Adaptive Retrieval for Reasoning-Intensive Retrieval

TL;DR

This paper introduces REPAIR, a framework that enhances reasoning-intensive retrieval by using reasoning plans as feedback to adaptively retrieve supportive documents, improving performance on complex QA tasks.

Contribution

REPAIR enables mid-course correction in retrieval pipelines by leveraging reasoning plans as dense feedback signals, addressing recall limitations in existing methods.

Findings

REPAIR outperforms baselines by 5.6% on reasoning-intensive retrieval tasks.

Using reasoning plans as feedback improves retrieval of supportive documents.

Adaptive retrieval with REPAIR enhances complex question answering performance.

Abstract

We study leveraging adaptive retrieval to ensure sufficient "bridge" documents are retrieved for reasoning-intensive retrieval. Bridge documents are those that contribute to the reasoning process yet are not directly relevant to the initial query. While existing reasoning-based reranker pipelines attempt to surface these documents in ranking, they suffer from bounded recall. Naive solution with adaptive retrieval into these pipelines often leads to planning error propagation. To address this, we propose REPAIR, a framework that bridges this gap by repurposing reasoning plans as dense feedback signals for adaptive retrieval. Our key distinction is enabling mid-course correction during reranking through selective adaptive retrieval, retrieving documents that support the pivotal plan. Experimental results on reasoning-intensive retrieval and complex QA tasks demonstrate that our method outperforms existing baselines by 5.6%pt.