Unifying O(3) Equivariant Neural Networks Design with Tensor-Network Formalism

Abstract

Many learning tasks, including learning potential energy surfaces from ab initio calculations, involve global spatial symmetries and permutational symmetry between atoms or general particles. Equivariant graph neural networks are a standard approach to such problems, with one of the most successful methods employing tensor products between various tensors that transform under the spatial group. However, as the number of different tensors and the complexity of relationships between them increase, maintaining parsimony and equivariance becomes increasingly challenging. In this paper, we propose using fusion diagrams, a technique widely employed in simulating SU($2$)-symmetric quantum many-body problems, to design new equivariant components for equivariant neural networks. This results in a diagrammatic approach to constructing novel neural network architectures. When applied to particles within a given local neighborhood, the resulting components, which we term "fusion blocks," serve as universal approximators of any continuous equivariant function defined in the neighborhood. We incorporate a fusion block into pre-existing equivariant architectures (Cormorant and MACE), leading to improved performance with fewer parameters on a range of challenging chemical problems. Furthermore, we apply group-equivariant neural networks to study non-adiabatic molecular dynamics of stilbene cis-trans isomerization. Our approach, which combines tensor networks with equivariant neural networks, suggests a potentially fruitful direction for designing more expressive equivariant neural networks.

10 pages + 12-page supplementary materials, many figures

References in corpus (24)

• DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Entanglement renormalization
• Entanglement Renormalization and Holography
• Holographic quantum error-correcting codes: Toy models for the bulk/boundary correspondence
• Tensor networks for complex quantum systems
• Renormalization algorithms for Quantum-Many Body Systems in two and higher dimensions
• A class of ansatz wave functions for 1D spin systems and their relation to DMRG
• MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
• On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
• Cormorant: Covariant Molecular Neural Networks
• Tensor network states and algorithms in the presence of a global SU(2) symmetry
• E(n) Equivariant Graph Neural Networks
• The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials
• TensorNetwork: A Library for Physics and Machine Learning
• TensorNetwork for Machine Learning
• Rotation Invariant Graph Neural Networks using Spin Convolutions
• Simulating the Sycamore quantum supremacy circuits
• A programming guide for tensor networks with global $SU(2)$ symmetry
• Theoretical Aspects of Group Equivariant Neural Networks
• SE(3) Equivariant Graph Neural Networks with Complete Local Frames
• Benchmarking global $SU(2)$ symmetry in 2d tensor network algorithms
• Symmetric Pruning in Quantum Neural Networks
• Introduction to SU(2) recoupling theory and graphical methods for loop quantum gravity

Cited by in corpus (1)

• Towards Symmetry-Aware Efficient Simulation of Quantum Systems and Beyond