Ryan Boustany

According to our database, Ryan Boustany authored at least 5 papers between 2022 and 2025.

Collaborative distances:

• Dijkstra number of five.
• Erdős number of four.

2025

When majority rules, minority loses: bias amplification of gradient descent.

François Bachoc, Jérôme Bolte, Ryan Boustany, Jean-Michel Loubes

CoRR, May, 2025

2024

On the numerical reliability of nonsmooth autodiff: a MaxPool case study.

Ryan Boustany

Trans. Mach. Learn. Res., 2024

A second-order-like optimizer with adaptive gradient scaling for deep learning.

Jérôme Bolte, Ryan Boustany, Edouard Pauwels, Andrei I. Purica

CoRR, 2024

2023

On the complexity of nonsmooth automatic differentiation.

Jérôme Bolte, Ryan Boustany, Edouard Pauwels, Béatrice Pesquet-Popescu

Proceedings of the Eleventh International Conference on Learning Representations, 2023

2022

Nonsmooth automatic differentiation: a cheap gradient principle and other complexity results.

Jérôme Bolte, Ryan Boustany, Edouard Pauwels, Béatrice Pesquet-Popescu

CoRR, 2022