Can machine learning identify the next high-temperature superconductor? Examining extrapolation performance for materials discovery
Journal Article
·
· Molecular Systems Design & Engineering
- Citrine Informatics, USA
- University of Chicago, USA, Argonne National Laboratory, USA
- Stanford University, USA
- National Institute of Standards and Technology, USA
- SLAC National Accelerator Laboratory, USA
Traditional machine learning (ML) metrics overestimate model performance for materials discovery. We introduce (1) leave-one-cluster-out cross-validation (LOCO CV) and (2) a simple nearest-neighbor benchmark to show that model performance in discovery applications strongly depends on the problem, data sampling, and extrapolation. Our results suggest that ML-guided iterative experimentation may outperform standard high-throughput screening for discovering breakthrough materials like high-Tc superconductors with ML.
- Research Organization:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Organization:
- USDOE; National Institute of Standards and Technology (NIST)
- Grant/Contract Number:
- AC02-06CH11357
- OSTI ID:
- 1464980
- Alternate ID(s):
- OSTI ID: 1484501
- Journal Information:
- Molecular Systems Design & Engineering, Journal Name: Molecular Systems Design & Engineering Vol. 3 Journal Issue: 5; ISSN 2058-9689
- Publisher:
- Royal Society of Chemistry (RSC)Copyright Statement
- Country of Publication:
- United Kingdom
- Language:
- English
Cited by: 90 works
Citation information provided by
Web of Science
Web of Science
Similar Records
Machine Learning for the Discovery, Design, and Engineering of Materials
Machine Learning Applied to Zeolite Synthesis: The Missing Link for Realizing High-Throughput Discovery
Multifidelity Information Fusion with Machine Learning: A Case Study of Dopant Formation Energies in Hafnia
Journal Article
·
Fri Jun 10 00:00:00 EDT 2022
· Annual Review of Chemical and Biomolecular Engineering
·
OSTI ID:1464980
Machine Learning Applied to Zeolite Synthesis: The Missing Link for Realizing High-Throughput Discovery
Journal Article
·
Wed Sep 25 00:00:00 EDT 2019
· Accounts of Chemical Research
·
OSTI ID:1464980
Multifidelity Information Fusion with Machine Learning: A Case Study of Dopant Formation Energies in Hafnia
Journal Article
·
Tue Apr 16 00:00:00 EDT 2019
· ACS Applied Materials and Interfaces
·
OSTI ID:1464980
+1 more