DEQuify your force field: More efficient simulations using deep equilibrium models
Not on the map. Why?
Abstract (arXiv)
Machine learning force fields show great promise in enabling more accurate molecular dynamics simulations compared to manually derived ones. Much of the progress in recent years was driven by exploiting prior knowledge about physical systems, in particular symmetries under rotation, translation, and reflections. In this paper, we argue that there is another important piece of prior information that, thus fa,r hasn't been explored: Simulating a molecular system is necessarily continuous, and successive states are therefore extremely similar. Our contribution is to show that we can exploit this information by recasting a state-of-the-art equivariant base model as a deep equilibrium model. This allows us to recycle intermediate neural network features from previous time steps, enabling us to improve both accuracy and speed by 10%-20% on the MD17, MD22, and OC20 200k datasets, compared to the non-DEQ base model. The training is also much more memory efficient, allowing us to train more expressive models on larger systems.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 26 of its 47 references that are in the map. First posted on arXiv 10 Sep 2025.
Builds on
26 of 47 references.
References · 26
- 2021How Attentive are Graph Attention Networks?cited by 835
- 2018Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point cloudscited by 746
- 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 708
- 2019Deep Equilibrium Modelscited by 615
- 2017SchNet: A continuous-filter convolutional neural network for modeling quantum interactionscited by 571
- 2020Directional Message Passing for Molecular Graphscited by 561
- 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
- 2021Equivariant message passing for the prediction of tensorial properties and molecular spectracited by 441
- 2021GemNet: Universal Directional Graph Neural Networks for Moleculescited by 343
- 2022Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphscited by 264
- 2019Cormorant: Covariant Molecular Neural Networkscited by 263
- 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 256
14 more
- 2023EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representationscited by 204
- 2021Spherical Message Passing for 3D Graph Networkscited by 168
- 2015A Theoretically Grounded Application of Dropout in Recurrent Neural Networkscited by 161
- 2023Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNscited by 117
- 2021JFB: Jacobian-Free Backpropagation for Implicit Networkscited by 86
- 2021On Training Implicit Modelscited by 75
- 2021Stabilizing Equilibrium Models by Jacobian Regularizationcited by 72
- 2022Deep Equilibrium Optical Flow Estimationcited by 64
- 2023One-Step Diffusion Distillation via Deep Equilibrium Modelscited by 61
- 2018Trellis Networks for Sequence Modelingcited by 52
- 2023From Molecules to Materials: Pre-training Large Generalizable Models for Atomic Property Predictioncited by 39
- 2023TorchDEQ: A Library for Deep Equilibrium Modelscited by 21
- 2023Deep Equilibrium Diffusion Restoration with Parallel Samplingcited by 17
- 2023Lightweight equivariant model for efficient machine learning interatomic potentialscited by 1
Built on by
Citing papers per year: 2026: 3.
3 papers.