AIConfPaper
← Search

Charles Guille-Escuret

6 accepted papers

2025

Understanding Adam Requires Better Rotation Dependent Assumptions

Tianyue H. Zhang, Lucas Maes, Alan Milligan, Alexia Jolicoeur-Martineau, Ioannis Mitliagkas, Damien Scieur +2
NeurIPS 2025poster

Despite its widespread adoption, Adam's advantage over Stochastic Gradient Descent (SGD) lacks a comprehensive theoretical explanation. This paper investigates Adam's sensitivity to rotations of the parameter space. We observe that Adam's performance in training transformers degrades under random ro…

Cited by 0Source ↗Scholar ↗
2024

Expecting The Unexpected: Towards Broad Out-Of-Distribution Detection

Charles Guille-Escuret, Pierre-Andre Noel, Ioannis Mitliagkas, David Vazquez, Joao Monteiro
NeurIPS 2024poster

Deployed machine learning systems require some mechanism to detect out-of-distribution (OOD) inputs. Existing research mainly focuses on one type of distribution shift: detecting samples from novel classes, absent from the training set. However, real-world systems encounter a broad variety of anomal…

Cited by 1Source ↗PDF ↗code ↗Scholar ↗
2024

No Wrong Turns: The Simple Geometry Of Neural Networks Optimization Paths

Charles Guille-Escuret, Hiroki Naganuma, Kilian FATRAS, Ioannis Mitliagkas
ICML 2024poster

Understanding the optimization dynamics of neural networks is necessary for closing the gap between theory and practice. Stochastic first-order optimization algorithms are known to efficiently locate favorable minima in deep neural networks. This efficiency, however, contrasts with the non-convex an…

Cited by 6Source ↗PDF ↗code ↗Scholar ↗
2023

CADet: Fully Self-Supervised Out-Of-Distribution Detection With Contrastive Learning

Charles Guille-Escuret, Pau Rodriguez, David Vazquez, Ioannis Mitliagkas, Joao Monteiro
NeurIPS 2023poster

Handling out-of-distribution (OOD) samples has become a major stake in the real-world deployment of machine learning systems. This work explores the use of self-supervised contrastive learning to the simultaneous detection of two types of OOD samples: unseen classes and adversarial perturbations. Fi…

Cited by 3Source ↗PDF ↗code ↗Scholar ↗
2022

Gradient Descent Is Optimal Under Lower Restricted Secant Inequality And Upper Error Bound

Charles Guille-Escuret, Adam Ibrahim, Baptiste Goujaud, Ioannis Mitliagkas
NeurIPS 2022accept

The study of first-order optimization is sensitive to the assumptions made on the objective functions. These assumptions induce complexity classes which play a key role in worst-case analysis, including the fundamental concept of algorithm optimality. Recent work argues that strong convexity and smo…

Cited by 18Source ↗PDF ↗Scholar ↗
2021

A Study of Condition Numbers for First-Order Optimization

Charles Guille-Escuret, Manuela Girotti, Baptiste Goujaud, Ioannis Mitliagkas
AISTATS 2021poster

In this work we introduce a new framework for the theoretical study of convergence and tuning of first-order optimization algorithms (FOA). The study of such algorithms typically requires assumptions on the objective functions: the most popular ones are probably smoothness and strong convexity. Thes…

Cited by 24Source ↗PDF ↗Scholar ↗