Reproduce “DynSHAP: Towards Explainable Dynamic Survival Analysis” 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
Nastasya Anokhina, Jonas Jürß, Pietro Liò
DynSHAP explains dynamic survival predictions by treating each binned time–feature pair as a SHAP player and estimating contributions at separate survival horizons. Its Temporal variant imputes missing pairs using learned temporal dependencies rather than independent background values. On synthetic data, it reduced attribution error for DySurv, but not DDH, where Kernel DynSHAP performed better.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.