skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning

Abstract

Abstract The current predictive modeling techniques applied to Density Functional Theory (DFT) computations have helped accelerate the process of materials discovery by providing significantly faster methods to scan materials candidates, thereby reducing the search space for future DFT computations and experiments. However, in addition to prediction error against DFT-computed properties, such predictive models also inherit the DFT-computation discrepancies against experimentally measured properties. To address this challenge, we demonstrate that using deep transfer learning, existing large DFT-computational data sets (such as the Open Quantum Materials Database (OQMD)) can be leveraged together with other smaller DFT-computed data sets as well as available experimental observations to build robust prediction models. We build a highly accurate model for predicting formation energy of materials from their compositions; using an experimental data set of $$1,643$$ 1 , 643 observations, the proposed approach yields a mean absolute error (MAE) of $$0.07$$ 0.07 eV/atom, which is significantly better than existing machine learning (ML) prediction modeling based on DFT computations and is comparable to the MAE of DFT-computation itself.

Authors:
; ; ; ; ; ;
Publication Date:
Research Org.:
Northwestern Univ., Evanston, IL (United States); National Inst. of Standards and Technology (NIST), Gaithersburg, MD (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR); Dept. of Commerce (United States)
OSTI Identifier:
1619594
Alternate Identifier(s):
OSTI ID: 1624217
Grant/Contract Number:  
SC0014330; SC0019358; 70NANB19H005
Resource Type:
Published Article
Journal Name:
Nature Communications
Additional Journal Information:
Journal Name: Nature Communications Journal Volume: 10 Journal Issue: 1; Journal ID: ISSN 2041-1723
Publisher:
Nature Publishing Group
Country of Publication:
United Kingdom
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING; 36 MATERIALS SCIENCE; 37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; theory and computation; materials chemistry; inorganic chemistry; density functional theory

Citation Formats

Jha, Dipendra, Choudhary, Kamal, Tavazza, Francesca, Liao, Wei-keng, Choudhary, Alok, Campbell, Carelyn, and Agrawal, Ankit. Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning. United Kingdom: N. p., 2019. Web. https://doi.org/10.1038/s41467-019-13297-w.
Jha, Dipendra, Choudhary, Kamal, Tavazza, Francesca, Liao, Wei-keng, Choudhary, Alok, Campbell, Carelyn, & Agrawal, Ankit. Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning. United Kingdom. https://doi.org/10.1038/s41467-019-13297-w
Jha, Dipendra, Choudhary, Kamal, Tavazza, Francesca, Liao, Wei-keng, Choudhary, Alok, Campbell, Carelyn, and Agrawal, Ankit. Fri . "Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning". United Kingdom. https://doi.org/10.1038/s41467-019-13297-w.
@article{osti_1619594,
title = {Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning},
author = {Jha, Dipendra and Choudhary, Kamal and Tavazza, Francesca and Liao, Wei-keng and Choudhary, Alok and Campbell, Carelyn and Agrawal, Ankit},
abstractNote = {Abstract The current predictive modeling techniques applied to Density Functional Theory (DFT) computations have helped accelerate the process of materials discovery by providing significantly faster methods to scan materials candidates, thereby reducing the search space for future DFT computations and experiments. However, in addition to prediction error against DFT-computed properties, such predictive models also inherit the DFT-computation discrepancies against experimentally measured properties. To address this challenge, we demonstrate that using deep transfer learning, existing large DFT-computational data sets (such as the Open Quantum Materials Database (OQMD)) can be leveraged together with other smaller DFT-computed data sets as well as available experimental observations to build robust prediction models. We build a highly accurate model for predicting formation energy of materials from their compositions; using an experimental data set of $$1,643$$ 1 , 643 observations, the proposed approach yields a mean absolute error (MAE) of $$0.07$$ 0.07 eV/atom, which is significantly better than existing machine learning (ML) prediction modeling based on DFT computations and is comparable to the MAE of DFT-computation itself.},
doi = {10.1038/s41467-019-13297-w},
journal = {Nature Communications},
number = 1,
volume = 10,
place = {United Kingdom},
year = {2019},
month = {11}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record
https://doi.org/10.1038/s41467-019-13297-w

Citation Metrics:
Cited by: 11 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

From Organized High-Throughput Data to Phenomenological Theory using Machine Learning: The Example of Dielectric Breakdown
journal, February 2016


Perspective: Materials informatics and big data: Realization of the “fourth paradigm” of science in materials science
journal, April 2016

  • Agrawal, Ankit; Choudhary, Alok
  • APL Materials, Vol. 4, Issue 5
  • DOI: 10.1063/1.4946894

MoleculeNet: a benchmark for molecular machine learning
journal, January 2018

  • Wu, Zhenqin; Ramsundar, Bharath; Feinberg, Evan N.
  • Chemical Science, Vol. 9, Issue 2
  • DOI: 10.1039/C7SC02664A

Deep learning
journal, May 2015

  • LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
  • Nature, Vol. 521, Issue 7553
  • DOI: 10.1038/nature14539

Elastic properties of bulk and low-dimensional materials using van der Waals density functional
journal, July 2018


Predicting density functional theory total energies and enthalpies of formation of metal-nonmetal compounds by linear regression
journal, February 2016


Diffusion of Ge below the Si(100) Surface: Theory and Experiment
journal, March 2000


Rapid and Accurate Machine Learning Recognition of High Performing Metal Organic Frameworks for CO 2 Capture
journal, August 2014

  • Fernandez, Michael; Boyd, Peter G.; Daff, Thomas D.
  • The Journal of Physical Chemistry Letters, Vol. 5, Issue 17
  • DOI: 10.1021/jz501331m

Including crystal structure attributes in machine learning models of formation energies via Voronoi tessellations
journal, July 2017


Dropout from Higher Education: A Theoretical Synthesis of Recent Research
journal, March 1975


Comparison of theory with quenching experiments for the entropy and enthalpy of vacancy formation in Si and Ge
journal, October 1976


Machine learning of molecular electronic properties in chemical compound space
journal, September 2013


Exploration of data science techniques to predict fatigue strength of steel from composition and processing parameters
journal, April 2014

  • Agrawal, Ankit; Deshpande, Parijat D.; Cecen, Ahmet
  • Integrating Materials and Manufacturing Innovation, Vol. 3, Issue 1
  • DOI: 10.1186/2193-9772-3-8

Optimizing transition states via kernel-based machine learning
journal, May 2012

  • Pozun, Zachary D.; Hansen, Katja; Sheppard, Daniel
  • The Journal of Chemical Physics, Vol. 136, Issue 17
  • DOI: 10.1063/1.4707167

SchNet – A deep learning architecture for molecules and materials
journal, June 2018

  • Schütt, K. T.; Sauceda, H. E.; Kindermans, P. -J.
  • The Journal of Chemical Physics, Vol. 148, Issue 24
  • DOI: 10.1063/1.5019779

Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
journal, May 2016

  • Shin, Hoo-Chang; Roth, Holger R.; Gao, Mingchen
  • IEEE Transactions on Medical Imaging, Vol. 35, Issue 5
  • DOI: 10.1109/TMI.2016.2528162

Computational screening of high-performance optoelectronic materials using OptB88vdW and TB-mBJ formalisms
journal, May 2018

  • Choudhary, Kamal; Zhang, Qin; Reid, Andrew C. E.
  • Scientific Data, Vol. 5, Issue 1
  • DOI: 10.1038/sdata.2018.82

Nobel Lecture: Electronic structure of matter—wave functions and density functionals
journal, October 1999


Properties of intrinsic point defects in silicon determined by zinc diffusion experiments under nonequilibrium conditions
journal, December 1995


Predicting the Thermodynamic Stability of Solids Combining Density Functional Theory and Machine Learning
journal, May 2017


Matrix- and tensor-based recommender systems for the discovery of currently unknown inorganic compounds
journal, January 2018


Understanding NOx formation in nonpremixed flames: Experiments and modeling
journal, January 1995


Extracting Grain Orientations from EBSD Patterns of Polycrystalline Materials Using Convolutional Neural Networks
journal, October 2018

  • Jha, Dipendra; Singh, Saransh; Al-Bahrani, Reda
  • Microscopy and Microanalysis, Vol. 24, Issue 5
  • DOI: 10.1017/S1431927618015131

Accelerating materials property predictions using machine learning
journal, September 2013

  • Pilania, Ghanshyam; Wang, Chenchen; Jiang, Xun
  • Scientific Reports, Vol. 3, Issue 1
  • DOI: 10.1038/srep02810

The Open Quantum Materials Database (OQMD): assessing the accuracy of DFT formation energies
journal, December 2015


How Chemical Composition Alone Can Predict Vibrational Free Energies and Entropies of Solids
journal, July 2017


Matminer: An open source toolkit for materials data mining
journal, September 2018


Machine learning with force-field-inspired descriptors for materials: Fast screening and mapping energy landscape
journal, August 2018


Toward Computational Materials Design: The Impact of Density Functional Theory on Materials Research
journal, September 2006

  • Hafner, Jürgen; Wolverton, Christopher; Ceder, Gerbrand
  • MRS Bulletin, Vol. 31, Issue 9
  • DOI: 10.1557/mrs2006.174

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation
journal, July 2013

  • Jain, Anubhav; Ong, Shyue Ping; Hautier, Geoffroy
  • APL Materials, Vol. 1, Issue 1
  • DOI: 10.1063/1.4812323

An introduction to ROC analysis
journal, June 2006


Recent progress in metallurgical thermochemistry
journal, January 1969


Deep materials informatics: Applications of deep learning in materials science
journal, June 2019


Representation of compounds for machine-learning prediction of physical properties
journal, April 2017


Materials Design and Discovery with High-Throughput Density Functional Theory: The Open Quantum Materials Database (OQMD)
journal, September 2013


Formation enthalpies by mixing GGA and GGA + U calculations
journal, July 2011


A high-throughput infrastructure for density functional theory calculations
journal, June 2011


High-Throughput Machine-Learning-Driven Synthesis of Full-Heusler Compounds
journal, October 2016


AFLOWLIB.ORG: A distributed materials properties repository from high-throughput ab initio calculations
journal, June 2012


Data-Driven Model for Estimation of Friction Coefficient Via Informatics Methods
journal, May 2012


A Survey on Transfer Learning
journal, October 2010

  • Pan, Sinno Jialin; Yang, Qiang
  • IEEE Transactions on Knowledge and Data Engineering, Vol. 22, Issue 10
  • DOI: 10.1109/TKDE.2009.191

Machine Learning Energies of 2 Million Elpasolite ( A B C 2 D 6 ) Crystals
journal, September 2016


IRNet: A General Purpose Deep Residual Regression Framework for Materials Discovery
conference, July 2019

  • Jha, Dipendra; Ward, Logan; Yang, Zijiang
  • KDD '19: The 25th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining
  • DOI: 10.1145/3292500.3330703

Machine-learning-assisted materials discovery using failed experiments
journal, May 2016

  • Raccuglia, Paul; Elbert, Katherine C.; Adler, Philip D. F.
  • Nature, Vol. 533, Issue 7601
  • DOI: 10.1038/nature17439

A general-purpose machine learning framework for predicting properties of inorganic materials
journal, August 2016


Universal fragment descriptors for predicting properties of inorganic crystals
journal, June 2017

  • Isayev, Olexandr; Oses, Corey; Toher, Cormac
  • Nature Communications, Vol. 8, Issue 1
  • DOI: 10.1038/ncomms15679

Combinatorial screening for new materials in unconstrained composition space with machine learning
journal, March 2014