English

TreeHop: Generate and Filter Next Query Embeddings Efficiently for Multi-hop Question Answering

Information Retrieval 2025-05-01 v2 Artificial Intelligence Human-Computer Interaction Machine Learning

Abstract

Retrieval-augmented generation (RAG) systems face significant challenges in multi-hop question answering (MHQA), where complex queries require synthesizing information across multiple document chunks. Existing approaches typically rely on iterative LLM-based query rewriting and routing, resulting in high computational costs due to repeated LLM invocations and multi-stage processes. To address these limitations, we propose TreeHop, an embedding-level framework without the need for LLMs in query refinement. TreeHop dynamically updates query embeddings by fusing semantic information from prior queries and retrieved documents, enabling iterative retrieval through embedding-space operations alone. This method replaces the traditional "Retrieve-Rewrite-Vectorize-Retrieve" cycle with a streamlined "Retrieve-Embed-Retrieve" loop, significantly reducing computational overhead. Moreover, a rule-based stop criterion is introduced to further prune redundant retrievals, balancing efficiency and recall rate. Experimental results show that TreeHop rivals advanced RAG methods across three open-domain MHQA datasets, achieving comparable performance with only 5\%-0.4\% of the model parameter size and reducing the query latency by approximately 99\% compared to concurrent approaches. This makes TreeHop a faster and more cost-effective solution for deployment in a range of knowledge-intensive applications. For reproducibility purposes, codes and data are available here: https://github.com/allen-li1231/TreeHop-RAG.

Keywords

Cite

@article{arxiv.2504.20114,
  title  = {TreeHop: Generate and Filter Next Query Embeddings Efficiently for Multi-hop Question Answering},
  author = {Zhonghao Li and Kunpeng Zhang and Jinghuai Ou and Shuliang Liu and Xuming Hu},
  journal= {arXiv preprint arXiv:2504.20114},
  year   = {2025}
}

Comments

9 pages

R2 v1 2026-06-28T23:14:17.570Z