Research map / Vision-language and multimodal / Visual question answering
Multi-Level Attention Networks for Visual Question Answering
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 |
|---|---|---|---|
| combine question-guided semantic attention and visual attention (ablation) | MLAN · VQA test-dev | accuracy: 62.50 -> 63.69 | Helped |
| add question-guided semantic attention to concept features (ablation) | Q + Concept · VQA test-dev | accuracy: 56.62 -> 59.28 | Helped |
| add context-aware region representations using bidirectional GRU (ablation) | Q + Visual Attention · VQA test-dev | accuracy: 62.29 -> 62.50 | Helped |
| use element-wise multiplication rather than addition to align question and region features (ablation) | visual attention · VQA test-dev | accuracy: 58.68 -> 62.29 | Helped |
| use MLAN rather than MCB+Att. | Visual7W · Visual7W test | accuracy (Avg): 62.2 -> 62.4 | Helped |
| use MLAN rather than MCB+Att. | Visual7W · Visual7W test, question types | accuracy: 60.3 -> 60.5; 70.4 -> 71.2; 79.5 -> 79.6; 69.2 -> 69.4; 58.2 -> 58.0; 51.1 -> 50.8 | Mixed |
Closest papers by meaning
- Hierarchical Question-Image Co-Attention for Visual Question Answering (NeurIPS 2016 · similarity 0.94)
- Multi-Question Learning for Visual Question Answering (AAAI 2020 · similarity 0.93)
- Deep Modular Co-Attention Networks for Visual Question Answering (CVPR 2019 · similarity 0.93)
- Stacked Attention Networks for Image Question Answering (CVPR 2016 · similarity 0.93)
- Re-Attention for Visual Question Answering (AAAI 2020 · similarity 0.92)
- FOCUS: Internal MLLM Representations for Efficient Fine-Grained Visual Question Answering (NeurIPS 2025 · similarity 0.92)
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.