Coarse-Grained Density Functional Theory Predictions via Deep Kernel Learning

Journal Article · 11 January 2022 · Journal of Chemical Theory and Computation

DOI: https://doi.org/10.1021/acs.jctc.1c01001 · OSTI ID:1880131

Sivaraman, Ganesh  ^[1]; Jackson, Nicholas E.  ^[2]

---

1. Argonne National Lab. (ANL), Argonne, IL (United States)
2. Univ. of Illinois at Urbana-Champaign, IL (United States)

Scalable electronic predictions are critical for soft materials design. Recently, the Electronic Coarse-Graining (ECG) method was introduced to renormalize all-atom quantum chemical (QC) predictions to coarse-grained (CG) resolutions using deep neural networks (DNNs). While DNNs can learn complex representations that prove challenging for kernel-based methods, they are susceptible to overfitting and the overconfidence of uncertainty estimations. Here, we develop ECG within a GPU-accelerated Deep Kernel Learning (DKL) framework to enable CG QC predictions using range-separated hybrid density functional theory (DFT), obtaining a 107 speedup relative to naive all-atom QC. By treating the predicted electronic properties as random Gaussian Processes, DKL incorporates CG mapping degeneracy by learning the distribution of electronic energies as a function of CG configuration. DKL-ECG accurately reproduces molecular orbital energies from range-separated DFT while facilitating efficient training via active learning using the uncertainties provided by DKL. Further, we show that while active learning algorithms enable efficient sampling of a more diverse configurational space relative to random sampling, all explored query methods exhibit comparable performance for the examined system. We attribute this result to the significant overlap of the feature space and output property distributions across multiple temperatures.

• Research Organization: Argonne National Laboratory (ANL), Argonne, IL (United States)
• Sponsoring Organization: USDOE Office of Science (SC); USDOE Laboratory Directed Research and Development (LDRD) Program
• Grant/Contract Number: AC02-06CH11357
• OSTI ID: 1880131
• Journal Information: Journal of Chemical Theory and Computation, Journal Name: Journal of Chemical Theory and Computation Journal Issue: 2 Vol. 18; ISSN 1549-9618
• Publisher: American Chemical SocietyCopyright Statement
• Country of Publication: United States
• Language: English