RMIT-ADM+S at the MMU-RAG NeurIPS 2025 Competition
#RMIT-ADM+S #NeurIPS 2025 #R2RAG #Retrieval-Augmented Generation #Information Retrieval #Dynamic Evaluation #Efficient AI
📌 Key Takeaways
- RMIT researchers won Best Dynamic Evaluation award at NeurIPS 2025 MMU-RAG Competition
- Their R2RAG system uses smaller LLMs to operate efficiently on consumer-grade GPUs
- The system dynamically adapts retrieval strategies based on query complexity
- Research builds on previous G-RAG system that won ACM SIGIR 2025
📖 Full Retelling
Researchers from RMIT led by Kun Ran and 11 co-authors won the Best Dynamic Evaluation award in the Open Source category at the NeurIPS 2025 MMU-RAG Competition with their RMIT-ADM+S system, as detailed in their paper submitted to arXiv on February 24, 2026. Their innovative Routing-to-RAG (R2RAG) architecture, built upon the previous G-RAG system that won the ACM SIGIR 2025 LiveRAG Challenge, demonstrates how smaller language models can effectively handle complex research tasks while operating on a single consumer-grade GPU. The system represents a significant advancement in retrieval-augmented generation technology by dynamically adapting retrieval strategies based on inferred query complexity and evidence sufficiency.
The R2RAG system distinguishes itself through its lightweight components that optimize performance across different types of queries without requiring substantial computational resources. This efficiency is particularly notable as it addresses growing concerns about the environmental impact and accessibility of AI systems that typically require massive computational infrastructure. The researchers enhanced their previous G-RAG architecture by incorporating modules informed by qualitative review of outputs, demonstrating how iterative improvement based on performance analysis can lead to breakthrough achievements in competitive AI environments.
The success of the RMIT-ADM+S team highlights the increasing importance of efficient AI design in research and industry settings. By proving that high-performance systems can operate on consumer-grade hardware, their work makes advanced AI capabilities more accessible to researchers and organizations with limited computational resources. This achievement comes at a critical time as the AI community grapples with balancing performance improvements with sustainability concerns and broader accessibility.
🏷️ Themes
Artificial Intelligence, Information Retrieval, Efficient Computing
📚 Related People & Topics
Information retrieval
Finding information for an information need
Information retrieval (IR) in computing and information science is the task of identifying and retrieving information system resources that are relevant to an information need. The information need can be specified in the form of a search query. In the case of document retrieval, queries can be base...
Entity Intersection Graph
Connections for Information retrieval:
🌐
Large language model
3 shared
🌐
Artificial intelligence
2 shared
🌐
Recommender system
2 shared
🌐
Efficiency
1 shared
🌐
Transparency
1 shared
Original Source
--> Computer Science > Information Retrieval arXiv:2602.20735 [Submitted on 24 Feb 2026] Title: RMIT-ADM+S at the MMU-RAG NeurIPS 2025 Competition Authors: Kun Ran , Marwah Alaofi , Danula Hettiachchi , Chenglong Ma , Khoi Nguyen Dinh Anh , Khoi Vo Nguyen , Sachin Pathiyan Cherumanal , Lida Rashidi , Falk Scholer , Damiano Spina , Shuoqi Sun , Oleg Zendel View a PDF of the paper titled RMIT-ADM+S at the MMU-RAG NeurIPS 2025 Competition, by Kun Ran and 11 other authors View PDF HTML 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 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. Comments: MMU-RAG NeurIPS 2025 winning system Subjects: Information Retrieval (cs.IR) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2602.20735 [cs.IR] (or arXiv:2602.20735v1 [cs.IR] for this version) https://doi.org/10.48550/arXiv.2602.20735 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Oleg Zendel [ view email ] [v1] Tue, 24 Feb 2026 09:58:25 UTC (93 KB) Full-text links: Access Paper: View a PDF of the paper titled RMIT-ADM+S at the MMU-RAG NeurIPS 2025 Competition, by Kun Ran and 11 other authors View PDF HTML TeX Source view license Current browse context: cs.IR < prev | next > new | recent | 2026-02 Change t...
Read full article at source