mlsanityData to 7 Oct 2026

Cycle-Consistency for Robust Visual Question Answering

arXiv 1902.05660 ↗First posted 15 Feb 2019cs.CVcited by 105

Not on the map. Why?

Cited by 26 in Multimodal LLMs / vision-language models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Despite significant progress in Visual Question Answering over the years, robustness of today's VQA models leave much to be desired. We introduce a new evaluation protocol and associated dataset (VQA-Rephrasings) and show that state-of-the-art VQA models are notoriously brittle to linguistic variations in questions. VQA-Rephrasings contains 3 human-provided rephrasings for 40k questions spanning 40k images from the VQA v2.0 validation dataset. As a step towards improving robustness of VQA models, we propose a model-agnostic framework that exploits cycle consistency. Specifically, we train a model to not only answer a question, but also generate a question conditioned on the answer, such that the answer predicted for the generated question is the same as the ground truth answer to the original question. Without the use of additional annotations, we show that our approach is significantly more robust to linguistic variations than state-of-the-art VQA models, when evaluated on the VQA-Rephrasings dataset. In addition, our approach outperforms state-of-the-art approaches on the standard VQA and Visual Question Generation tasks on the challenging VQA v2.0 dataset.

Key facts. Cited by 105 papers in the map (data through 7 Oct 2026). First posted on arXiv 15 Feb 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 24; 2022: 21; 2023: 20; 2024: 19; 2025: 9; 2026: 12.

105 papers.

Sort:

By area:

  1. 2023Nougat: Neural Optical Understanding for Academic Documentscited by 240
  2. 2022Towards Adversarial Attack on Vision-Language Pre-training Modelscited by 142
  3. 2022Vision-Language Pre-training: Basics, Recent Advances, and Future Trendscited by 117
  4. 2022Reliable Visual Question Answering: Abstain Rather Than Answer Incorrectlycited by 86
  5. 2021Breaking the Dilemma of Medical Image-to-image Translationcited by 82
  6. 2023Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Modelscited by 80
  7. 2021Adversarial VQA: A New Benchmark for Evaluating the Robustness of VQA Modelscited by 74
  8. 2024Natural Language Understanding and Inference with MLLM in Visual Question Answering: A Surveycited by 74
  9. 2021Beyond Question-Based Biases: Assessing Multimodal Shortcut Learning in Visual Question Answeringcited by 71
  10. 2021BiaSwap: Removing dataset bias with bias-tailored swapping augmentationcited by 70
  11. 2022All You May Need for VQA are Image Captionscited by 65
  12. 2021The Spotlight: A General Method for Discovering Systematic Errors in Deep Learning Modelscited by 63
38 more
  1. 2021Human-Adversarial Visual Question Answeringcited by 57
  2. 2023COLA: A Benchmark for Compositional Text-to-image Retrievalcited by 53
  3. 2022Counterfactual Cycle-Consistent Learning for Instruction Following and Generation in Vision-Language Navigationcited by 50
  4. 2023Bi-directional Training for Composed Image Retrieval via Text Prompt Learningcited by 43
  5. 2021VX2TEXT: End-to-End Learning of Video-Based Text Generation From Multimodal Inputscited by 42
  6. 2023BenchLMM: Benchmarking Cross-style Visual Capability of Large Multimodal Modelscited by 41
  7. 2024One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Modelscited by 39
  8. 2022Hierarchical Local-Global Transformer for Temporal Sentence Groundingcited by 38
  9. 2021Back to Square One: Artifact Detection, Training and Commonsense Disentanglement in the Winograd Schemacited by 32
  10. 2022CLIP-TD: CLIP Targeted Distillation for Vision-Language Taskscited by 32
  11. 2024Consistency and Uncertainty: Identifying Unreliable Responses From Black-Box Vision-Language Models for Selective Visual Question Answeringcited by 31
  12. 2022Learning to Answer Visual Questions from Web Videoscited by 30
  13. 2024VideoQA in the Era of LLMs: An Empirical Studycited by 29
  14. 2024Modality-Fair Preference Optimization for Trustworthy MLLM Alignmentcited by 26
  15. 2025Explain Before You Answer: A Survey on Compositional Visual Reasoningcited by 26
  16. 2021Separating Skills and Concepts for Novel Visual Question Answeringcited by 24
  17. 2023Robust Visual Question Answering: Datasets, Methods, and Future Challengescited by 23
  18. 2021Counterfactual Samples Synthesizing and Training for Robust Visual Question Answeringcited by 22
  19. 2022Multimodal Adaptive Distillation for Leveraging Unimodal Encoders for Vision-Language Taskscited by 22
  20. 2024Dynamic Multimodal Evaluation with Flexible Complexity by Vision-Language Bootstrappingcited by 22
  21. 2025Cycle Consistency as Reward: Learning Image-Text Alignment without Human Preferencescited by 22
  22. 2024LOVA3: Learning to Visual Question Answering, Asking and Assessmentcited by 21
  23. 2021Multimodal Integration of Human-Like Attention in Visual Question Answeringcited by 20
  24. 2022VLC-BERT: Visual Question Answering with Contextualized Commonsense Knowledgecited by 20
  25. 2023Q: How to Specialize Large Vision-Language Models to Data-Scarce VQA Tasks? A: Self-Train on Unlabeled Images!cited by 19
  26. 2021LPF: A Language-Prior Feedback Objective Function for De-biased Visual Question Answeringcited by 17
  27. 2022Selecting Better Samples from Pre-trained LLMs: A Case Study on Question Generationcited by 17
  28. 2021VQA-MHUG: A Gaze Dataset to Study Multimodal Neural Attention in Visual Question Answeringcited by 16
  29. 2021Are VQA Systems RAD? Measuring Robustness to Augmented Data with Focused Interventionscited by 15
  30. 2022Measuring Compositional Consistency for Video Question Answeringcited by 15
  31. 2024Improving Generalization in Visual Reasoning via Self-Ensemblecited by 15
  32. 2021Discovering the Unknown Knowns: Turning Implicit Knowledge in the Dataset into Explicit Training Examples for Visual Question Answeringcited by 14
  33. 2022A Feature-space Multimodal Data Augmentation Technique for Text-video Retrievalcited by 13
  34. 2022Why Did the Chicken Cross the Road? Rephrasing and Analyzing Ambiguous Questions in VQAcited by 13
  35. 2021Contrastive Fine-tuning Improves Robustness for Neural Rankerscited by 12
  36. 2022TAG: Boosting Text-VQA via Text-aware Visual Question-answer Generationcited by 12
  37. 2021Domain-robust VQA with diverse datasets and methods but no target labelscited by 11
  38. 2022CARETS: A Consistency And Robustness Evaluative Test Suite for VQAcited by 11

50 most cited of 105 All 105 (JSON)