Reproduce “Gone Fishing: Neural Active Learning with Fisher Embeddings” 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

Jordan T. Ash, Surbhi Goel, Akshay Krishnamurthy, Sham Kakade

BAIT selects neural-network label batches by estimating how candidate labels inform model parameters, using Fisher information from candidates and the unlabeled pool. It recomputes this information as the model’s representation changes. Across classification and regression experiments, it performs as well as or better than methods; its clearest advantage is on fixed random-feature MNIST, a constructed setting that does not establish universal superiority.

Summary by alphaXiv.

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