MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields

Explore a comprehensive lecture on MACE (Higher Order Equivariant Message Passing Neural Networks) for fast and accurate force fields in computational chemistry and materials science. Delve into representations of interacting particle clouds, focusing on O(3) symmetry in chemistry. Examine the MACE model's message expansion technique and its efficient application to point cloud machine learning. Analyze MACE's impressive results and participate in an engaging Q&A session. Learn how this innovative approach addresses limitations of traditional MPNNs, achieving state-of-the-art accuracy with improved computational efficiency and scalability.

- Intro
- Representations of clouds of particles in interaction
- The case of O93 for chemistry
- MACE: Message expansion
- Efficient machine learning on point clouds
: MACE results
- Q+A and Discussion