MARAG-R1: Beyond Single Retriever via Reinforcement-Learned Multi-Tool Agentic Retrieval

TL;DR

MARAG-R1 introduces a reinforcement-learned multi-tool retrieval framework for LLMs, enabling dynamic coordination of multiple retrieval methods to improve factual accuracy and reasoning in knowledge-intensive tasks.

Contribution

It proposes a novel multi-tool retrieval system with reinforcement learning, allowing LLMs to better access and synthesize external knowledge for complex reasoning.

Findings

Outperforms strong baselines on GlobalQA, HotpotQA, and 2WikiMultiHopQA.

Achieves state-of-the-art results in corpus-level reasoning.

Demonstrates effective coordination of multiple retrieval tools.

Abstract

Large Language Models (LLMs) excel at reasoning and generation but are inherently limited by static pretraining data, resulting in factual inaccuracies and weak adaptability to new information. Retrieval-Augmented Generation (RAG) addresses this issue by grounding LLMs in external knowledge; However, the effectiveness of RAG critically depends on whether the model can adequately access relevant information. Existing RAG systems rely on a single retriever with fixed top-k selection, restricting access to a narrow and static subset of the corpus. As a result, this single-retriever paradigm has become the primary bottleneck for comprehensive external information acquisition, especially in tasks requiring corpus-level reasoning. To overcome this limitation, we propose MARAG-R1, a reinforcement-learned multi-tool RAG framework that enables LLMs to dynamically coordinate multiple retrieval mechanisms for broader and more precise information access. MARAG-R1 equips the model with four retrieval tools -- semantic search, keyword search, filtering, and aggregation -- and learns both how and when to use them through a two-stage training process: supervised fine-tuning followed by reinforcement learning. This design allows the model to interleave reasoning and retrieval, progressively gathering sufficient evidence for corpus-level synthesis. Experiments on GlobalQA, HotpotQA, and 2WikiMultiHopQA demonstrate that MARAG-R1 substantially outperforms strong baselines and achieves new state-of-the-art results in corpus-level reasoning tasks.