Research map / Vision-language and multimodal / Visual question answering
Fully Authentic Visual Question Answering Dataset from Online Communities
Chongyan Chen, Mengchen Liu, Noel C Codella, Yunsheng Li, Lu Yuan, Danna Gurari
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add one in-context exemplar (random or topic-matched, with or without image) (ablation) | mPLUG-Owl · VQAonline test | METEOR: 0.12 -> 0.108 (best one-shot variant) | Hurt |
| use topic-matched instead of random one-shot exemplar (ablation) | mPLUG-Owl one-shot · VQAonline test | METEOR: 0.106 -> 0.108 (no image) | Improved |
| exclude images from one-shot exemplars (ablation) | mPLUG-Owl one-shot · VQAonline test | METEOR: 0.105 -> 0.108 (matched topic) | Improved |
| add context input (ablation) | mPLUG-Owl · VQAonline test | METEOR: 0.07 -> 0.12 | Improved |
| add question input (ablation) | mPLUG-Owl · VQAonline test | ROUGE-L: 0.14 -> 0.16 | Improved |
| add image input (ablation) | mPLUG-Owl · VQAonline test | METEOR: 0.11 -> 0.12 | Improved |
| remove context input (Q+I) when scored by CLIP-S (ablation) | mPLUG-Owl · VQAonline test | CLIP-S: 0.72 -> 0.73 | Improved |
| evaluate on Science versus Culture & Recreation topics | mPLUG-Owl · VQAonline test | METEOR: 0.11 -> 0.14 | Improved |
| use LLaMA2 prompted scoring instead of CLIP-S as evaluation metric | evaluation metrics · Human judgment alignment | Pearson correlation: 0.662 -> 0.975 | Improved |
Closest papers by meaning
- Answer Them All! Toward Universal Visual Question Answering Models (CVPR 2019 · similarity 0.92)
- FVQA: Fact-Based Visual Question Answering (TPAMI 2017 · similarity 0.91)
- VQA: Visual Question Answering (IJCV 2016 · similarity 0.90)
- Robust Visual Question Answering: Datasets, Methods, and Future Challenges (TPAMI 2024 · similarity 0.90)
- KVQA: Knowledge-Aware Visual Question Answering (AAAI 2019 · similarity 0.90)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.88)
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.
