TaSR-RAG: Taxonomy-guided Structured Reasoning for Retrieval-Augmented Generation

Researchers propose TaSR-RAG, a new framework that improves Retrieval-Augmented Generation systems by using taxonomy-guided structured reasoning for better evidence selection. The system decomposes complex questions into triple sub-queries and performs step-wise evidence matching, achieving up to 14% performance improvements over existing RAG baselines on multi-hop question answering benchmarks.

• →TaSR-RAG addresses limitations in current RAG systems including redundant context and brittle multi-hop reasoning.
• →The framework represents queries and documents as relational triples with a lightweight two-level taxonomy structure.
• →It decomposes complex questions into ordered sequences of triple sub-queries for better evidence selection.
• →The system maintains an explicit entity binding table to reduce entity conflation without requiring graph construction.
• →Experimental results show consistent 14% improvements over strong RAG baselines with clearer evidence attribution.