Insights Over-squashing, Bottlenecks, and Graph Ricci curvature

Over-squashing is a common plight of Graph Neural Networks. In this post, we discuss how this phenomenon can be understood and remedied through the concept of Ricci curvature.

Insights GNNs through the lens of differential geometry and algebraic topology

In this post, we show how tools from the fields of differential geometry and algebraic topology can be used to reinterpret GNNs and address some of their common plights in a principled way.

Insights Graph Neural Networks as Neural Diffusion PDEs

In this post, we will discuss our recent work on neural graph diffusion networks.

Insights Provably expressive graph neural networks

Twitter describes how to design local and computationally efficient provably powerful graph neural networks.

Insights ICLR Invited Talk on Geometric Deep Learning

This blog post is based on Michael Bronstein’s keynote talk at ICLR 2021 and the paper M. M. Bronstein et al., Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges (2021)

Insights Simple scalable graph neural networks

In this post, we describe a graph neural network architecture (SIGN) that is of simple implementation and that works on very large graphs.

Insights Deep learning on dynamic graphs

A new neural network architecture for dynamic graphs

Insights Graph ML at Twitter

How we do Graph Machine Learning at Twitter