Research map / Vision-language and multimodal / Visual question answering
VQA With No Questions-Answers Training
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use question-to-graph parsing and modular visual estimators without question-answer training | UnCoRd · CLEVR test set | overall accuracy: 99.74 vs 99.1 | Helped |
| Use question-to-graph parsing and modular visual estimators without question-answer training | UnCoRd · COCO, non-VQA-v2 questions | overall accuracy: 35.0 -> 76.0 | Helped |
| Use question-to-graph parsing and modular visual estimators without question-answer training | UnCoRd · VQA-v2 sample | overall accuracy: 77.0 -> 92.0 | Helped |
| Add generated question examples to the Minimal mapper training data (ablation) | UnCoRd question-to-graph mapper · CLEVR test set | overall accuracy: 99.68 -> 99.46 | Hurt |
| Add generated question examples to the Extended mapper training data (ablation) | UnCoRd question-to-graph mapper · CLEVR test set | overall accuracy: 94.80 -> 99.49 | Helped |
| Add generated question examples to the VG mapper training data (ablation) | UnCoRd question-to-graph mapper · CLEVR test set | overall accuracy: 94.81 -> 97.49 | Helped |
| Train the mapper with broader VG vocabulary and question variability (ablation) | UnCoRd question-to-graph mapper · VQA validation sample | graph representation accuracy: 1 -> 50 | Helped |
| Train the mapper on VG categories rather than cleaner, more restricted category sets (ablation) | UnCoRd question-to-graph mapper · CLEVR test set | overall accuracy: 99.74 -> 97.49 | Hurt |
| Increase training-data variability with Extended question and vocabulary modifications (ablation) | UnCoRd question-to-graph mapper · CLEVR test set | overall accuracy: 99.74 -> 94.80 (Basic); 99.74 -> 99.49 (Enhanced) | Mixed |
| Use a richer vocabulary and generalized mapper for CLEVR-Humans questions | UnCoRd · CLEVR-Humans test set | accuracy: 57.4 -> 60.59 | Helped |
Closest papers by meaning
- VQ-VA World: Towards High-Quality Visual Question-Visual Answering (CVPR 2026 · similarity 0.92)
- Answer Them All! Toward Universal Visual Question Answering Models (CVPR 2019 · similarity 0.91)
- An Empirical Evaluation of Visual Question Answering for Novel Objects (CVPR 2017 · similarity 0.91)
- Ask Me Anything: Free-Form Visual Question Answering Based on Knowledge From External Sources (CVPR 2016 · similarity 0.91)
- Right this way: Can VLMs Guide Us to See More to Answer Questions? (NeurIPS 2024 · similarity 0.90)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.