A Domain Incremental Continual Learning Benchmark for ICU Time Series Model Transportability

Ryan King, Conrad Krueger, Ethan Veselka, Tianbao Yang, and Bobak J. Mortazavi · 2026

Read A Domain Incremental Continual Learning Benchmark for ICU Time Series Model Transportability by Ryan King, Conrad Krueger, Ethan Veselka, Tianbao Yang, and Bobak J. Mortazavi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-03832.txt.

Description

Ryan King and four co-authors build a benchmark for moving intensive care prediction models between hospitals. Training clinical models takes data, labeling, and computing power that small hospitals may lack, so transferring a model from a large hospital is attractive. Yet models trained at one site often generalize poorly, because measurement patterns differ across United States regions.

Their benchmark frames regional transfer as domain-incremental continual learning: a model must absorb each new region's data while retaining what the original domain taught it, without catastrophic forgetting. The benchmark tests transportability of intensive care time series models across regions, giving hospitals and researchers a shared yardstick for judging whether a transferred model is safe to adapt locally.

Catalog record

CREATORRyan King, Conrad Krueger, Ethan Veselka, Tianbao Yang, and Bobak J. Mortazavi
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.03832

Browse the full library · About the Tower of Babel