AISTATS 2015