Reproduce “Retriever-Free Retrieval: Navigating Documents with Decision Models” 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
Chao Wan, Paul Jünger, Beixi Du, Bodong Liu, Navaneetha Krishnan, Kilian Q. Weinberger
DecisionRAG retrieves evidence by making parallel relevance judgments over candidates in a corpus hierarchy, rather than relying on a conventional first-stage retriever or generating navigation text. Its pretrained, task-untrained Jev model reaches 47.4% R@1 on MultiDoc2Dial and 65.4% Hit@1 on Natural Questions, exceeding dense retrieval plus reranking; results depend on hierarchy, and pruning can hide evidence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.