Research map / Vision-language and multimodal / Human and embodied modeling
Learning Dictionary for Visual Attention
Read the paperCheck 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 |
|---|---|---|---|
| replace SA with Dic-Attn | ViT · CIFAR-10 | Top-1 accuracy: 90.91 -> 91.55 | Helped |
| replace SA with Dic-Attn | ViT · CIFAR-100 | Top-1 accuracy: 66.67 -> 68.79 | Helped |
| replace SA with Dic-Attn | ViT · CIFAR-100 parameters | Params (M): 22.74 -> 19.33 | Helped |
| replace SA with Dic-Attn | PCT · ModelNet40 | Top-1 accuracy: 92.95 -> 94.60 | Helped |
| replace SA with Dic-Attn | PCT · ScanObjectNN | Top-1 accuracy: 80.60 -> 88.96 | Helped |
| add Dic-Attn to Point-MAE | Point-MAE · ModelNet40 | Top-1 accuracy: 93.80 -> 94.46 | Helped |
| add Dic-Attn to Point-MAE | Point-MAE · ScanObjectNN | Top-1 accuracy: 88.30 -> 89.41 | Helped |
| replace SA with Dic-Attn | Segmenter · ADE20K | mIoU: +1.00% | Helped |
| replace SA with Dic-Attn | Segmenter · ADE20K | mean pixel accuracy: +1.25% | Helped |
| replace SA with Dic-Attn | RVT · CIFAR-100 clean | accuracy: 64.80 -> 55.93 | Hurt |
| replace SA with Dic-Attn | RVT · CIFAR-100 FGSM | attack accuracy: 0.20 -> 10.00 | Helped |
| replace SA with Dic-Attn | attention module · efficiency | MACs (G): 292.00 -> 20.00 | Helped |
Closest papers by meaning
- Sparsifiner: Learning Sparse Instance-Dependent Attention for Efficient Vision Transformers (CVPR 2023 · similarity 0.86)
- Attention-aware Inference Optimizations for Large Vision-Language Models with Memory-efficient Decoding (CVPR 2026 · similarity 0.84)
- Bridging the Divide: Reconsidering Softmax and Linear Attention (NeurIPS 2024 · similarity 0.84)
- LookHere: Vision Transformers with Directed Attention Generalize and Extrapolate (NeurIPS 2024 · similarity 0.84)
- CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up (NeurIPS 2025 · similarity 0.84)
- Learning 1D Causal Visual Representation with De-focus Attention Networks (NeurIPS 2024 · similarity 0.84)
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.