From development to deployment: dataset shift, causality, and shift-stable models in health AI

Biostatistics. 2020 Apr 1;21(2):345-352. doi: 10.1093/biostatistics/kxz041.

Authors

Adarsh Subbaswamy¹, Suchi Saria²

Affiliations

¹ Department of Computer Science, Johns Hopkins University, 160 Malone Hall, 3400 N. Charles Street, Baltimore, MD, USA.
² Department of Computer Science; Department of Applied Math & Statistics, and Department of Health Policy & Management, Johns Hopkins University, 160 Malone Hall, 3400 N. Charles Street, Baltimore, MD, USA.

PMID: 31742354
DOI: 10.1093/biostatistics/kxz041

No abstract available

Keywords: Causal inference; Dataset shift; Generalizability; Machine learning.

MeSH terms

Artificial Intelligence*
Biomedical Research* / methods
Biomedical Research* / standards
Biostatistics* / methods
Causality
Datasets as Topic*
Humans
Models, Theoretical*
Mortality