Lightweight and effective tensor sensitivity for atomistic neural networks

Journal Article · Tue May 09 00:00:00 EDT 2023 · Journal of Chemical Physics

DOI:https://doi.org/10.1063/5.0142127· OSTI ID:2441369

Chigaev, Michael Anatol ^[1]; Smith, Justin S. ^[2]; Anaya, Steven Ray ^[1]; Nebgen, Benjamin Tyler ^[1]; Bettencourt, Matthew ^[3]; Barros, Kipton Marcos ^[1]; Lubbers, Nicholas Edward ^[1]

---

1. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
2. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States); NVIDIA Corporation, Santa Clara, CA (United States)
3. NVIDIA Corporation, Santa Clara, CA (United States)

Atomistic machine learning focuses on the creation of models that obey fundamental symmetries of atomistic configurations, such as permutation, translation, and rotation invariances. In many of these schemes, translation and rotation invariance are achieved by building on scalar invariants, e.g., distances between atom pairs. There is growing interest in molecular representations that work internally with higher rank rotational tensors, e.g., vector displacements between atoms, and tensor products thereof. Here, we present a framework for extending the Hierarchically Interacting Particle Neural Network (HIP-NN) with Tensor Sensitivity information (HIP-NN-TS) from each local atomic environment. Crucially, the method employs a weight tying strategy that allows direct incorporation of many-body information while adding very few model parameters. We show that HIP-NN-TS is more accurate than HIP-NN, with negligible increase in parameter count, for several datasets and network sizes. As the dataset becomes more complex, tensor sensitivities provide greater improvements to model accuracy. In particular, HIP-NN-TS achieves a record mean absolute error of 0.927 $$\frac{\text{kcal}}{\text{mol}}$$ for conformational energy variation on the challenging COMP6 benchmark, which includes a broad set of organic molecules. We also compare the computational performance of HIP-NN-TS to HIP-NN and other models in the literature.

• Research Organization: Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
• Sponsoring Organization: USDOE Laboratory Directed Research and Development (LDRD) Program; USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC), Basic Energy Sciences (BES). Chemical Sciences, Geosciences & Biosciences Division (CSGB); USDOE Office of Science (SC), Fusion Energy Sciences (FES)
• Grant/Contract Number: 89233218CNA000001
• OSTI ID: 2441369
• Report Number(s): LA-UR--22-24861
• Journal Information: Journal of Chemical Physics, Journal Name: Journal of Chemical Physics Journal Issue: 18 Vol. 158; ISSN 0021-9606
• Publisher: American Institute of Physics (AIP)
• Country of Publication: United States
• Language: English

Related Subjects

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY
artificial neural networks
machine learning
potential energy surfaces