mlsanityData to 14 Oct 2026

Uncertainty-Aware Decision Making in Multimodal Large Language Models

arXiv 2608.17084 ↗First posted 17 Aug 2026cs.CLcited by 0

On the map: Multimodal LLMs / vision-language models → MLLM training and visual instruction tuning

Abstract (arXiv)

Multimodal large language models (MLLMs) increasingly answer questions whose correctness depends on visual, textual, temporal, acoustic, document, chart, or embodied evidence. Their failures are therefore not only linguistic. A fluent answer may conceal poor input quality, a perceptual error, weak grounding, conflict between modalities, unstable reasoning, distribution shift, or a question that is not answerable from the supplied evidence. This survey organizes the literature on uncertainty-aware MLLMs around a decision-centered framework: uncertainty sources give rise to observable signals, signals must be calibrated or controlled for risk, and calibrated uncertainty should determine the system action. We review work on token and logit uncertainty, semantic disagreement, perturbation instability, grounding and attribution scores, verbalized confidence, verifier and judge scores, conformal prediction, selective answering, abstention, clarification, retrieval, self-checking, and escalation. The central argument is that uncertainty should not be evaluated only as a confidence number; it should be evaluated by whether it improves behavior under insufficient, conflicting, shifted, or high-risk multimodal evidence. We position this survey against text-only uncertainty and abstention surveys, broad MLLM surveys, MLLM hallucination surveys, and safety-oriented reviews. We conclude with open problems in source-aware decomposition, action-aware benchmarks, calibration under shift, black-box uncertainty estimation, broader modality coverage, reproducible reporting, and human-centered uncertainty communication.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in MLLM training and visual instruction tuning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 64 of its 76 references that are in the map. First posted on arXiv 17 Aug 2026.

Builds on

64 of 76 references.

Closest ancestors · 13

Same direction, newest first

  1. 2026Instinct vs. Reflection: Unifying Token and Verbalized Confidence in Multimodal Large Models19 Apr
  2. 2026Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volume27 Feb
  3. 2025When Modalities Conflict: How Unimodal Reasoning Uncertainty Governs Preference Dynamics in MLLMs4 Nov
  4. 2025FESTA: Functionally Equivalent Sampling for Trust Assessment of Multimodal LLMs20 Sep
  5. 2025Calibrating MLLM-as-a-judge via Multimodal Bayesian Prompt Ensembles10 Sep
  6. 2025Uncertainty-o: One Model-agnostic Framework for Unveiling Uncertainty in Large Multimodal Models9 Jun
  7. 2025MMBoundary: Advancing MLLM Knowledge Boundary Awareness through Reasoning Step Confidence Calibration29 May
  8. 2025Calibrating Uncertainty Quantification of Multi-Modal LLMs using Grounding30 Apr
  9. 2024Unveiling Uncertainty: A Deep Dive into Calibration and Performance of Multimodal Large Language Models19 Dec
  10. 2024Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios5 Nov
  11. 2024Sample then Identify: A General Framework for Risk Control and Assessment in Multimodal Large Language Models10 Oct
  12. 2024MM-SAP: A Comprehensive Benchmark for Assessing Self-Awareness of Multimodal Large Language Models in Perception15 Jan
1 more
  1. 2023A Survey on Multimodal Large Language Models23 Jun

Other references · 51

  1. 2024Hallucination of Multimodal Large Language Models: A Surveycited by 388
  2. 2024A Survey on Hallucination in Large Vision-Language Modelscited by 357
  3. 2024Know Your Limits: A Survey of Abstention in Large Language Modelscited by 121
  4. 2024A Survey on Uncertainty Quantification of Large Language Models: Taxonomy, Open Research Challenges, and Future Directionscited by 121
  5. 2025ChartQAPro: A More Diverse and Challenging Benchmark for Chart Question Answeringcited by 108
  6. 2025Uncertainty Quantification and Confidence Calibration in Large Language Models: A Surveycited by 95
  7. 2024Safety of Multimodal Large Language Models on Images and Textscited by 75
  8. 2024VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimationcited by 73
  9. 2025A Survey of Uncertainty Estimation Methods on Large Language Modelscited by 62
  10. 2024Selective "Selective Prediction": Reducing Unnecessary Abstention in Vision-Language Reasoningcited by 40
  11. 2023Towards Calibrated Robust Fine-Tuning of Vision-Language Modelscited by 35
  12. 2024Uncertainty-Aware Evaluation for Vision-Language Modelscited by 29
38 more
  1. 2025A Survey of Multimodal Hallucination Evaluation and Detectioncited by 27
  2. 2024The First to Know: How Token Distributions Reveal Hidden Knowledge in Large Vision-Language Models?cited by 26
  3. 2025Learning Conformal Abstention Policies for Adaptive Risk Management in Large Language and Vision-Language Modelscited by 21
  4. 2025Grounding Multimodal LLMs to Embodied Agents that Ask for Help with Reinforcement Learningcited by 21
  5. 2025Seeing is Believing, but How Much? A Comprehensive Analysis of Verbalized Calibration in Vision-Language Modelscited by 16
  6. 2024Enhancing Vision-Language Model Reliability with Uncertainty-Guided Dropout Decodingcited by 15
  7. 2024HEIE: MLLM-Based Hierarchical Explainable AIGC Image Implausibility Evaluatorcited by 13
  8. 2025Seeing and Reasoning with Confidence: Supercharging Multimodal LLMs with an Uncertainty-Aware Agentic Frameworkcited by 12
  9. 2025Consensus Entropy: Harnessing Multi-VLM Agreement for Self-Verifying and Self-Improving OCRcited by 12
  10. 2025Mitigating Hallucination in Large Vision-Language Models via Adaptive Attention Calibrationcited by 11
  11. 2025Teaching Vision-Language Models to Ask: Resolving Ambiguity in Visual Questionscited by 11
  12. 2024Right this way: Can VLMs Guide Us to See More to Answer Questions?cited by 10
  13. 2025When Robots Should Say "I Don't Know": Benchmarking Abstention in Embodied Question Answeringcited by 10
  14. 2026Adaptive Vision-Language Model Routing for Computer Use Agentscited by 8
  15. 2026VL-Calibration: Decoupled Confidence Calibration for Large Vision-Language Models Reasoningcited by 8
  16. 2025Conformal Prediction and MLLM aided Uncertainty Quantification in Scene Graph Generationcited by 7
  17. 2026Knowing When Not to Answer: Evaluating Abstention in Multimodal Reasoning Systemscited by 7
  18. 2024Can We Predict Performance of Large Models across Vision-Language Tasks?cited by 6
  19. 2025Can Video LLMs Refuse to Answer? Alignment for Answerability in Video Large Language Modelscited by 6
  20. 2025Prompt4Trust: A Reinforcement Learning Prompt Augmentation Framework for Clinically-Aligned Confidence Calibration in Multimodal Large Language Modelscited by 6
  21. 2026Omni-Modal Dissonance Benchmark: Systematically Breaking Modality Consensus to Probe Robustness and Calibrated Abstentioncited by 6
  22. 2024Selectively Answering Visual Questionscited by 5
  23. 2025Variational Visual Question Answering for Uncertainty-Aware Selective Predictioncited by 5
  24. 2026VAUQ: Vision-Aware Uncertainty Quantification for LVLM Self-Evaluationcited by 5
  25. 2026Grounded or Guessing? LVLM Confidence Estimation via Blind-Image Contrastive Rankingcited by 5
  26. 2025Refine and Align: Confidence Calibration through Multi-Agent Interaction in VQAcited by 4
  27. 2025EmoCaliber: Advancing Reliable Visual Emotion Comprehension via Confidence Verbalization and Calibrationcited by 4
  28. 2026Walking Through Uncertainty: An Empirical Study of Uncertainty Estimation for Audio-Aware Large Language Modelscited by 4
  29. 2026Confidence Calibration for Multimodal LLMs: An Empirical Study through Medical VQAcited by 4
  30. 2025Data-Driven Calibration of Prediction Sets in Large Vision-Language Models Based on Inductive Conformal Predictioncited by 3
  31. 2025CLASH: A Benchmark for Cross-Modal Contradiction Detectioncited by 3
  32. 2025Explicit Abstention Knobs for Predictable Reliability in Video Question Answeringcited by 3
  33. 2026VLM-UQBench: A Benchmark for Modality-Specific and Cross-Modality Uncertainties in Vision Language Modelscited by 3
  34. 2026Multimodal Adaptive Retrieval Augmented Generation through Internal Representation Learningcited by 3
  35. 2025The Art of Saying "Maybe": A Conformal Lens for Uncertainty Benchmarking in VLMscited by 2
  36. 2026Knowing When to Answer: Adaptive Confidence Refinement for Reliable Audio-Visual Question Answeringcited by 2
  37. 2025Analysis of Image-and-Text Uncertainty Propagation in Multimodal Large Language Models with Cardiac MR-Based Applicationscited by 1
  38. 2026From Attribution to Abstention: Training-Free Attention-Based Auditing for Clinical Summarizationcited by 1

50 most cited of 51 All 51 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Dissecting Training-Free Uncertainty Estimation in Multimodal Large Language Models
  2. 2026CoMet: Context and Multiplicity Decomposition for Multimodal Uncertainty Estimation
  3. 2025Revisiting Uncertainty Estimation and Calibration of Large Language Models
  4. 2026Beyond "I Don't Know": Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty
  5. 2026Uncertainty Quantification for LLM Agents: A Taxonomy, an Evaluation Protocol, and an Empirical Study
  6. 2025Can LLMs Detect Their Confabulations? Estimating Reliability in Uncertainty-Aware Language Models
  7. 2026Estimating Uncertainty from Reasoning: A Large-Scale Study of Multi- and Crosslingual MCQA Performance in LLMs
  8. 2026Estimating Uncertainty of Omnimodal Large Language Models via External Omnimodal Embeddings