Reproduce “In Defense of Core-set: A Density-aware Core-set Selection for Active Learning” 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
Yeachan Kim, Bonggun Shin
DACS steers core-set selection toward sparse feature regions by estimating density with locality-sensitive hashing and assigning more queries to smaller density groups. On constructed Repeated MNIST, at the reported 2.0%-labeled endpoint, it reaches 97.6% test accuracy versus 96.7% for CDAL. Sparse-only selection generally beats dense-only in tested settings, though not at RMNIST’s first displayed point.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.