Equivariant Models | Open Catalyst Intro Series | Ep. 6

Показать описание

Episode 6: In this episode, we explore ML models that have equivariant representations. These model representations are quite fascinating, since they change predictably given changes in the input. For instance, if the input atoms are rotated, the model’s internal representation will also “rotate”. We’ll discuss how a special set of basis functions called spherical harmonics are used in equivariant models to represent atom neighborhoods and what makes them so mathematically interesting.

This video series is aimed at machine learning and AI researchers interested in gaining a better understanding of how to explore machine learning problems in chemistry and material science.

#opencatalyst #ai4science #climatechange

Additional materials:

Videos on Fourier Transforms:

Some equivariant model papers:
E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials:

Open Catalyst Project

Рекомендации по теме

Комментарии

Brilliant explanation! My only regret is, that I haven't found this earlier.

paranoid_android

Awesome video. May I ask how might I apply the Wigner D matrix?

I have the sum wave function. Then right now, I am trying to shift it by A that's just a scalar/number and not a matrix. It works for one wave function, but for two it's not moving in unison and it seems to be moving independantly. (Not an issue, as expected).

But, now, how do I go about getting the Wigner D and how to apply it? I figured I can do J=1/2, but it feels like I need to configure it with alpha, beta, gamma but what would those values be in this context?

landland

I love these brave women! Thanks for standing up!

raguaviva

Equivariant Models | Open Catalyst Intro Series | Ep. 6

Equivariant Models | Open Catalyst Intro Series | Ep. 6

Invariant Models | Open Catalyst Intro Series | Ep. 5

Why model atoms? | Open Catalyst Intro Series | Ep. 1

A New Perspective on Building Efficient and Expressive 3D Equivariant Graph Neural Networks

2021-04-16 - Zachary Ulissi, Larry Zitnick - The Open Catalyst 2020 Dataset & Community Challeng...

Equivariant Neural Networks | Part 3/3 - Transformers and GNNs

NeurIPS 2021: Open Catalyst Challenge | 2nd place — Innopolis AI

Open Catalyst Project Tutorial: An Introduction to Machine Learning for Material Discovery

NeurIPS 2021: Open Catalyst Challenge | 1st place — Microsoft Research Asia

NeurIPS 2022: Open Catalyst Challenge Event

Group Equivariant Deep Learning - Lecture 3.1: Motivation for SE(3) equivariant graph NNs

Learning Local Equivariant Representations for Large-Scale Atomistic Dynamics | Albert Musaelian

Local Equivariant Representations for Large-Scale Atomistic Dynamics | S. Batzner and A. Musaelian

Materials Project Seminar Series Episode 2: Prof Zachary W. Ulissi

Modeling of Complex Energy Materials with Machine Learning

Materials Project Seminars – Zachary Ulissi, 'Accelerating Catalyst Discovery Using General [.....

Improving the Speed and Accuracy of Neural Network Interatomic Potentials Across Chemical Domains

The Frontiers of Modeling Atoms with AI—Larry Zitnick (Meta, Inc.)

Generative Language Models in Molecular Discovery: Regression Transformer, GT4SD and Beyond

Boris Kozinsky - Uncertainty-aware machine learning models of many-body atomic interactions

Examining GNNs for Crystal Structures: Limitations and Opportunities for Capturing Periodicity

Towards Predicting Equilibrium Distributions for Molecular Systems with Deep Learning | Shuxin Zheng

Yuanqing Wang - Parameterization of Extended Force Field using Graph Neural Nets

ML4Mol: Graph Neural Network Part 1