Reproduce “Self-Directed Apple Tasting” 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

Tanmay Devale, Reed Phillips, Arvind Ramaswami

The paper studies a learner that chooses the order of a fixed set of cases, receiving the true label only when it predicts positive. In the realizable setting, optimal expected mistakes depend on VC dimension, the ∅-centered star number, and self-directed dimension: rates are constant, unbounded logarithmic, square-root up to logarithms, or linear. The logarithmic rate is not exact.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.