Reproduce “Generalized Neural Distributional Regression” 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
Natan Hilario da Silva, Vicente Garibay Cancho, Adriano Kamimura Suzuki
GNDR predicts distributional outcomes from complex inputs and supports inference by freezing neural features before estimating a final layer. This permits Fisher-information and Delta-method uncertainty conditional on those features. In hospital-stay data, model (d) had a held-out Anderson–Darling statistic of 0.1585, versus 0.6846 for model (b), indicating better residual calibration in that comparison. UTKFace test interval coverage was about 89.35%, below nominal 95%.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.