Reproduce “Error-Correcting Factorization” 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

Miguel Angel Bautista, Oriol Pujol, Fernando de la Torre, Sergio Escalera

ECF designs codes for classifying four sports-ball categories. It uses a design matrix to allocate pairwise protection toward classes that are hard to distinguish, rather than relying on a global correction measure. In experiments with RBF-kernel SVMs, ECF-H wins 62.1% by support-vector count and 70.0% by classifier count; these are win rates, not accuracy gains, and do not directly test binary errors.

Summary by alphaXiv.

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