mlsanityData to 14 Oct 2026

Uncertainty Quantification for Retrieval-Augmented Reasoning

arXiv 2510.11483 ↗First posted 13 Oct 2025cs.IRcited by 6

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

also in 1 more
Abstract (arXiv)

Retrieval-augmented reasoning (RAR) is a recent evolution of retrieval-augmented generation (RAG) that employs multiple reasoning steps for retrieval and generation. While effective for some complex queries, RAR remains vulnerable to errors and misleading outputs. Uncertainty quantification (UQ) offers methods to estimate the confidence of systems' outputs. These methods, however, often handle simple queries with no retrieval or single-step retrieval, without properly handling RAR setup. Accurate estimation of UQ for RAR requires accounting for all sources of uncertainty, including those arising from retrieval and generation. In this paper, we account for all these sources and introduce Retrieval-Augmented Reasoning Consistency (R2C)--a novel UQ method for RAR. The core idea of R2C is to perturb the multi-step reasoning process by applying various actions to reasoning steps. These perturbations alter the retriever's input, which shifts its output and consequently modifies the generator's input at the next step. Through this iterative feedback loop, the retriever and generator continuously reshape one another's inputs, enabling us to capture uncertainty arising from both components. Experiments on five popular RAR systems across diverse QA datasets show that R2C improves AUROC by over 5% on average compared to the state-of-the-art UQ baselines. Extrinsic evaluations using R2C as an external signal further confirm its effectiveness for two downstream tasks: in Abstention, it achieves ~5% gains in both F1Abstain and AccAbstain; in Model Selection, it improves the exact match by ~7% over single models and ~3% over selection methods.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 55 of its 89 references that are in the map. First posted on arXiv 13 Oct 2025.

Builds on

55 of 89 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025Batched Self-Consistency Improves LLM Relevance Assessment and Ranking18 May
  2. 2025ReARTeR: Retrieval-Augmented Reasoning with Trustworthy Process Rewarding14 Jan
  3. 2024RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinement17 Dec
  4. 2024RARE: Retrieval-Augmented Reasoning Enhancement for Large Language Models3 Dec
  5. 2024Retrieving, Rethinking and Revising: The Chain-of-Verification Can Improve Retrieval Augmented Generation8 Oct
  6. 2024Open-RAG: Enhanced Retrieval-Augmented Reasoning with Open-Source Large Language Models2 Oct
  7. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions20 Dec

Other references · 48

  1. 2024Qwen2.5 Technical Reportcited by 4,942
  2. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  3. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  4. 2017What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?cited by 2,352
  5. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  6. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  7. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,128
  8. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  9. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  10. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
  11. 2023Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedbackcited by 812
  12. 2021A Survey of Uncertainty in Deep Neural Networkscited by 680
36 more
  1. 2023LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusioncited by 564
  2. 2023Active Retrieval Augmented Generationcited by 516
  3. 2023Generating with Confidence: Uncertainty Quantification for Black-box Large Language Modelscited by 293
  4. 2024The Power of Noise: Redefining Retrieval for RAG Systemscited by 281
  5. 2022Out-of-Distribution Detection and Selective Generation for Conditional Language Modelscited by 197
  6. 2024Don't Hallucinate, Abstain: Identifying LLM Knowledge Gaps via Multi-LLM Collaborationcited by 180
  7. 2024Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantificationcited by 162
  8. 2023Shifting Attention to Relevance: Towards the Predictive Uncertainty Quantification of Free-Form Large Language Modelscited by 150
  9. 2024Benchmarking LLMs via Uncertainty Quantificationcited by 126
  10. 2023Decomposing Uncertainty for Large Language Models through Input Clarification Ensemblingcited by 114
  11. 2025Uncertainty Quantification and Confidence Calibration in Large Language Models: A Surveycited by 95
  12. 2024Evaluating Retrieval Quality in Retrieval-Augmented Generationcited by 90
  13. 2025Harnessing Multiple Large Language Models: A Survey on LLM Ensemblecited by 86
  14. 2024Uncertainty Estimation and Quantification for LLMs: A Simple Supervised Approachcited by 76
  15. 2024ConU: Conformal Uncertainty in Large Language Models with Correctness Coverage Guaranteescited by 65
  16. 2024DRAGIN: Dynamic Retrieval Augmented Generation based on the Information Needs of Large Language Modelscited by 63
  17. 2024Fine Tuning vs. Retrieval Augmented Generation for Less Popular Knowledgecited by 50
  18. 2024MARS: Meaning-Aware Response Scoring for Uncertainty Estimation in Generative LLMscited by 46
  19. 2025s3: You Don't Need That Much Data to Train a Search Agent via RLcited by 42
  20. 2025UProp: Investigating the Uncertainty Propagation of LLMs in Multi-Step Agentic Decision-Makingcited by 42
  21. 2025Adaptive Retrieval Without Self-Knowledge? Bringing Uncertainty Back Homecited by 33
  22. 2024Smoothie: Label Free Language Model Routingcited by 31
  23. 2024Cycles of Thought: Measuring LLM Confidence through Stable Explanationscited by 29
  24. 2021Directed Diversity: Leveraging Language Embedding Distances for Collective Creativity in Crowd Ideationcited by 23
  25. 2024A Survey on Recent Advances in Conversational Data Generationcited by 23
  26. 2024Do Not Design, Learn: A Trainable Scoring Function for Uncertainty Estimation in Generative LLMscited by 23
  27. 2025CoT-UQ: Improving Response-wise Uncertainty Quantification in LLMs with Chain-of-Thoughtcited by 23
  28. 2025Reconsidering LLM Uncertainty Estimation Methods in the Wildcited by 17
  29. 2025Why Uncertainty Estimation Methods Fall Short in RAG: An Axiomatic Analysiscited by 14
  30. 2025Uncertainty Quantification in Retrieval Augmented Question Answeringcited by 8
  31. 2025Evaluating the Evaluation of Diversity in Commonsense Generationcited by 5
  32. 2025Adaptive Orchestration of Modular Generative Information Access Systemscited by 4
  33. 2025TruthTorchLM: A Comprehensive Library for Predicting Truthfulness in LLM Outputscited by 4
  34. 2025CIIR@LiveRAG 2025: Optimizing Multi-Agent Retrieval Augmented Generation through Self-Trainingcited by 2
  35. 2025Revisiting RAG Ensemble: A Theoretical and Mechanistic Analysis of Multi-RAG System Collaborationcited by 2
  36. 2025Ask, Retrieve, Summarize: A Modular Pipeline for Scientific Literature Summarizationcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026NOVA: NOise-aware Verbal Confidence CAlibration for Robust Large Language Models in RAG Systemscited by 6
  2. 2026When Deep Research Agents Stagnate: Enhancing Reasoning with Retrieval-Aware Agent Controlcited by 3
  3. 2026Learning to Reason for Multi-Step Retrieval of Personal Context in Personalized Question Answeringcited by 2
  4. 2026When Confidence Takes the Wrong Path: Diagnosing Retrieval-State Lock-In in RAGcited by 2
  5. 2026Uncertainty Quantification for Multimodal Retrieval Augmented Generationcited by 0
  6. 2026Evaluating RAG Reliability under Clean, Misleading, and Mixed Retrievalcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026URAG: A Benchmark for Uncertainty Quantification in Retrieval-Augmented Large Language Models
  2. 2026Interpretable Uncertainty for Adaptive Retrieval and Reasoning in Question Answering
  3. 2026Decide Then Retrieve: A Training-Free Framework with Uncertainty-Guided Triggering and Dual-Path Retrieval
  4. 2024Improving Retrieval Augmented Language Model with Self-Reasoning
  5. 2025LIR3AG: A Lightweight Rerank Reasoning Strategy Framework for Retrieval-Augmented Generation
  6. 2025MCTS-RAG: Enhancing Retrieval-Augmented Generation with Monte Carlo Tree Search
  7. 2025Thinking Forward and Backward: Multi-Objective Reinforcement Learning for Retrieval-Augmented Reasoning
  8. 2026ERA: Evidence-based Reliability Alignment for Honest Retrieval-Augmented Generation