 Information-theoretic reduction of deep neural networks to linear models in the overparametrized proportional regimeAnnual Conference Computational Learning Theory (COLT), 2025 |
 Local Loss Optimization in the Infinite Width: Stable Parameterization of Predictive Coding Networks and Target PropagationInternational Conference on Learning Representations (ICLR), 2024 |
 Phase diagram of early training dynamics in deep neural networks: effect
of the learning rate, depth, and widthNeural Information Processing Systems (NeurIPS), 2023 |
 Relative stability toward diffeomorphisms indicates performance in deep
netsNeural Information Processing Systems (NeurIPS), 2021 |
 Double-descent curves in neural networks: a new perspective using
Gaussian processesAAAI Conference on Artificial Intelligence (AAAI), 2021 |
 Tilting the playing field: Dynamical loss functions for machine learningInternational Conference on Machine Learning (ICML), 2021 |
