mlsanityData to 14 Oct 2026

Online-Optimized RAG for Tool Use and Function Calling

arXiv 2509.20415 ↗First posted 24 Sep 2025cs.SEcited by 2

On the map: Retrieval-augmented generation → RAG pipeline design and evaluation

also in 1 more
Abstract (arXiv)

In many applications, retrieval-augmented generation (RAG) drives tool use and function calling by embedding the (user) queries and matching them to pre-specified tool/function descriptions. In this paper, we address an embedding misalignment issue that often arises in practical applications due to imperfect embedding models or noisy descriptions; such misalignment may lead to incorrect retrieval and task failure. We introduce Online-Optimized RAG, a deployment-time framework that continually adapts retrieval embeddings from live interactions using minimal feedback (e.g., task success). Online-Optimized RAG applies lightweight online gradient updates with negligible per-query latency and requires no changes to the underlying LLM. The method is plug-and-play: it supports both single- and multi-hop tool use, dynamic tool inventories, and K-retrieval with re-ranking. We provide a problem-dependent theoretical analysis that quantifies how the method's performance depends on the initialization quality of the embeddings and other related quantities. Across diverse tool-use and document-retrieval scenarios, our Online-Optimized RAG consistently improves tool selection accuracy and end-task success, thus providing a simple, practical path to robust, self-improving RAG systems.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in RAG pipeline design and evaluation, one of 15 directions in Retrieval-augmented generation. Builds on 24 of its 44 references that are in the map. First posted on arXiv 24 Sep 2025.

Builds on

24 of 44 references.

References · 24

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  3. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,947
  4. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  5. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
  6. 2021BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Modelscited by 1,252
  7. 2023HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Facecited by 1,190
  8. 2019Introduction to Online Convex Optimizationcited by 803
  9. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMscited by 591
  10. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 451
  11. 2024RAFT: Adapting Language Model to Domain Specific RAGcited by 262
  12. 2023ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddingscited by 240
12 more
  1. 2024Kernel Language Entropy: Fine-grained Uncertainty Quantification for LLMs from Semantic Similaritiescited by 159
  2. 2024EASYTOOL: Enhancing LLM-based Agents with Concise Tool Instructioncited by 154
  3. 2023Quantifying Uncertainty in Answers from any Language Model and Enhancing their Trustworthinesscited by 122
  4. 2024Planning, Creation, Usage: Benchmarking LLMs for Comprehensive Tool Utilization in Real-World Complex Scenarioscited by 86
  5. 2024Uncertainty Estimation and Quantification for LLMs: A Simple Supervised Approachcited by 76
  6. 2024ToolGen: Unified Tool Retrieval and Calling via Generationcited by 74
  7. 2024MBA-RAG: a Bandit Approach for Adaptive Retrieval-Augmented Generation through Question Complexitycited by 29
  8. 2024Toolshed: Scale Tool-Equipped Agents with Advanced RAG-Tool Fusion and Tool Knowledge Basescited by 23
  9. 2024NUDGE: Lightweight Non-Parametric Fine-Tuning of Embeddings for Retrievalcited by 5
  10. 2025Query Drift Compensation: Enabling Compatibility in Continual Learning of Retrieval Embedding Modelscited by 5
  11. 2025FLAIR: Feedback Learning for Adaptive Information Retrievalcited by 4
  12. 2023ReFIT: Relevance Feedback from a Reranker during Inferencecited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026PhGPO: Pheromone-Guided Policy Optimization for Long-Horizon Tool Planningcited by 9
  2. 2026Multi-Field Tool Retrievalcited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2026Mapping the RAG Landscape: A Four Axis Taxonomy of Efficiency, Defense, Interactivity, and Reasoning
  2. 2025Fast or Better? Balancing Accuracy and Cost in Retrieval-Augmented Generation with Flexible User Control
  3. 2026Revisiting RAG Retrievers: An Information Theoretic Benchmark
  4. 2023Context Tuning for Retrieval Augmented Generation
  5. 2024Toward Optimal Search and Retrieval for RAG
  6. 2025An Analysis of Hyper-Parameter Optimization Methods for Retrieval Augmented Generation
  7. 2026RAISE: RAG Design as an Architecture Search Problem
  8. 2025Less Is More: Elevating RAG via Performance-Driven Context Compression