Research map / Vision-language and multimodal / Visual question answering
IVQA: Inverse Visual Question Answering
Feng Liu, Tao Xiang, Timothy M. Hospedales, Wankou Yang, Changyin Sun
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 dynamic multi-modal attention to the image-and-answer model (ablation) | iVQA · iVQA validation | Acc@1: 29.929 -> 32.547 | Improved |
| add global semantic image features (ablation) | iVQA with dynamic attention · iVQA validation | Acc@1: 32.547 -> 33.636 | Improved |
| add dynamic multi-modal attention and global semantic image features (ablation) | I+A · iVQA validation | CIDEr: 1.541 -> 1.750 | Improved |
| combine VQA answers with iVQA scores using score fusion | MLB-att · VQA 2.0 validation | accuracy: 57.85 -> 58.86 | Improved |
| combine VQA answers with iVQA scores using score fusion | MLB-att · VQA 2.0 validation, number type | accuracy: 34.94 -> 38.71 | Improved |
| use image and answer rather than answer alone | iVQA · iVQA test | Acc@1: 14.589 -> 32.899 | Improved |
| provide the answer type in addition to the image | iVQA image-only model · iVQA test | Acc@1: 13.012 -> 20.277 | Improved |
| use image-plus-answer-type instead of the full image-and-answer model | iVQA · iVQA test | Acc@1: 20.277 -> 32.899 | Hurt |
| use the proposed dynamic multi-modal iVQA model instead of SAT | iVQA · iVQA test | Acc@1: 29.722 -> 32.899 | Improved |
| use the proposed dynamic multi-modal iVQA model instead of VQG+VQA | iVQA · iVQA test | Acc@1: 16.529 -> 32.899 | Improved |
Closest papers by meaning
- Inverse Visual Question Answering: A New Benchmark and VQA Diagnosis Tool (TPAMI 2018 · similarity 0.96)
- An Analysis of Visual Question Answering Algorithms (ICCV 2017 · similarity 0.92)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.91)
- Visual Question Answering as a Meta Learning Task (ECCV 2018 · similarity 0.91)
- Answer Them All! Toward Universal Visual Question Answering Models (CVPR 2019 · similarity 0.91)
- A-OKVQA: A Benchmark for Visual Question Answering Using World Knowledge (ECCV 2022 · similarity 0.91)
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.
