mlsanityData to 7 Oct 2026

Certainly Uncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness

arXiv 2407.01942 ↗First posted 2 Jul 2024cs.AI cs.CL cs.CVcited by 13

Not on the map. Why?

Abstract (arXiv)

The ability to acknowledge the inevitable uncertainty in their knowledge and reasoning is a prerequisite for AI systems to be truly truthful and reliable. In this paper, we present a taxonomy of uncertainty specific to vision-language AI systems, distinguishing between epistemic uncertainty (arising from a lack of information) and aleatoric uncertainty (due to inherent unpredictability), and further explore finer categories within. Based on this taxonomy, we synthesize a benchmark dataset, CertainlyUncertain, featuring 178K visual question answering (VQA) samples as contrastive pairs. This is achieved by 1) inpainting images to make previously answerable questions into unanswerable ones; and 2) using image captions to prompt large language models for both answerable and unanswerable questions. Additionally, we introduce a new metric confidence-weighted accuracy, that is well correlated with both accuracy and calibration error, to address the shortcomings of existing metrics.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 63 of its 74 references that are in the map. First posted on arXiv 2 Jul 2024.

Builds on

63 of 74 references.

References · 63

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2023Visual Instruction Tuningcited by 7,792
  4. 2023Mistral 7Bcited by 3,065
  5. 2016Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answeringcited by 2,498
  6. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 2,179
  7. 2023MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGIcited by 2,044
  8. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,696
  9. 2023HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Facecited by 1,173
  10. 2024Grounded SAM: Assembling Open-World Models for Diverse Visual Taskscited by 914
  11. 2018VizWiz Grand Challenge: Answering Visual Questions from Blind Peoplecited by 852
  12. 2023SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Modelscited by 787
38 more
  1. 2021Resolution-robust Large Mask Inpainting with Fourier Convolutionscited by 730
  2. 2023Aligning Large Multimodal Models with Factually Augmented RLHFcited by 703
  3. 2021Retrieval Augmentation Reduces Hallucination in Conversationcited by 691
  4. 2018Object Hallucination in Image Captioningcited by 661
  5. 2023ViperGPT: Visual Inference via Python Execution for Reasoningcited by 591
  6. 2022Socratic Models: Composing Zero-Shot Multimodal Reasoning with Languagecited by 513
  7. 2023MM-REACT: Prompting ChatGPT for Multimodal Reasoning and Actioncited by 511
  8. 2022Visual Programming: Compositional visual reasoning without trainingcited by 507
  9. 2023Mitigating Hallucination in Large Multi-Modal Models via Robust Instruction Tuningcited by 439
  10. 2024Hallucination of Multimodal Large Language Models: A Surveycited by 380
  11. 2023HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Modelscited by 329
  12. 2023Analyzing and Mitigating Object Hallucination in Large Vision-Language Modelscited by 314
  13. 2023AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluationcited by 293
  14. 2023AlignScore: Evaluating Factual Consistency with a Unified Alignment Functioncited by 284
  15. 2023Detecting and Preventing Hallucinations in Large Vision Language Modelscited by 254
  16. 2020Selective Question Answering under Domain Shiftcited by 228
  17. 2023Woodpecker: Hallucination Correction for Multimodal Large Language Modelscited by 223
  18. 2024Multi-Modal Hallucination Control by Visual Information Groundingcited by 196
  19. 2015DenseCap: Fully Convolutional Localization Networks for Dense Captioningcited by 176
  20. 2023QA-LoRA: Quantization-Aware Low-Rank Adaptation of Large Language Modelscited by 154
  21. 2020Detecting Hallucinated Content in Conditional Neural Sequence Generationcited by 151
  22. 2024Fine-grained Hallucination Detection and Editing for Language Modelscited by 146
  23. 2021A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generationcited by 137
  24. 2024DOCCI: Descriptions of Connected and Contrasting Imagescited by 130
  25. 2021Hallucinated but Factual! Inspecting the Factuality of Hallucinations in Abstractive Summarizationcited by 128
  26. 2023Silkie: Preference Distillation for Large Visual Language Modelscited by 128
  27. 2023HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Datacited by 124
  28. 2021Neural Path Hunter: Reducing Hallucination in Dialogue Systems via Path Groundingcited by 106
  29. 2023R-Tuning: Instructing Large Language Models to Say `I Don't Know'cited by 106
  30. 2023Evaluation and Analysis of Hallucination in Large Vision-Language Modelscited by 99
  31. 2023The Generative AI Paradox: "What It Can Create, It May Not Understand"cited by 97
  32. 2017An Analysis of Visual Question Answering Algorithmscited by 90
  33. 2022Reliable Visual Question Answering: Abstain Rather Than Answer Incorrectlycited by 86
  34. 2023Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Modelscited by 80
  35. 2022Investigating Selective Prediction Approaches Across Several Tasks in IID, OOD, and Adversarial Settingscited by 66
  36. 2023Mitigating Fine-Grained Hallucination by Fine-Tuning Large Vision-Language Models with Caption Rewritescited by 61
  37. 2023Mitigating Hallucination in Visual Language Models with Visual Supervisioncited by 49
  38. 2022RHO (ρ): Reducing Hallucination in Open-domain Dialogues with Knowledge Groundingcited by 47

50 most cited of 63 All 63 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 3; 2026: 9.

13 papers.

Sort:

By area:

  1. 2025QuestBench: Can LLMs ask the right question to acquire information in reasoning tasks?cited by 56
  2. 2024MAQA: Evaluating Uncertainty Quantification in LLMs Regarding Data Uncertaintycited by 25
  3. 2025CLIP-UP: CLIP-Based Unanswerable Problem Detection for Visual Question Answeringcited by 6
  4. 2026VLM-UQBench: A Benchmark for Modality-Specific and Cross-Modality Uncertainties in Vision Language Modelscited by 3
  5. 2026TRAPSBench: Vision-Language Models Encode but Fail to Express Epistemic Restraintcited by 3
  6. 2025CORA: Consistency-Guided Semi-Supervised Framework for Reasoning Segmentationcited by 1
  7. 2026Failing to See or Failing to Know? Attributing Errors in Vision-Language Modelscited by 1
  8. 2026CCTVBench: Contrastive Consistency Traffic VideoQA Benchmark for Multimodal LLMscited by 0
  9. 2026Inducing Artificial Uncertainty in Language Modelscited by 0
  10. 2026Uncertainty Is Not a Safety Net for Clinical VQA, but Can It Anticipate Model Failure?cited by 0
  11. 2026Whose record is this? Diagnosing and authorizing record use in personalized multimodal modelscited by 0
  12. 2026From Knowing to Abstaining: Bridging the Representation-Action Gap in Vision-Language Modelscited by 0
1 more
  1. 2026When Does an Image Determine the Answer? Benchmarking Visual Answerability across Charts and Scenescited by 0