Learning from Failure: Predicting Electronic Structure Calculation Outcomes with Machine Learning Models
Name
Learning_from_Failure__Predicting_Electronic_Structure_Calculation_Outcomes_with_Machine_Learning_Models_v1.pdf
Description
Accepted version
Size
6.88 MB
Format
Adobe PDF
Checksum (MD5)
829ad7ba4880cae074221b04d577c29e
Author(s)
Duan, Chenru
Janet, Jon Paul
Liu, Fang
Nandy, Aditya
Kulik, Heather Janine
Date Issued
March 2019
Journal
Journal of Chemical Theory and Computation
Publisher
American Chemical Society (ACS)
Citation
Duan, Chenru et al., "Learning from Failure: Predicting Electronic Structure Calculation Outcomes with Machine Learning Models." Journal of Chemical Theory and Computation 15, 4 (April 2019): 2331–2345 doi. 10.1021/acs.jctc.9b00057 ©2019 Authors
Version
Author's final manuscript
Abstract
High-throughput computational screening for chemical discovery mandates the automated and unsupervised simulation of thousands of new molecules and materials. In challenging materials spaces, such as open shell transition metal chemistry, characterization requires time-consuming first-principles simulation that often necessitates human intervention. These calculations can frequently lead to a null result, e.g., the calculation does not converge or the molecule does not stay intact during a geometry optimization. To overcome this challenge toward realizing fully automated chemical discovery in transition metal chemistry, we have developed the first machine learning models that predict the likelihood of successful simulation outcomes. We train support vector machine and artificial neural network classifiers to predict simulation outcomes (i.e., geometry optimization result and degree of S 2 deviation) for a chosen electronic structure method based on chemical composition. For these static models, we achieve an area under the curve of at least 0.95, minimizing computational time spent on nonproductive simulations and therefore enabling efficient chemical space exploration. We introduce a metric of model uncertainty based on the distribution of points in the latent space to systematically improve model prediction confidence. In a complementary approach, we train a convolutional neural network classification model on simulation output electronic and geometric structure time series data. This dynamic model generalizes more readily than the static classifier by becoming more predictive as input simulation length increases. Finally, we describe approaches for using these models to enable autonomous job control in transition metal complex discovery. ©2019 American Chemical Society.
Subjects
Physical and Theoretical Chemistry
Computer Science Applications
MIT Department
Massachusetts Institute of Technology. Department of Chemical Engineering
Massachusetts Institute of Technology. Department of Chemistry
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1021/acs.jctc.9b00057
