Reproduce “When, Where, How: Active In-Context Learning for Tabular Foundation Models” 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
Christian M. M. Frey
EPIC selects labels for a frozen tabular model by estimating how much each candidate’s unknown label would inform predictions on available target inputs. TANN offers a simpler alternative, choosing pool rows near the targets where the model is least certain. On 20 OpenML classification benchmarks, EPIC had the highest mean final Cohen’s κ, but its advantage over Hybrid did not meet the Holm-corrected 0.05 threshold.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.