Thrilled to share the recording from my Simons Foundation Presidential lecture given last Wednesday. What else than: Statistical Physics of Machine Learning :). youtu.be/TLHYwbrhGJc?si…

Amazing talk: The statistical physics of machine learning

youtu.be/TLHYwbrhGJc?si…

The Sherrington-Kirkpatrick model is a prototypical model of a complex landscape. We answer some of the open questions about the behaviour of greedy descents in: arxiv.org/pdf/2405.04267 Curiously, being most reluctantly greedy seems to get you to the ground state energy density.

Former PhD students starting their faculty jobs is always some of the best news to learn about! Congrats Stefano!

Presidential Lecture 5/1: Lenka Zdeborová (Lenka Zdeborova) on 'Statistical Physics of Machine Learning.' NYC, 6pm. Free to attend but registration is required: simonsfoundation.org/event/statisti…

Starting tomorrow: If you are around MIT join for the Simons Lectures Series 2024 math.mit.edu

Dynamical phase transitions in graph cellular automata, Freya Behrens, Barbora Hudcová, and Lenka Zdeborová #Networks #ComplexSystems #PhaseTransitions Lenka Zdeborova EPFL go.aps.org/4b1DoHu

Talking to many junior faculty members and students in AI lately. Many seem to be somewhat lost with all the seemingly fast progresses made by the industry. My suggestion to them is: It is industry's job to find how to do better, but academia is to find out how to do it right.

Doesn’t everyone feel like that? Those who do not either don’t care about AI or fake it.

Dear Twitter, there was recently a paper showing that different diffusion models generate very similar images when starting with the same initial noise. I did not note the reference. Somebody, please, has it?

Hello Twitter, were the accepted ICML 2024 workshops announced already?

Almost everyone confuses knowledge with intelligence. There’s fundamental difference between absorbing existing knowledge and discovering new one, like the difference between a student and a scientist. That is the reason why existing AI systems require good data to begin with.