Reproduce “Adversarial Online Collaborative Filtering” 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
Stephen Pasteris, Fabio Vitale, Mark Herbster, Claudio Gentile, Andre' Panisson
ORCA treats recommendation as sequential collaboration with no-repeat items, learning from likes and dislikes. It uses shared user and item preference patterns to achieve expected regret of order min(C,D)(M+N) in perfectly clustered settings. A matching lower bound applies when min(C,D) is at most square root of min(M,N). On MovieLens, ORCA* scored 0.746 versus WRMF’s 0.597 in a 50-item test with random user order.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.