TriForces: Augmenting Atomistic GNNs for Transferable Representations
For materials science researchers needing accurate MLIPs with limited task-specific data, TriForces provides a transferable representation that reduces the need for expensive DFT calculations.
TriForces introduces a model-agnostic three-stream framework with self-supervised learning to improve transferability of atomistic GNNs for interatomic potentials. It reduces energy MAE by 57% on OMat24 with only 20K samples and improves force MAE across sample sizes, while outperforming baselines on MatBench and QM9 without DFT labels.
Machine learning interatomic potentials (MLIPs) achieve excellent accuracy when trained on large Density Functional Theory (DFT) data. To be useful in practice, they must often be adapted to target chemistries using small and expensive task-specific datasets. However, MLIPs transfer inconsistently across domains, with representations that often loose accessible composition and structure information. To address this, we present TriForces, a model-agnostic three-stream framework that separates composition and structure information, combined with self-supervised learning to preserve transferable representations. TriForces improves performance on MatBench and QM9 over baselines without needing DFT labels and enables efficient similar structure retrieval through its learned latent space. On OMat24, in limited-data training regime, TriForces reduces energy MAE by 57% at 20K samples only and improves force MAE across sample sizes. We release pretrained TriForces variants across multiple MLIP architectures with code at https://github.com/Ramlaoui/triforces.