Publications
- Opening the Black Box: Reverse-Engineering of Sparse Neural Networks, V. Castin, R. Gribonval, GRETSI 2025. Diapositives, Poster
- A Unified Perspective on the Dynamics of Deep Transformers, V. Castin, P. Ablin, J. A. Carrillo, G. Peyré, Foundations of Computational Mathematics, 2025. Diapositives, Poster
- How Smooth Is Attention?, V. Castin, P. Ablin, G. Peyré, ICML 2024. Diapositives, Poster
Exposés sur invitation
- Séminaire SIERRA, INRIA Paris, 2026
- Minisymposium Partial Differential Equations, Inverse Problems and Machine Learning, Universidad de Alicante, janvier 2026, Diapositives
- Workshop Mathematics of Transformers, DESY Hambourg, septembre 2025. Diapositives
- Workshop Interpretability in LLMs Using Geometrical and Statistical Methods, Université d’Amsterdam, mai 2025, Diapositives
- Séminaire MLSP, ENS de Lyon, janvier 2025. Diapositives
Conférences et écoles d’été