mlsanityData to 14 Oct 2026

Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size

arXiv 2304.10061 ↗First posted 20 Apr 2023physics.comp-ph cs.LG physics.chem-phcited by 33

Not on the map. Why?

Abstract (arXiv)

This work brings the leading accuracy, sample efficiency, and robustness of deep equivariant neural networks to the extreme computational scale. This is achieved through a combination of innovative model architecture, massive parallelization, and models and implementations optimized for efficient GPU utilization. The resulting Allegro architecture bridges the accuracy-speed tradeoff of atomistic simulations and enables description of dynamics in structures of unprecedented complexity at quantum fidelity. To illustrate the scalability of Allegro, we perform nanoseconds-long stable simulations of protein dynamics and scale up to a 44-million atom structure of a complete, all-atom, explicitly solvated HIV capsid on the Perlmutter supercomputer. We demonstrate excellent strong scaling up to 100 million atoms and 70% weak scaling to 5120 A100 GPUs.

Key facts. Cited by 33 papers in the map (data through 14 Oct 2026). Builds on 18 of its 44 references that are in the map. First posted on arXiv 20 Apr 2023.

Builds on

18 of 44 references.

References · 18

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2021E(n) Equivariant Graph Neural Networkscited by 965
  3. 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 708
  4. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
  5. 2021Equivariant message passing for the prediction of tensorial properties and molecular spectracited by 441
  6. 2019Cormorant: Covariant Molecular Neural Networkscited by 263
  7. 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 256
  8. 2020Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium Moleculescited by 255
  9. 2017Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanicscited by 214
  10. 2022TorchMD-NET: Equivariant Transformers for Neural Network based Molecular Potentialscited by 182
  11. 2022e3nn: Euclidean Neural Networkscited by 174
  12. 2021SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effectscited by 104
6 more
  1. 2022The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentialscited by 101
  2. 2022Forces are not Enough: Benchmark and Critical Evaluation for Machine Learning Force Fields with Molecular Simulationscited by 95
  3. 2022SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentialscited by 82
  4. 2022Fast Uncertainty Estimates in Deep Learning Interatomic Potentialscited by 22
  5. 2022Accurate Machine Learned Quantum-Mechanical Force Fields for Biomolecular Simulationscited by 9
  6. 2021Towards Electronic Structure-Based Ab-Initio Molecular Dynamics Simulations with Hundreds of Millions of Atomscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 6; 2024: 10; 2025: 10; 2026: 7.

33 papers.

Sort:

By area:

  1. 2023TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentialscited by 54
  2. 2023LagrangeBench: A Lagrangian Fluid Mechanics Benchmarking Suitecited by 27
  3. 2025A practical guide to machine learning interatomic potentials -- Status and futurecited by 26
  4. 2023Evaluation of the MACE Force Field Architecture: from Medicinal Chemistry to Materials Sciencecited by 25
  5. 2023Navigating protein landscapes with a machine-learned transferable coarse-grained modelcited by 23
  6. 2024Generalizing Denoising to Non-Equilibrium Structures Improves Equivariant Force Fieldscited by 18
  7. 2025Force-Free Molecular Dynamics Through Autoregressive Equivariant Networkscited by 15
  8. 2026EquiformerV3: Scaling Efficient, Expressive, and General SE(3)-Equivariant Graph Attention Transformerscited by 11
  9. 2023Variational Monte Carlo on a Budget -- Fine-tuning pre-trained Neural Wavefunctionscited by 9
  10. 2024Stability-Aware Training of Machine Learning Force Fields with Differentiable Boltzmann Estimatorscited by 8
  11. 2025MOFSimBench: Evaluating Universal Machine Learning Interatomic Potentials In Metal--Organic Framework Molecular Modelingcited by 7
  12. 2024Overcoming systematic softening in universal machine learning interatomic potentials by fine-tuningcited by 6
21 more
  1. 2025chemtrain-deploy: A parallel and scalable framework for machine learning potentials in million-atom MD simulationscited by 6
  2. 2023Learning Interatomic Potentials at Multiple Scalescited by 5
  3. 2024A Multi-Grained Symmetric Differential Equation Model for Learning Protein-Ligand Binding Dynamicscited by 5
  4. 2025Rao-Blackwell Gradient Estimators for Equivariant Denoising Diffusioncited by 5
  5. 2025DistMLIP: A Distributed Inference Platform for Machine Learning Interatomic Potentialscited by 5
  6. 2024LTAU-FF: Loss Trajectory Analysis for Uncertainty in Atomistic Force Fieldscited by 4
  7. 2024On the design space between molecular mechanics and machine learning force fieldscited by 4
  8. 2024Grappa -- A Machine Learned Molecular Mechanics Force Fieldcited by 3
  9. 2024Learning Macroscopic Dynamics from Partial Microscopic Observationscited by 3
  10. 2026Beyond Adam: SOAP and Muon for Faster, Label-Efficient Training of Machine Learning Interatomic Potentialscited by 3
  11. 2026Smooth Dynamic Cutoffs for Machine Learning Interatomic Potentialscited by 2
  12. 2025Towards scientific machine learning for granular material simulations -- challenges and opportunitiescited by 1
  13. 2025Efficient Parallelization of Message Passing Neural Network Potentials for Large-scale Molecular Dynamicscited by 1
  14. 2024Considerations in the use of ML interaction potentials for free energy calculationscited by 0
  15. 2024Implicit Delta Learning of High Fidelity Neural Network Potentialscited by 0
  16. 2025Distributed Equivariant Graph Neural Networks for Large-Scale Electronic Structure Predictioncited by 0
  17. 2025Learning 3D Anisotropic Noise Distributions Improves Molecular Force Field Modelingcited by 0
  18. 2026Equivariant Evidential Deep Learning for Interatomic Potentialscited by 0
  19. 2026Multi-objective optimization and quantum hybridization of equivariant deep learning interatomic potentialscited by 0
  20. 2026MALOQ: Massively Accelerated Learning of Operators for Quantum Transportcited by 0
  21. 2026BranchIP: Learning Adaptive Equivariant Computation for Interatomic Potentialscited by 0