Retro*: Optimizing LLMs for Reasoning-Intensive Document Retrieval

• URL: http://arxiv.org/abs/2509.24869v2
• Date: Sun, 12 Oct 2025 09:37:17 GMT
• Title: Retro*: Optimizing LLMs for Reasoning-Intensive Document Retrieval
• Authors: Junwei Lan, Jianlyu Chen, Zheng Liu, Chaofan Li, Siqi Bao, Defu Lian,
• Abstract summary: Retro* is a novel approach for reasoning-intensive document retrieval.<n>We introduce a rubric-based relevance scoring mechanism, enabling the model to reason about the relationship between a task and a document.<n>Our experiments show that Retro* outperforms existing document retrieval methods with notable advantages.
• Score: 44.680580989270965
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: With the growing popularity of LLM agents and RAG, it has become increasingly important to retrieve documents that are essential for solving a task, even when their connection to the task is indirect or implicit. Addressing this problem requires fine-grained reasoning to accurately assess the relevance between the task and each candidate document. This capability, however, poses a significant challenge for existing IR techniques. Despite recent progress in reasoning-enhanced IR, existing approaches still face significant challenges in applicability, scalability, and efficiency. In this work, we propose Retro*, a novel approach for reasoning-intensive document retrieval. Our method introduces a rubric-based relevance scoring mechanism, enabling the model to reason about the relationship between a task and a document based on explicitly defined criteria, whereby producing a fine-grained, interpretable relevance score. Retro* also supports test-time scaling by combining multiple reasoning trajectories via score integration, which produces more reliable relevance estimates. To optimize Retro*'s reasoning capabilities, we introduce a novel reinforcement learning algorithm tailored for its relevance scoring mechanism, which employs two composite rewards to fully exploit the trajectories of each training sample. Our experiments show that Retro* outperforms existing document retrieval methods with notable advantages, leading to state-of-the-art performance on the BRIGHT benchmark.

Related papers

• Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration [49.9937230730202]
  We propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention.<n>Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories.<n>We show that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales.
  arXiv  Detail & Related papers  (2026-02-03T15:32:09Z)
• Multi-hop Reasoning via Early Knowledge Alignment [68.28168992785896]
  Early Knowledge Alignment (EKA) aims to align Large Language Models with contextually relevant retrieved knowledge.<n>EKA significantly improves retrieval precision, reduces cascading errors, and enhances both performance and efficiency.<n>EKA proves effective as a versatile, training-free inference strategy that scales seamlessly to large models.
  arXiv  Detail & Related papers  (2025-12-23T08:14:44Z)
• RouteRAG: Efficient Retrieval-Augmented Generation from Text and Graph via Reinforcement Learning [69.87510139069218]
  Retrieval-Augmented Generation (RAG) integrates non-parametric knowledge into Large Language Models (LLMs)<n>Recent progress has advanced text-based RAG to multi-turn reasoning through Reinforcement Learning (RL)<n>We introduce model, an RL-based framework that enables LLMs to perform multi-turn and adaptive graph-text hybrid RAG.
  arXiv  Detail & Related papers  (2025-12-10T10:05:31Z)
• Embedding-Based Context-Aware Reranker [11.885086835801523]
  Retrieval-Augmented Generation (RAG) systems rely on retrieving relevant evidence from a corpus to support downstream generation.<n>We propose Embedding-Based Context-Aware Reranker (EBCAR), a lightweight reranking framework operating directly on embeddings of retrieved passages.<n>We evaluate EBCAR against SOTA rerankers on the ConTEB benchmark, demonstrating its effectiveness for information retrieval requiring cross-passage inference.
  arXiv  Detail & Related papers  (2025-10-15T09:14:04Z)
• Improving Context Fidelity via Native Retrieval-Augmented Reasoning [35.50952279309109]
  Large language models (LLMs) often struggle with context fidelity, producing inconsistent answers when responding to questions based on provided information.<n>We propose CARE, a novel native retrieval-augmented reasoning framework that teaches LLMs to explicitly integrate in-context evidence within their reasoning process with the model's own retrieval capabilities.<n>Our method requires limited labeled evidence data while significantly enhancing both retrieval accuracy and answer generation performance through strategically retrieved in-context tokens in the reasoning chain.
  arXiv  Detail & Related papers  (2025-09-17T04:28:07Z)
• DIVER: A Multi-Stage Approach for Reasoning-intensive Information Retrieval [36.38599923075882]
  DIVER is a retrieval pipeline designed for reasoning-intensive information retrieval.<n>It consists of four components: the document preprocessing stage, the query expansion stage, the retrieval stage and the reranking stage.<n>On the BRIGHT benchmark, DIVER achieves state-of-the-art nDCG@10 scores of 45.8 overall and 28.9 on original queries, consistently outperforming competitive reasoning-aware models.
  arXiv  Detail & Related papers  (2025-08-11T13:57:49Z)
• FrugalRAG: Learning to retrieve and reason for multi-hop QA [10.193015391271535]
  Large-scale fine-tuning is not needed to improve RAG metrics.<n>Supervised and RL-based fine-tuning can help RAG from the perspective of frugality.
  arXiv  Detail & Related papers  (2025-07-10T11:02:13Z)
• Breaking the Lens of the Telescope: Online Relevance Estimation over Large Retrieval Sets [14.494301139974455]
  We propose a novel paradigm for re-ranking called online relevance estimation.<n>Online relevance estimation continuously updates relevance estimates for a query throughout the ranking process.<n>We validate our approach on TREC benchmarks under two scenarios: hybrid retrieval and adaptive retrieval.
  arXiv  Detail & Related papers  (2025-04-12T22:05:50Z)
• Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval [49.669503570350166]
  Generative information retrieval (GenIR) is a promising neural retrieval paradigm that formulates document retrieval as a document identifier (docid) generation task.<n>Existing GenIR models suffer from token-level misalignment, where models trained to predict the next token often fail to capture document-level relevance effectively.<n>We propose direct document relevance optimization (DDRO), which aligns token-level docid generation with document-level relevance estimation through direct optimization via pairwise ranking.
  arXiv  Detail & Related papers  (2025-04-07T15:27:37Z)
• Is Relevance Propagated from Retriever to Generator in RAG? [21.82171240511567]
  RAG is a framework for incorporating external knowledge, usually in the form of a set of documents retrieved from a collection.<n>We empirically investigate whether a RAG context comprised of topically relevant documents leads to improved downstream performance.
  arXiv  Detail & Related papers  (2025-02-20T20:21:46Z)
• JudgeRank: Leveraging Large Language Models for Reasoning-Intensive Reranking [81.88787401178378]
  We introduce JudgeRank, a novel agentic reranker that emulates human cognitive processes when assessing document relevance. We evaluate JudgeRank on the reasoning-intensive BRIGHT benchmark, demonstrating substantial performance improvements over first-stage retrieval methods. In addition, JudgeRank performs on par with fine-tuned state-of-the-art rerankers on the popular BEIR benchmark, validating its zero-shot generalization capability.
  arXiv  Detail & Related papers  (2024-10-31T18:43:12Z)
• REAR: A Relevance-Aware Retrieval-Augmented Framework for Open-Domain Question Answering [115.72130322143275]
  REAR is a RElevance-Aware Retrieval-augmented approach for open-domain question answering (QA) We develop a novel architecture for LLM-based RAG systems, by incorporating a specially designed assessment module. Experiments on four open-domain QA tasks show that REAR significantly outperforms previous a number of competitive RAG approaches.
  arXiv  Detail & Related papers  (2024-02-27T13:22:51Z)
• SAIS: Supervising and Augmenting Intermediate Steps for Document-Level Relation Extraction [51.27558374091491]
  We propose to explicitly teach the model to capture relevant contexts and entity types by supervising and augmenting intermediate steps (SAIS) for relation extraction. Based on a broad spectrum of carefully designed tasks, our proposed SAIS method not only extracts relations of better quality due to more effective supervision, but also retrieves the corresponding supporting evidence more accurately.
  arXiv  Detail & Related papers  (2021-09-24T17:37:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.