Reproduce “Conformal Prediction for Distribution-to-Distribution 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
Trung-Khang Tran, Tuan A. Hoang, Viet-Hoang Tran, Tan M. Nguyen
The method makes conformal prediction usable for distribution-to-distribution regression by representing output distributions with finitely many orthonormal-basis coefficients, calibrating discrepancies in that space, and mapping prediction sets back to functions. These ellipsoids have tractable volume and uniform sampling. Coverage holds for truncated empirical outputs under exchangeability; adaptive conditional coverage is asymptotic, with deviations at most 0.06 in one synthetic setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.