Multimodal Survival Analysis with Locally Deployable Large Language Models

Moritz Gögl, Christopher Yau · 2026

Read Multimodal Survival Analysis with Locally Deployable Large Language Models by Moritz Gögl, Christopher Yau (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-2603-22158.txt.

Description

Gögl and Yau study survival prediction that mixes three kinds of patient data. The inputs are clinical text, tabular variables, and genomic profiles. Their method runs on local language models, not cloud services. That suits hospitals with tight budgets and strict privacy rules. The method estimates calibrated survival chances for each patient. It also writes short prognosis text grounded in the evidence. Training uses teacher-student distillation and a principled fusion of the three inputs. On a TCGA cohort, the authors report it beats standard baselines.

Catalog record

CREATORMoritz Gögl, Christopher Yau
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27847
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.22158

Browse the full library · About the Tower of Babel