Search papers, labs, and topics across Lattice.
This paper introduces $\Psi$-RAG, a novel tree-based retrieval-augmented generation framework designed for cross-document multi-hop question answering. It addresses limitations of existing Tree-RAG methods by using an iterative "merging and collapse" process to build a hierarchical abstract tree index that adapts to data distributions. Experiments on cross-document multi-hop QA benchmarks demonstrate that $\Psi$-RAG significantly outperforms existing methods like RAPTOR and HippoRAG, achieving up to a 25.9% improvement in average F1 score.
Tree-based RAG gets a major upgrade: $\Psi$-RAG's adaptive hierarchical index and multi-granular retrieval agent leapfrog existing methods on complex, cross-document reasoning tasks.
Retrieval-augmented generation (RAG) enhances large language models with external knowledge, and tree-based RAG organizes documents into hierarchical indexes to support queries at multiple granularities. However, existing Tree-RAG methods designed for single-document retrieval face critical challenges in scaling to cross-document multi-hop questions: (1) poor distribution adaptability, where $k$-means clustering introduces noise due to rigid distribution assumptions; (2) structural isolation, as tree indexes lack explicit cross-document connections; and (3) coarse abstraction, which obscures fine-grained details. To address these limitations, we propose $\Psi$-RAG, a tree-RAG framework with two key components. First, a hierarchical abstract tree index built through an iterative"merging and collapse"process that adapts to data distributions without a priori assumption. Second, a multi-granular retrieval agent that intelligently interacts with the knowledge base with reorganized queries and an agent-powered hybrid retriever. $\Psi$-RAG supports diverse tasks from token-level question answering to document-level summarization. On cross-document multi-hop QA benchmarks, it outperforms RAPTOR by 25.9% and HippoRAG 2 by 7.4% in average F1 score. Code is available at https://github.com/Newiz430/Psi-RAG.
