Reproduce “ZeroRAG: Adaptive Semantic Peer Routing for Decentralized Retrieval-Augmented Generation” 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
Petru Neague, Quinten Stokkink, Martijn de Vos, Rafael Pires, Mathis Randl, Anne-Marie Kermarrec, Johan Pouwelse
ZeroRAG routes retrieval queries among peers without a central index, while documents remain with their owners. Peers use advertised document embeddings, recent query embeddings, and referrals to guide searches, then adapt local connections from results. In PubMedQA, it reaches 93.0% recall and 77% answer accuracy with five retrieved documents, below centralized HNSW’s 97.3% and 79%.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.