I feel the same about generating functions.

Cantor-Schröder-Bernstein's theorem tells that given sets A and B, if there're injective functions f : A → B and g : B → A, then there's a bijection h : A → B. A famous proof, given by J. König is strikingly similar to the Khun's algorithm for max bipartite matching.

Why do GATs work worse than MPNN for neural algorithmic reasoning? (experiments done for “learning” Bellman-Ford’s algorithm) Any light on that, NAR people?

I loved this session with @mathildepapillon.bsky.social. Here it's the YouTube recording: youtu.be/g3fecArn-2A.

December 5th our ML theory group at Cohere For AI is hosting @mathildepapillon.bsky.social to discuss their recent review arxiv.org/abs/2407.09468 on geometric/topological/algebraic ML. Join us online 💫

A starter pack for researchers interested in Geometric Deep Learning - in the broadest sense possible! Let me know if you would like to be listed. :) Thanks @sharvaree.bsky.social for the idea! go.bsky.app/7h8sek