Stéphane d'Ascoli

Stéphane d'Ascoli

Ph.D. Student

ENS and FAIR Paris


Hi! I’m a Ph.D. student working on deep learning, jointly supervised by Giulio Biroli (ENS Paris) and Levent Sagun (FAIR Paris). Before that, I studied Theoretical Physics and worked with NASA on black hole mergers.

My research focuses on understanding how deep neural networks are able to generalize despite being heavily overparametrized. On one hand, I use tools from statistical mechanics to study simple models, and try to understand when and why they overfit. On the other hand, I investigate how different types of inductive biases affect learning, from fully-connected networks to convolutional networks to transformers. I am also interested in bio-inspired alternatives to backpropagation.

I love communicating science to the general audience, and wrote a couple books for this purpose. In my spare time I also play music, feel free to check out my videos!


  • PhD in Artificial Intelligence, 2018-

    Ecole Normale Supérieure, Paris

  • Master's in Theoretical Physics, 2018

    Ecole Normale Supérieure, Paris

  • Bachelor's in Physics, 2016

    Ecole Normale Supérieure, Paris

Recent publications

ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases

More data or more parameters? Investigating the effect of data structure on generalization

Transformed CNNs: recasting pre-trained convolutional layers with self-attention

Scaling description of generalization with number of parameters in deep learning

The dynamics of learning with feedback alignment


Comprendre la révolution de l’Intelligence Artificielle

L’Intelligence Artificielle en 5 minutes par jour

Voyage au Coeur de l’Espace-Temps


L’espace-temps est courbe: qu’est-ce à dire?

La Conversation Scientifique with Etienne Klein, France Culture

Big Bang et Trous Noirs

Minute Papillon with Sidonie Bonnec, France Bleu



  • 24 rue Lhomond, Paris, 75005