papers

Publications (15)

cond-mat.mtrl-sci2024

Orb: A Fast, Scalable Neural Network Potential

Mark Neumann, James Gin, Benjamin Rhodes +5

We introduce Orb, a family of universal interatomic potentials for atomistic modelling of materials. Orb models are 3-6 times faster than existing universal potentials, stable unde…

cs.LG2021

Automap: Towards Ergonomic Automated Parallelism for ML Models

Michael Schaarschmidt, Dominik Grewe, Dimitrios Vytiniotis +8

The rapid rise in demand for training large neural network architectures has brought into focus the need for partitioning strategies, for example by using data, model, or pipeline…

physics.comp-ph2023

Band-gap regression with architecture-optimized message-passing neural networks

Tim Bechtel, Daniel T. Speckhard, Jonathan Godwin +1

Graph-based neural networks and, specifically, message-passing neural networks (MPNNs) have shown great potential in predicting physical properties of solids. In this work, we trai…

cs.LG2022

Simple GNN Regularisation for 3D Molecular Property Prediction & Beyond

Jonathan Godwin, Michael Schaarschmidt, Alexander Gaunt +5

In this paper we show that simple noise regularisation can be an effective way to address GNN oversmoothing. First we argue that regularisers addressing oversmoothing should both p…

cond-mat.mtrl-sci2025

Orb-v3: atomistic simulation at scale

Benjamin Rhodes, Sander Vandenhaute, Vaidotas Å imkus +4

We introduce Orb-v3, the next generation of the Orb family of universal interatomic potentials. Models in this family expand the performance-speed-memory Pareto frontier, offering…

cs.LG2022

Pre-training via Denoising for Molecular Property Prediction

Sheheryar Zaidi, Michael Schaarschmidt, James Martens +6

Many important problems involving molecular property prediction from 3D structures have limited data, posing a generalization challenge for neural networks. In this paper, we descr…