A Caltech Library Service

Inverse Design of Solid-State Materials via a Continuous Representation

Noh, Juhwan and Kim, Jaehoon and Stein, Helge S. and Sanchez-Lengeling, Benjamin and Gregoire, John M. and Aspuru-Guzik, Alán and Jung, Yousung (2019) Inverse Design of Solid-State Materials via a Continuous Representation. Matter, 1 (5). pp. 1370-1384. ISSN 2590-2385. doi:10.1016/j.matt.2019.08.017.

[img] PDF (Document S1. Supplemental Experimental Procedures, Figures S1–S11, and Table S1) - Supplemental Material
See Usage Policy.


Use this Persistent URL to link to this item:


The non-serendipitous discovery of materials with targeted properties is the ultimate goal of materials research, but to date, materials design lacks the incorporation of all available knowledge to plan the synthesis of the next material. This work presents a framework for learning a continuous representation of materials and building a model for new discovery using latent space representation. The ability of autoencoders to generate experimental materials is demonstrated with vanadium oxides via rediscovery of experimentally known structures when the model was trained without them. Approximately 20,000 hypothetical materials are generated, leading to several completely new metastable V_xO_y materials that may be synthesizable. Comparison with genetic algorithms suggests computational efficiency of generative models that can explore chemical compositional space effectively by learning the distributions of known materials for crystal structure prediction. These results are an important step toward machine-learned inverse design of inorganic functional materials using generative models.

Item Type:Article
Related URLs:
URLURL TypeDescription
Noh, Juhwan0000-0003-1183-9955
Stein, Helge S.0000-0002-3461-0232
Gregoire, John M.0000-0002-2863-5265
Jung, Yousung0000-0003-2615-8394
Additional Information:© 2019 Elsevier. Received 22 July 2019, Revised 6 August 2019, Accepted 17 August 2019, Available online 2 October 2019. We acknowledge the support from the National Research Foundation of Korea (NRF-2017R1A2B3010176) and Korea Institute of Energy Technology Evaluation and Planning (KETEP-20188500000440) grants from the Korean Government, and a generous supercomputing time from Korea Insitute of Science and Technology Information (KISTI). H.S.S. and J.M.G. are supported through the Office of Science of the U.S. Department of Energy under award no. DE-SC0004993. A.A.-G. thanks the Canada 150 Research Chairs Program, Natural Resources Canada, and the Vannevar Bush Faculty Fellowship Program for support. A.A.-G. acknowledges the generous support of Anders G. Frøseth. Author Contributions: J.N., J.K., A.A.-G., and Y.J. designed the project. J.N. performed the machine-learning simulations, DFT calculations, and analyses. J.N. and Y.J. analyzed the results and wrote the manuscript. H.S.S. and J.M.G. assisted with data analysis and interpretation of the generated materials. B.S.-L. and A.A.-G. assisted with the machine-learning model construction. All authors contributed to the discussion and editing of the manuscript. Y.J. supervised the project. Data and Code Availability: The datasets used to train the model and the generated crystal structures are available at Source codes and trained parameters are available at The authors declare no competing interests.
Funding AgencyGrant Number
National Research Foundation of KoreaNRF-2017R1A2B3010176
Korea Institute of Energy Technology Evaluation and Planning20188500000440
Department of Energy (DOE)DE-SC0004993
Canada Research Chairs ProgramUNSPECIFIED
Natural Resources CanadaUNSPECIFIED
Vannevar Bush FellowshipUNSPECIFIED
Anders G. FrøsethUNSPECIFIED
Subject Keywords:MAP2: Benchmark; inverse design; generative model; machine learning; autoencoder; vanadium oxides; inorganic materials
Issue or Number:5
Record Number:CaltechAUTHORS:20191002-094950933
Persistent URL:
Official Citation:Juhwan Noh, Jaehoon Kim, Helge S. Stein, Benjamin Sanchez-Lengeling, John M. Gregoire, Alan Aspuru-Guzik, Yousung Jung, Inverse Design of Solid-State Materials via a Continuous Representation, Matter, Volume 1, Issue 5, 2019, Pages 1370-1384, ISSN 2590-2385, (
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:99012
Deposited By: George Porter
Deposited On:02 Oct 2019 20:22
Last Modified:16 Nov 2021 17:43

Repository Staff Only: item control page