Am I on the Right Track? What Can Predicted Query Performance Tell Us about the Search Behaviour of Agentic RAG

TL;DR

This paper investigates how query performance prediction can inform the search behavior of Agentic Retrieval-Augmented Generation models, showing that better query estimates correlate with higher answer quality and enabling adaptive retrieval strategies.

Contribution

It introduces the application of query performance prediction to Agentic RAG models, demonstrating its potential to improve answer quality and inform retrieval decisions.

Findings

Effective retrievers lead to higher answer quality.

QPP estimates correlate positively with final answer quality.

QPP can inform adaptive retrieval strategies.

Abstract

Agentic Retrieval-Augmented Generation (RAG) is a new paradigm where the reasoning model decides when to invoke a retriever (as a "tool") when answering a question. This paradigm, exemplified by recent research works such as Search-R1, enables the model to decide when to search and obtain external information. However, the queries generated by such Agentic RAG models and the role of the retriever in obtaining high-quality answers remain understudied. To this end, this initial study examines the applicability of query performance prediction (QPP) within the recent Agentic RAG models Search-R1 and R1-Searcher. We find that applying effective retrievers can achieve higher answer quality within a shorter reasoning process. Moreover, the QPP estimates of the generated queries, used as an approximation of their retrieval quality, are positively correlated with the quality of the final answer. Ultimately, our work is a step towards adaptive retrieval within Agentic RAG, where QPP is used to inform the model if the retrieved results are likely to be useful.