Reproduce “Hidden Data Challenges in Clinical Machine Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Carolin Cissée, Dennis Fast, Raneen Younis, Alexander Löser, Wolfgang Nejdl, Zahra Ahmadi
The authors audit how preprocessing choices in MIMIC-IV affect prediction tasks. They distinguish event time from information availability, and examine semantic and population alignment. In a 48-hour ICU mortality study, eight configurations have similar aggregate scores but different feature attributions; this shows predictive agreement need not mean models rely on same signals. Storetime remains only a proxy for clinician knowledge.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.