Equivariant message passing for the prediction of tensorial properties and molecular spectra

2021/02/05 by Kristof T. Schütt, Oliver T. Unke, Schütt, Kristof T. +3 · 56 citations
Biochemistry, Genetics and Molecular Biology · Computer Science · Materials Science · #Chemical Physics (physics.chem-ph) #Computational Drug Discovery Methods #FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (cs.LG) #Machine Learning in Materials Science #Protein Structure and Dynamics

paper · pdf · doi:10.48550/arxiv.2102.03150

openalex publication_date 2021/02/05 · openalex created_date 2021/02/15 · openalex updated_date 2026/07/28

Abstract

Message passing neural networks have become a method of choice for learning on graphs, in particular the prediction of chemical properties and the acceleration of molecular dynamics studies. While they readily scale to large training data sets, previous approaches have proven to be less data efficient than kernel methods. We identify limitations of invariant representations as a major reason and extend the message passing formulation to rotationally equivariant representations. On this basis, we propose the polarizable atom interaction neural network (PaiNN) and improve on common molecule benchmarks over previous networks, while reducing model size and inference time. We leverage the equivariant atomwise representations obtained by PaiNN for the prediction of tensorial properties. Finally, we apply this to the simulation of molecular spectra, achieving speedups of 4-5 orders of magnitude compared to the electronic structure reference.

Cited by

• mCGCNN: A Dual-Stream Crystal Graph Convolutional Neural Network for the Efficient Prediction of Magnetic Properties of Crystalline Materials
• DeepHartree: A Poisson-Coupled Neural Field for One-Shot Density Functional Theory
• QE-Catalytic: A Graph-Language Multimodal Base Model for Relaxed-Energy Prediction in Catalytic Adsorption
• Long-range electrostatics for machine learning interatomic potentials is easier than we thought
• How accurate are foundational machine learning interatomic potentials for heterogeneous catalysis?
• Accelerating High-Throughput Catalyst Screening by Direct Generation of Equilibrium Adsorption Structures
• GoMS: Graph of Molecule Substructure Network for Molecule Property Prediction
• Characterizing Defect Dynamics in Silicon Carbide Using Symmetry-Adapted Collective Variables and Machine Learning Interatomic Potentials
• QuantumCanvas: A Multimodal Benchmark for Visual Learning of Atomic Interactions
• CORGI: GNNs with Convolutional Residual Global Interactions for Lagrangian Simulation
• Equivariant Atomic and Lattice Modeling Using Geometric Deep Learning for Crystal Structure Optimization
• A Circular Argument : Does RoPE need to be Equivariant for Vision?
• Hierarchical Direction Perception via Atomic Dot-Product Operators for Rotation-Invariant Point Clouds Learning
• Magnitude-Modulated Equivariant Adapter for Parameter-Efficient Fine-Tuning of Equivariant Graph Neural Networks
• Design, Assessment, and Application of Machine Learning Potential Energy Surfaces
• Atomistic Simulations of Oxide-Water Interfaces using Machine Learning Potentials
• Cross Learning between Electronic Structure Theories for Unifying Molecular, Surface, and Inorganic Crystal Foundation Force Fields
• Knowledge Distillation of a Protein Language Model Yields a Foundational Implicit Solvent Model
• Learning 3D Anisotropic Noise Distributions Improves Molecular Force Field Modeling
• HollowFlow: Efficient Sample Likelihood Evaluation using Hollow Message Passing
• 3D-GSRD: 3D Molecular Graph Auto-Encoder with Selective Re-mask Decoding
• FLOWR.root: A flow matching based foundation model for joint multi-purpose structure-aware 3D ligand generation and affinity prediction
• GO-Diff: Data-free and amortized global structure optimization
• SimPoly: Simulation of Polymers with Machine Learning Force Fields Derived from First Principles
• Reciprocal Space Attention for Learning Long-Range Interactions
• Universally Invariant Learning in Equivariant GNNs
• Catalyst GFlowNet for electrocatalyst design: A hydrogen evolution reaction case study
• Simulating enzyme catalysis with electrostatically embedded machine learning potentials
• When Does Global Attention Help? A Unified Empirical Study on Atomistic Graph Learning
• Flow-Matching Based Refiner for Molecular Conformer Generation
• Atomistic Machine Learning with Irreducible Cartesian Natural Tensors
• A Universal Deep Learning Force Field for Molecular Dynamic Simulation and Vibrational Spectra Prediction
• Are neural scaling laws leading quantum chemistry astray?
• Learning the Neighborhood: Contrast-Free Multimodal Self-Supervised Molecular Graph Pretraining
• MCGM: Multi-stage Clustered Global Modeling for Long-range Interactions in Molecules
• Scalable Foundation Interatomic Potentials via Message-Passing Pruning and Graph Partitioning
• Shoot from the HIP: Hessian Interatomic Potentials without derivatives
• Chem World: A Large-Scale Benchmark and Physics-Informed Framework for Trustworthy Chemical Property Prediction
• Artificial Intelligence for Direct Prediction of Molecular Dynamics across Chemical Space
• Graph atomic cluster expansion for foundational machine learning interatomic potentials
• DEQuify your force field: More efficient simulations using deep equilibrium models
• DiffPlace: A Conditional Diffusion Framework for Simultaneous VLSI Placement Beyond Sequential Paradigms
• Bayesian Prior Construction for Uncertainty Quantification in First-Principles Statistical Mechanics
• Quasi-van Hove singularities informed approach improving DOS/pDOS predictions in GNN
• Equivariant U-Shaped Neural Operators for the Cahn-Hilliard Phase-Field Model
• Node-Equivariant Message Passing for Efficient and Accurate Machine Learning Interatomic Potentials
• Challenges in Non-Polymeric Crystal Structure Prediction: Why a Geometric, Permutation-Invariant Loss is Needed
• MACE4IRmol: An uncertainty-aware foundation model for molecular infrared spectroscopy
• LFaB: Low fidelity as Bias for Active Learning in the chemical configuration space
• Equi-mRNA: Protein Translation Equivariant Encoding for mRNA Language Models
• BLIPs: Bayesian Learned Interatomic Potentials
• Constructing 3D Rotational Invariance and Equivariance with Symmetric Tensor Networks
• RAFFLE: active learning accelerated interface structure prediction
• The Open DAC 2025 Dataset for Sorbent Discovery in Direct Air Capture
• Open Molecular Crystals 2025 (OMC25) Dataset and Models
• Iterative Pretraining Framework for Interatomic Potentials

Related

• Direct Molecular Polarizability Prediction with SO(3) Equivariant Local Frame GNNs
• TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials
• Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing
• Broadening the Scope of Neural Network Potentials through Direct Inclusion of Additional Molecular Attributes
• Node-Equivariant Message Passing for Efficient and Accurate Machine Learning Interatomic Potentials