Research map / Vision-language and multimodal / Visual question answering
Ask, Attend and Answer: Exploring Question-Guided Spatial Attention for Visual Question Answering
Huijuan Xu, Kate Saenko
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 a second attention hop over the whole question (ablation) | SMem-VQA · DAQUAR | accuracy (%): 36.03 -> 40.07 | Improved |
| add a second attention hop over the whole question (ablation) | SMem-VQA · VQA test-dev | overall accuracy (%): 56.56 -> 57.99 | Improved |
| add spatial attention (Two-Hop) over iBOWIMG | iBOWIMG · VQA test-standard | overall accuracy (%): 2.35% | Improved |
| add spatial attention (Two-Hop) over iBOWIMG | iBOWIMG · VQA test-dev | overall accuracy (%): 2.27% | Improved |
| add a third attention hop (ablation) | SMem-VQA Two-Hop · VQA | accuracy | No clear effect |
| add one-hop word-guided spatial attention over mean-pooled features | iBOWIMG · Synthetic absolute position | accuracy (%): 75 -> 100 | Improved |
| add one-hop word-guided spatial attention over mean-pooled features | iBOWIMG · Synthetic relative position (cat images) | accuracy (%): 75 -> 96 | Improved |
| add one-hop spatial attention instead of mean pooling | iBOWIMG · VQA test-dev | overall accuracy (%): 55.72 -> 56.56 | Improved |
| add one-hop spatial attention instead of image concatenation | IMG+BOW · DAQUAR | accuracy (%): 34.17 -> 36.03 | Improved |
Closest papers by meaning
- Multi-Level Attention Networks for Visual Question Answering (CVPR 2017 · similarity 0.94)
- Hierarchical Question-Image Co-Attention for Visual Question Answering (NeurIPS 2016 · similarity 0.93)
- Image Captioning and Visual Question Answering Based on Attributes and External Knowledge (TPAMI 2017 · similarity 0.91)
- Structured Attentions for Visual Question Answering (ICCV 2017 · similarity 0.91)
- An Analysis of Visual Question Answering Algorithms (ICCV 2017 · similarity 0.91)
- FOCUS: Internal MLLM Representations for Efficient Fine-Grained Visual Question Answering (NeurIPS 2025 · similarity 0.89)
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.
