mlsanityData to 7 Oct 2026

Adversarial VQA: A New Benchmark for Evaluating the Robustness of VQA Models

arXiv 2106.00245 ↗First posted 1 Jun 2021cs.CV cs.CLcited by 74

Not on the map. Why?

Cited by 25 in Multimodal LLMs / vision-language models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Benefiting from large-scale pre-training, we have witnessed significant performance boost on the popular Visual Question Answering (VQA) task. Despite rapid progress, it remains unclear whether these state-of-the-art (SOTA) models are robust when encountering examples in the wild. To study this, we introduce Adversarial VQA, a new large-scale VQA benchmark, collected iteratively via an adversarial human-and-model-in-the-loop procedure. Through this new benchmark, we discover several interesting findings. (i) Surprisingly, we find that during dataset collection, non-expert annotators can easily attack SOTA VQA models successfully. (ii) Both large-scale pre-trained models and adversarial training methods achieve far worse performance on the new benchmark than over standard VQA v2 dataset, revealing the fragility of these models while demonstrating the effectiveness of our adversarial dataset. (iii) When used for data augmentation, our dataset can effectively boost model performance on other robust VQA benchmarks. We hope our Adversarial VQA dataset can shed new light on robustness study in the community and serve as a valuable benchmark for future work.

Key facts. Cited by 74 papers in the map (data through 7 Oct 2026). Builds on 50 of its 57 references that are in the map. First posted on arXiv 1 Jun 2021.

Builds on

50 of 57 references.

References · 50

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  4. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  5. 2019XLNet: Generalized Autoregressive Pretraining for Language Understandingcited by 2,801
  6. 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,772
  7. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,665
  8. 2016Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotationscited by 2,646
  9. 2016Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answeringcited by 2,498
  10. 2015VQA: Visual Question Answeringcited by 2,489
  11. 2019ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Taskscited by 2,219
  12. 2015A large annotated corpus for learning natural language inferencecited by 1,837
38 more
  1. 2019LXMERT: Learning Cross-Modality Encoder Representations from Transformerscited by 1,548
  2. 2015Microsoft COCO Captions: Data Collection and Evaluation Servercited by 1,447
  3. 2017Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answeringcited by 1,310
  4. 2014Deep Visual-Semantic Alignments for Generating Image Descriptionscited by 966
  5. 2019VL-BERT: Pre-training of Generic Visual-Linguistic Representationscited by 954
  6. 2019Adversarial NLI: A New Benchmark for Natural Language Understandingcited by 798
  7. 2019Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailmentcited by 663
  8. 2018Annotation Artifacts in Natural Language Inference Datacited by 656
  9. 2018From Recognition to Cognition: Visual Commonsense Reasoningcited by 553
  10. 2021Less is More: ClipBERT for Video-and-Language Learning via Sparse Samplingcited by 484
  11. 2019Deep Modular Co-Attention Networks for Visual Question Answeringcited by 338
  12. 2017Don't Just Assume; Look and Answer: Overcoming Priors for Visual Question Answeringcited by 337
  13. 2018SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inferencecited by 333
  14. 2020Large-Scale Adversarial Training for Vision-and-Language Representation Learningcited by 327
  15. 2018Hypothesis Only Baselines in Natural Language Inferencecited by 291
  16. 2020Pixel-BERT: Aligning Image Pixels with Text by Deep Multi-Modal Transformerscited by 288
  17. 2019Adversarial Examples Improve Image Recognitioncited by 283
  18. 2020In Defense of Grid Features for Visual Question Answeringcited by 155
  19. 2020Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehensioncited by 131
  20. 2019Relation-Aware Graph Attention Network for Visual Question Answeringcited by 119
  21. 2018Dynamic Fusion with Intra- and Inter- Modality Attention Flow for Visual Question Answeringcited by 113
  22. 2021VinVL: Revisiting Visual Representations in Vision-Language Modelscited by 109
  23. 2019Towards Causal VQA: Revealing and Reducing Spurious Correlations by Invariant and Covariant Semantic Editingcited by 108
  24. 2019Cycle-Consistency for Robust Visual Question Answeringcited by 105
  25. 2020Behind the Scene: Revealing the Secrets of Pre-trained Vision-and-Language Modelscited by 90
  26. 2019GQA: A New Dataset for Real-World Visual Reasoning and Compositional Question Answeringcited by 77
  27. 2020On the Value of Out-of-Distribution Testing: An Example of Goodhart's Lawcited by 74
  28. 2020Roses Are Red, Violets Are Blue... but Should Vqa Expect Them To?cited by 69
  29. 2021UC2: Universal Cross-lingual Cross-modal Vision-and-Language Pre-trainingcited by 64
  30. 2018Pythia v0.1: the Winning Entry to the VQA Challenge 2018cited by 62
  31. 2020What is More Likely to Happen Next? Video-and-Language Future Event Predictioncited by 61
  32. 2019r/Fakeddit: A New Multimodal Benchmark Dataset for Fine-grained Fake News Detectioncited by 53
  33. 2020VQA-LOL: Visual Question Answering under the Lens of Logiccited by 48
  34. 2021LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrievalcited by 44
  35. 2020A Closer Look at the Robustness of Vision-and-Language Pre-trained Modelscited by 37
  36. 2020MoVie: Revisiting Modulated Convolutions for Visual Counting and Beyondcited by 11
  37. 2017Mastering the Dungeon: Grounded Language Learning by Mechanical Turker Descentcited by 10
  38. 2020SQuINTing at VQA Models: Introspecting VQA Models with Sub-Questionscited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 6; 2022: 15; 2023: 18; 2024: 11; 2025: 14; 2026: 10.

74 papers.

Sort:

By area:

  1. 2022Multimodal Learning with Transformers: A Surveycited by 399
  2. 2023On Evaluating Adversarial Robustness of Large Vision-Language Modelscited by 281
  3. 2022Vision-Language Pre-training: Basics, Recent Advances, and Future Trendscited by 117
  4. 2024NaturalBench: Evaluating Vision-Language Models on Natural Adversarial Samplescited by 94
  5. 2023On the Robustness of Large Multimodal Models Against Image Adversarial Attackscited by 89
  6. 2022Video Question Answering: Datasets, Algorithms and Challengescited by 84
  7. 2023Negative Object Presence Evaluation (NOPE) to Measure Object Hallucination in Vision-Language Modelscited by 80
  8. 2021Human-Adversarial Visual Question Answeringcited by 57
  9. 2023Equivariant Similarity for Vision-Language Foundation Modelscited by 51
  10. 2023Benchmarks for Automated Commonsense Reasoning: A Surveycited by 40
  11. 2024An Image Is Worth 1000 Lies: Adversarial Transferability across Prompts on Vision-Language Modelscited by 40
  12. 2024Do Vision-Language Models Really Understand Visual Language?cited by 29
38 more
  1. 2021Coarse-to-Fine Reasoning for Visual Question Answeringcited by 25
  2. 2022Causal Reasoning Meets Visual Representation Learning: A Prospective Studycited by 24
  3. 2022Are Deep Neural Networks SMARTer than Second Graders?cited by 24
  4. 2023Improving Selective Visual Question Answering by Learning from Your Peerscited by 24
  5. 2022Generalized but not Robust? Comparing the Effects of Data Modification Methods on Out-of-Domain Generalization and Adversarial Robustnesscited by 23
  6. 2023Robust Visual Question Answering: Datasets, Methods, and Future Challengescited by 23
  7. 2023UNK-VQA: A Dataset and a Probe into the Abstention Ability of Multi-modal Large Modelscited by 23
  8. 2022A Survey of Adversarial Defences and Robustness in NLPcited by 22
  9. 2022Reassessing Evaluation Practices in Visual Question Answering: A Case Study on Out-of-Distribution Generalizationcited by 20
  10. 2022WinoGAViL: Gamified Association Benchmark to Challenge Vision-and-Language Modelscited by 20
  11. 2023Q: How to Specialize Large Vision-Language Models to Data-Scarce VQA Tasks? A: Self-Train on Unlabeled Images!cited by 19
  12. 2024MirrorCheck: Efficient Adversarial Defense for Vision-Language Modelscited by 19
  13. 2024Look, Listen, and Answer: Overcoming Biases for Audio-Visual Question Answeringcited by 18
  14. 2022Dynatask: A Framework for Creating Dynamic AI Benchmark Taskscited by 16
  15. 2024Visual Robustness Benchmark for Visual Question Answering (VQA)cited by 12
  16. 2025Code-Vision: Evaluating Multimodal LLMs Logic Understanding and Code Generation Capabilitiescited by 11
  17. 2023Adventures of Trustworthy Vision-Language Models: A Surveycited by 10
  18. 2021Semantically Distributed Robust Optimization for Vision-and-Language Inferencecited by 9
  19. 2022MaXM: Towards Multilingual Visual Question Answeringcited by 9
  20. 2025UniGame: Turning a Unified Multimodal Model Into Its Own Adversarycited by 9
  21. 2022Correlation Information Bottleneck: Towards Adapting Pretrained Multimodal Models for Robust Visual Question Answeringcited by 7
  22. 2023RoCOCO: Robustness Benchmark of MS-COCO to Stress-test Image-Text Matching Modelscited by 7
  23. 2026Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volumecited by 7
  24. 2021Unsupervised Natural Language Inference Using PHL Triplet Generationcited by 6
  25. 2021MLP Architectures for Vision-and-Language Modeling: An Empirical Studycited by 6
  26. 2023Reconstruction Distortion of Learned Image Compression with Imperceptible Perturbationscited by 5
  27. 2024Benchmark Granularity and Model Robustness for Image-Text Retrievalcited by 5
  28. 2024Identifying Crucial Objects in Blind and Low-Vision Individuals' Navigationcited by 5
  29. 2025FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answeringcited by 5
  30. 2022Robustness of Fusion-based Multimodal Classifiers to Cross-Modal Content Dilutionscited by 4
  31. 2024A Dataset for Crucial Object Recognition in Blind and Low-Vision Individuals' Navigationcited by 4
  32. 2024A Hitchhikers Guide to Fine-Grained Face Forgery Detection Using Common Sense Reasoningcited by 4
  33. 2025TabAttackBench: A Benchmark for Adversarial Attacks on Tabular Datacited by 4
  34. 2025Mitigating Easy Option Bias in Multiple-Choice Question Answeringcited by 4
  35. 2025The Artificial Intelligence Cognitive Examination: A Survey on the Evolution of Multimodal Evaluation from Recognition to Reasoningcited by 4
  36. 2022Measuring CLEVRness: Blackbox testing of Visual Reasoning Modelscited by 3
  37. 2023Cross-Modal Attribute Insertions for Assessing the Robustness of Vision-and-Language Learningcited by 3
  38. 2024Evaluating Durability: Benchmark Insights into Multimodal Watermarkingcited by 3

50 most cited of 74 All 74 (JSON)