RMIT-ADM+S at the MMU-RAG NeurIPS 2025 Competition

TL;DR

The paper introduces RMIT-ADM+S, a lightweight, adaptive retrieval-augmented generation system that dynamically adjusts retrieval strategies based on query complexity, achieving high performance on the NeurIPS 2025 competition with minimal resources.

Contribution

It presents Routing-to-RAG (R2RAG), a novel adaptive RAG architecture that improves retrieval efficiency and effectiveness using smaller LLMs and dynamic strategies, extending previous systems like G-RAG.

Findings

Won the Best Dynamic Evaluation award at NeurIPS 2025

Operates efficiently on a single consumer-grade GPU

Demonstrates high effectiveness in complex research tasks

Abstract

This paper presents the award-winning RMIT-ADM+S system for the Text-to-Text track of the NeurIPS~2025 MMU-RAG Competition. We introduce Routing-to-RAG (R2RAG), a research-focused retrieval-augmented generation (RAG) architecture composed of lightweight components that dynamically adapt the retrieval strategy based on inferred query complexity and evidence sufficiency. The system uses smaller LLMs, enabling operation on a single consumer-grade GPU while supporting complex research tasks. It builds on the G-RAG system, winner of the ACM~SIGIR~2025 LiveRAG Challenge, and extends it with modules informed by qualitative review of outputs. R2RAG won the Best Dynamic Evaluation award in the Open Source category, demonstrating high effectiveness with careful design and efficient use of resources.