Reproduce “Structure-Aware Graph Abstention for Reliable Selective Forecasting” 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
Jianxiang Xie, Belal Alsinglawi
SASF adds a relational check to multivariate forecast release decisions, scoring whether predicted channel trajectories remain consistent with a learned global graph. It complements whole-forecast risk scoring rather than replacing it. On PatchTST Weather, validation-set thresholds transferred to test at 50% target coverage yield selective MSE 0.1443, versus 0.1792 for TEM, in one seed-2024 comparison; gains are not universal.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.