Embedding free multi hop retrieval achieves strong results with bm25 and jev
SAG2: Embedding-Free Multi-Hop Retrieval with BM25 and JEV
Computation and Language
Summary
Finding relevant information that requires linking multiple pieces of knowledge can be slow and complicated. The authors introduce SAG2, a new system that quickly finds connected information without precomputing complex data or using specialized servers. It relies on a fast keyword search method (BM25) combined with a smart decision process (JEV) done at search time. Tests show SAG2 matches or beats other methods while being easy and quick to set up.
What this means in practice
- •For information retrieval engineers: Build efficient multi-hop search systems that need no offline embedding computation or complex infrastructure.
- •For enterprise search teams: Deploy fast and accurate document retrieval for large internal document collections using a lightweight, single database approach.
Authors
Yuchao Wu
Abstract
Strong multi-hop retrieval should be practical to deploy without expensive corpus preparation or complex database infrastructure. We present SAG2, an embedding-free retrieval architecture built around a single SQLite database, BM25, and contextual evidence decisions from JEV. SAG2 follows the query-time association principle of SAG (SQL-Retrieval Augmented Generation), while replacing its event-entity and dense indexes with a lexical index over original passages. Selected passages supply terms for further retrieval, and newly retrieved evidence guides subsequent decisions. This moves semantic evidence selection to query time, eliminating offline embedding computation, LLM-based event-entity extraction, and dedicated database servers. The resulting system combines lightweight preparation with strong retrieval accuracy: on 1,000-question subsets of MuSiQue, 2WikiMultiHopQA, and HotpotQA, SAG2 achieves the highest Recall@5 and Recall@10 among the compared methods, with Recall@10 of 86.62%, 97.43%, and 99.00%. It also reaches 81.10% document Recall@10 over the 511,962-document EnterpriseRAG-Bench corpus. Indexing all 11,656 MuSiQue passages takes 2.36 seconds, produces a 25.87 MiB database, and requires no model calls. SAG2 demonstrates that accurate multi-hop evidence retrieval can be supported by compact, rapidly prepared infrastructure, making it a practical foundation for knowledge retrieval applications.