FAENet: Frame Averaging Equivariant GNN for Materials Modeling
This addresses the need for more efficient and expressive symmetry-preserving models in materials and molecular modeling, offering a novel approach that is not incremental but provides a new paradigm.
The paper tackled the problem of enforcing symmetries in graph neural networks for materials modeling, which often reduces expressivity and scalability, by introducing a flexible framework using stochastic frame-averaging to achieve E(3)-equivariance or invariance through data transformations, and demonstrated superior accuracy and computational scalability on datasets like OC20 and QM9.
Applications of machine learning techniques for materials modeling typically involve functions known to be equivariant or invariant to specific symmetries. While graph neural networks (GNNs) have proven successful in such tasks, they enforce symmetries via the model architecture, which often reduces their expressivity, scalability and comprehensibility. In this paper, we introduce (1) a flexible framework relying on stochastic frame-averaging (SFA) to make any model E(3)-equivariant or invariant through data transformations. (2) FAENet: a simple, fast and expressive GNN, optimized for SFA, that processes geometric information without any symmetrypreserving design constraints. We prove the validity of our method theoretically and empirically demonstrate its superior accuracy and computational scalability in materials modeling on the OC20 dataset (S2EF, IS2RE) as well as common molecular modeling tasks (QM9, QM7-X). A package implementation is available at https://faenet.readthedocs.io.