skip to main content

Explore Research Products in the NSF-PAR

Title: Convergence acceleration in machine learning potentials for atomistic simulations
Machine learning potentials (MLPs) for atomistic simulations have an enormous prospective impact on materials modeling, offering orders of magnitude speedup over density functional theory (DFT) calculations without appreciably sacrificing accuracy in the prediction of material properties. However, the generation of large datasets needed for training MLPs is daunting. Herein, we show that MLP-based material property predictions converge faster with respect to precision for Brillouin zone integrations than DFT-based property predictions. We demonstrate that this phenomenon is robust across material properties for different metallic systems. Further, we provide statistical error metrics to accurately determine a priori the precision level required of DFT training datasets for MLPs to ensure accelerated convergence of material property predictions, thus significantly reducing the computational expense of MLP development.  more » « less
Award ID(s):
2003808
NSF-PAR ID:
10417870
Author(s) / Creator(s):
; ; ;
Date Published:
Journal Name:
Digital Discovery
Volume:
1
Issue:
1
ISSN:
2635-098X
Page Range / eLocation ID:
61 to 69
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ( , Digital discovery)

    Machine learning atomistic potentials (MLPs) trained using density functional theory (DFT) datasets allow for the modeling of complex material properties with near-DFT accuracy while imposing a fraction of its computational cost.

     
    more » « less
  2. ; ; ; ; ( , npj Computational Materials)
    Abstract

    Despite the machine learning (ML) methods have been largely used recently, the predicted materials properties usually cannot exceed the range of original training data. We deployed a boundless objective-free exploration approach to combine traditional ML and density functional theory (DFT) in searching extreme material properties. This combination not only improves the efficiency for screening large-scale materials with minimal DFT inquiry, but also yields properties beyond original training range. We use Stein novelty to recommend outliers and then verify using DFT. Validated data are then added into the training dataset for next round iteration. We test the loop of training-recommendation-validation in mechanical property space. By screening 85,707 crystal structures, we identify 21 ultrahigh hardness structures and 11 negative Poisson’s ratio structures. The algorithm is very promising for future materials discovery that can push materials properties to the limit with minimal DFT calculations on only ~1% of the structures in the screening pool.

     
    more » « less
  3. ; ; ; ( , NMR in Biomedicine)
    Abstract

    Nuclear magnetic resonance (NMR) spectroscopy is a powerful tool for quantitative metabolomics; however, quantification of metabolites from NMR data is often a slow and tedious process requiring user input and expertise. In this study, we propose a neural network approach for rapid, automated lipid identification and quantification from NMR data. Multilayered perceptron (MLP) networks were developed with NMR spectra as the input and lipid concentrations as output. Three large synthetic datasets were generated, each with 55,000 spectra from an original 30 scans of reference standards, by using linear combinations of standards and simulating experimental‐like modifications (line broadening, noise, peak shifts, baseline shifts) and common interference signals (water, tetramethylsilane, extraction solvent), and were used to train MLPs for robust prediction of lipid concentrations. The performances of MLPS were first validated on various synthetic datasets to assess the effect of incorporating different modifications on their accuracy. The MLPs were then evaluated on experimentally acquired data from complex lipid mixtures. The MLP‐derived lipid concentrations showed high correlations and slopes close to unity for most of the quantified lipid metabolites in experimental mixtures compared with ground‐truth concentrations. The most accurate, robust MLP was used to profile lipids in lipophilic hepatic extracts from a rat metabolomics study. The MLP lipid results analyzed by two‐way ANOVA for dietary and sex differences were similar to those obtained with a conventional NMR quantification method. In conclusion, this study demonstrates the potential and feasibility of a neural network approach for improving speed and automation in NMR lipid profiling and this approach can be easily tailored to other quantitative, targeted spectroscopic analyses in academia or industry.

     
    more » « less
  4. ; ; ( , Mathematical and Scientific Machine Learning)
    A major factor in the success of deep neural networks is the use of sophisticated architectures rather than the classical multilayer perceptron (MLP). Residual networks (ResNets) stand out among these powerful modern architectures. Previous works focused on the optimization advantages of deep ResNets over deep MLPs. In this paper, we show another distinction between the two models, namely, a tendency of ResNets to promote smoother interpolations than MLPs. We analyze this phenomenon via the neural tangent kernel (NTK) approach. First, we compute the NTK for a considered ResNet model and prove its stability during gradient descent training. Then, we show by various evaluation methodologies that for ReLU activations the NTK of ResNet, and its kernel regression results, are smoother than the ones of MLP. The better smoothness observed in our analysis may explain the better generalization ability of ResNets and the practice of moderately attenuating the residual blocks. 
    more » « less
  5. ; ; ( , Mathematical and Scientific Machine Learning)
    null (Ed.)
    A major factor in the success of deep neural networks is the use of sophisticated architectures rather than the classical multilayer perceptron (MLP). Residual networks (ResNets) stand out among these powerful modern architectures. Previous works focused on the optimization advantages of deep ResNets over deep MLPs. In this paper, we show another distinction between the two models, namely, a tendency of ResNets to promote smoother interpolations than MLPs. We analyze this phenomenon via the neural tangent kernel (NTK) approach. First, we compute the NTK for a considered ResNet model and prove its stability during gradient descent training. Then, we show by various evaluation methodologies that for ReLU activations the NTK of ResNet, and its kernel regression results, are smoother than the ones of MLP. The better smoothness observed in our analysis may explain the better generalization ability of ResNets and the practice of moderately attenuating the residual blocks. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email