A new multimodal representation learning framework explicitly uses informative missingness patterns in structured and textual clinical data via Bayesian filtering to improve treatment policy learning and mortality prediction on sepsis ICU datasets.
Lehman and Matthieu Komorowski and Matthieu Komorowski and Aldo Faisal and Leo Anthony Celi and David Sontag and Finale Doshi-Velez , title =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Learning Dynamic Representations and Policies from Multimodal Clinical Time-Series with Informative Missingness
A new multimodal representation learning framework explicitly uses informative missingness patterns in structured and textual clinical data via Bayesian filtering to improve treatment policy learning and mortality prediction on sepsis ICU datasets.