CaltechAUTHORS
Published June 2019 | Version public
Journal Article Metadata-only

CRYSTAL: a multi-agent AI system for automated mapping of materials' crystal structures

Creators

  • 1. ROR icon Cornell University
  • 2. ROR icon Purdue University West Lafayette
  • 3. ROR icon Joint Center for Artificial Photosynthesis
  • 4. ROR icon Toyota Motor Corporation (United States)

Abstract

We introduce CRYSTAL, a multi-agent AI system for crystal-structure phase mapping. CRYSTAL is the first system that can automatically generate a portfolio of physically meaningful phase diagrams for expert-user exploration and selection. CRYSTAL outperforms previous methods to solve the example Pd-Rh-Ta phase diagram, enabling the discovery of a mixed-intermetallic methanol oxidation electrocatalyst. The integration of multiple data-knowledge sources and learning and reasoning algorithms, combined with the exploitation of problem decompositions, relaxations, and parallelism, empowers AI to supersede human scientific data interpretation capabilities and enable otherwise inaccessible scientific discovery in materials science and beyond.

Additional Information

© 2019 Materials Research Society. Received 18 January 2019; accepted 8 April 2019. This work was supported by NSF awards CCF-1522054 and CNS-0832782 (Expeditions), CNS-1059284 (Infrastructure), and IIS-1344201 (INSPIRE); ARO awards W911NF-14-1-0498 and W911NF-17-1-0187; AFOSR Multidisciplinary University Research Initiatives (MURI) Program FA9550-18-1-0136, Toyota Research Institute award; and US DOE Award No. DE-SC0004993. Use of SSRL is supported by DOE Contract No. DE-AC02-76SF00515. Use of CHESS is supported by the NSF award DMR-1332208. The authors thank A. Mehta, D. G. Van Campen, M. Tague, and D. Dale for assistance with data collection. Author contributions: The authors' contributions are as follows: C.P.G., R.B.vD., and J.M.G. conceived and managed the project. C.P.G. conceived CRYSTAL'S multiple knowledge source approach. J.Ba., JBj., C.P.G., and Y.X. designed the bots' algorithms. J.Ba., C.P.G., J.M.G., B.H.R., and Y.X. designed the Diagram Rendering bot. J.Ba. implemented the IAFD bots, phase matching bot, and phase analysis bot. B.H.R. implemented the diagram rendering algorithm, and Analysis & Reporting and Visualizer & Interface bots. S.K. performed the comparison with NMFK. R.A.B. assisted with programming in several components of CRYSTAL. R.B.vD. and J.M.G. acquired Pd-Rh-Ta data, and S.K.S. and J.M.G. acquired Nb-Cu-V data with assistance as noted in the Acknowledgments. S.K.S. and J.M.G. served as human experts for both systems. C.P.G. and J.M.G. were the primary authors of the manuscript. S.A., J.Ba., J.Bj., C.P.G., J.G.M., B.H.R., and Y.X. were the primary authors of the Methods and Supplementary Information. Data availability: The raw data for the Pd-Rh-Ta along with CRYSTAL's results and reports will be available at http://www.udiscover.it/resources/data/. Further documentation and source code for IAFD can be found at http://www.udiscover.it/resources/software/.

Additional details

Identifiers

Eprint ID
97941
DOI
10.1557/mrc.2019.50
Resolver ID
CaltechAUTHORS:20190816-090926308

Related works

Describes
10.1557/mrc.2019.50 (DOI)

Funding

NSF
CCF-1522054
NSF
CNS-0832782
NSF
CNS-1059284
NSF
IIS-1344201
Army Research Office (ARO)
W911NF-14-1-0498
Army Research Office (ARO)
W911NF-17-1-0187
Air Force Office of Scientific Research (AFOSR)
FA9550-18-1-0136
Toyota Research Institute
Department of Energy (DOE)
DE-SC0004993
Department of Energy (DOE)
DE-AC02-76SF00515
NSF
DMR-1332208

Dates

Created
2019-08-16
Created from EPrint's datestamp field
Updated
2021-11-16
Created from EPrint's last_modified field

Caltech Custom Metadata

Caltech groups
JCAP