mlsanityData to 13 Oct 2026

Do Explanations make VQA Models more Predictable to a Human?

arXiv 1810.12366 ↗First posted 29 Oct 2018cs.AI cs.CL cs.CVcited by 52

Not on the map. Why?

Abstract (arXiv)

A rich line of research attempts to make deep neural networks more transparent by generating human-interpretable 'explanations' of their decision process, especially for interactive tasks like Visual Question Answering (VQA). In this work, we analyze if existing explanations indeed make a VQA model -- its responses as well as failures -- more predictable to a human. Surprisingly, we find that they do not. On the other hand, we find that human-in-the-loop approaches that treat the model as a black-box do.

Key facts. Cited by 52 papers in the map (data through 13 Oct 2026). First posted on arXiv 29 Oct 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2021: 12; 2022: 13; 2023: 9; 2024: 11; 2025: 4; 2026: 3.

52 papers.

Sort:

By area:

  1. 2025Reasoning Models Don't Always Say What They Thinkcited by 404
  2. 2022Foundations and Trends in Multimodal Machine Learning: Principles, Challenges, and Open Questionscited by 116
  3. 2021What I Cannot Predict, I Do Not Understand: A Human-Centered Evaluation Framework for Explainability Methodscited by 98
  4. 2022Interpreting Language Models with Contrastive Explanationscited by 92
  5. 2021The effectiveness of feature attribution methods and its correlation with automatic evaluation scorescited by 88
  6. 2021Towards a Science of Human-AI Decision Making: A Survey of Empirical Studiescited by 86
  7. 2023Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanationscited by 85
  8. 2022Towards Human-centered Explainable AI: A Survey of User Studies for Model Explanationscited by 84
  9. 2023Appropriate Reliance on AI Advice: Conceptualization and the Effect of Explanationscited by 83
  10. 2023In Search of Verifiability: Explanations Rarely Enable Complementary Performance in AI-Advised Decision Makingcited by 50
  11. 2021A Review on Explainability in Multimodal Deep Neural Netscited by 46
  12. 2022MultiViz: Towards Visualizing and Understanding Multimodal Modelscited by 36
38 more
  1. 2022DIME: Fine-grained Interpretations of Multimodal Models via Disentangled Local Explanationscited by 34
  2. 2021Teaching Humans When To Defer to a Classifier via Exemplarscited by 29
  3. 2022A Meta-Analysis of the Utility of Explainable Artificial Intelligence in Human-AI Decision-Makingcited by 29
  4. 2022Should I Follow AI-based Advice? Measuring Appropriate Reliance in Human-AI Decision-Makingcited by 28
  5. 2021Explain, Edit, and Understand: Rethinking User Study Design for Evaluating Model Explanationscited by 26
  6. 2021Connecting Attributions and QA Model Behavior on Realistic Counterfactualscited by 20
  7. 2021Can Explanations Be Useful for Calibrating Black Box Models?cited by 20
  8. 2022Machine Explanations and Human Understandingcited by 20
  9. 2022Learning to Scaffold: Optimizing Model Explanations for Teachingcited by 20
  10. 2023Effective Human-AI Teams via Learned Natural Language Rules and Onboardingcited by 20
  11. 2021VisQA: X-raying Vision and Language Reasoning in Transformerscited by 19
  12. 2022Use-Case-Grounded Simulations for Explanation Evaluationcited by 19
  13. 2023Can Language Models Teach Weaker Agents? Teacher Explanations Improve Students via Personalizationcited by 16
  14. 2022Saliency Map Verbalization: Comparing Feature Importance Representations from Model-free and Instruction-based Methodscited by 13
  15. 2023Are Human Explanations Always Helpful? Towards Objective Evaluation of Human Natural Language Explanationscited by 13
  16. 2021Improving Human Sequential Decision-Making with Reinforcement Learningcited by 12
  17. 2023Patchwork Learning: A Paradigm Towards Integrative Analysis across Diverse Biomedical Data Sourcescited by 10
  18. 2022Calibrating Trust of Multi-Hop Question Answering Systems with Decompositional Probescited by 9
  19. 2024Utilizing Human Behavior Modeling to Manipulate Explanations in AI-Assisted Decision Making: The Good, the Bad, and the Scarycited by 9
  20. 2024To Err Is AI! Debugging as an Intervention to Facilitate Appropriate Reliance on AI Systemscited by 7
  21. 2025Evaluating Model Explanations without Ground Truthcited by 7
  22. 2026Do LLM Self-Explanations Help Users Predict Model Behavior? Evaluating Counterfactual Simulatability with Pragmatic Perturbationscited by 7
  23. 2024How explainable AI affects human performance: A systematic review of the behavioural consequences of saliency mapscited by 6
  24. 2021Improving Users' Mental Model with Attention-directed Counterfactual Editscited by 5
  25. 2023Efficient Shapley Values Estimation by Amortization for Text Classificationcited by 4
  26. 2024XForecast: Evaluating Natural Language Explanations for Time Series Forecastingcited by 4
  27. 2021Generating and Evaluating Explanations of Attended and Error-Inducing Input Regions for VQA Modelscited by 2
  28. 2024Multimodal Rationales for Explainable Visual Question Answeringcited by 2
  29. 2024OpenHEXAI: An Open-Source Framework for Human-Centered Evaluation of Explainable Machine Learningcited by 2
  30. 2024Explainable Interface for Human-Autonomy Teaming: A Surveycited by 2
  31. 2024Interpretable Tensor Fusioncited by 2
  32. 2024A Sim2Real Approach for Identifying Task-Relevant Properties in Interpretable Machine Learningcited by 2
  33. 2023Quantifying the Intrinsic Usefulness of Attributional Explanations for Graph Neural Networks with Artificial Simulatability Studiescited by 1
  34. 2024SimpsonsVQA: Enhancing Inquiry-Based Learning with a Tailored Datasetcited by 1
  35. 2025Explainable AI the Latest Advancements and New Trendscited by 1
  36. 2026Not All Explanations Simulate Equally: Comparing Verbalized Feature Attributions and Self-Generated Rationalescited by 1
  37. 2022Tensions Between the Proxies of Human Values in AIcited by 0
  38. 2024Underspecified Human Decision Experiments Considered Harmfulcited by 0

50 most cited of 52 All 52 (JSON)