Research map / Label-efficient and robust learning / Knowledge distillation
Semi-attention Partition for Occluded Person Re-identification
Mengxi Jia, Yifan Sun, Yunpeng Zhai, Xinhua Cheng, Yi Yang, Ying Li
This paper tried a common change: Add dropout
add supervision semantic dropout · Occluded-Duke · mAP: 61.8 -> 62.2 Improved
What 795 accepted papers report for this change
- 549 improved
- 87 no clear effect
- 12 mixed
- 147 hurt
- Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference ICML 2026 · apply layer dropout to whole transformer blocks instead of attention and FFN sub-blocks separately · 20 TPP val loss, 271M · val loss: 2.377 -> 2.367Improved
- Quantifying the Uncertainty of Foundation Models with Singular Value Ensembles ICML 2026 · use MC Dropout · vision and SST-2 · accuracy: 86.3 -> 76.6 on Flowers102Hurt
- Motion Planning in Compressed Representation Spaces ICML 2026 · add nested dropout during training · WOMD guided generation · goal FDE: 4.145 -> 1.104Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 patch-to-patch semi-attention (ablation) | Baseline re-ID model · Occluded-Duke | mAP: 56.0 -> 59.6 | Improved |
| add proxy-to-patch semi-attention (ablation) | Baseline re-ID model · Occluded-Duke | mAP: 56.0 -> 60.1 | Improved |
| add both patch-to-patch and proxy-to-patch semi-attention (ablation) | Baseline re-ID model · Occluded-Duke | mAP: 56.0 -> 62.2 | Improved |
| add mutual absorbing partition refinement (ablation) | SAP · Occluded-Duke | mAP: 60.8 -> 62.2 | Improved |
| add supervision semantic dropout (ablation) | SAP · Occluded-Duke | mAP: 61.8 -> 62.2 | Improved |
| use Pascal rather than ATR or lip for human-parsing teacher pretraining (ablation) | SAP · Occluded-Duke | mAP: 61.4, 61.6 -> 62.2 | Improved |
| use a noisy semantic-partition teacher with SAP distillation (ablation) | Transformer re-ID student · Occluded-Duke | mAP: 56.0 -> 62.2 | Improved |
| use SAP rather than PFD | Occluded-REID · Occluded-REID | mAP: 83.0 -> 76.8 | Hurt |
| use SAP rather than FED | Occluded-REID · Occluded-REID | Rank-1: 86.3 -> 83.0 | Hurt |
| use SAP rather than PFD | Market-1501 · Market-1501 | mAP: 89.7 -> 90.5 | Improved |
| use SAP rather than lip | Market-1501 · Market-1501 | Rank-1: 95.7 -> 96.0 | Improved |
| use SAP rather than TransReID | MSMT17 · MSMT17 | mAP: 67.4 -> 67.8 | Improved |
Closest papers by meaning
- Relation-Aware Global Attention for Person Re-Identification (CVPR 2020 · similarity 0.87)
- Scalable Person Re-Identification by Harmonious Attention (IJCV 2019 · similarity 0.87)
- Harmonious Attention Network for Person Re-Identification (CVPR 2018 · similarity 0.87)
- Second-Order Non-Local Attention Networks for Person Re-Identification (ICCV 2019 · similarity 0.87)
- ABD-Net: Attentive but Diverse Person Re-Identification (ICCV 2019 · similarity 0.87)
- Dynamic Feature Pruning and Consolidation for Occluded Person Re-identification (AAAI 2024 · similarity 0.87)
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.
