Research map / Label-efficient and robust learning / Recognition and long-tail
Dual-Phase Visual-Language Pretraining and Adaptation for Long-Tailed Multi-Label Recognition
Yongcheng Li, Xuekuan Wang, Zhifei Zhang, Cairong Zhao
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 STAG LLM-generated descriptions to GCL baseline (ablation) | DP-VLPA · COCO-LT | mAP: 48.23 -> 64.01 | Improved |
| add STAG LLM-generated descriptions to GCL baseline (ablation) | DP-VLPA · VOC-LT | mAP: 82.84 -> 83.21 | Improved |
| add QCL fine-grained alignment to GCL baseline (ablation) | DP-VLPA · COCO-LT | mAP: 48.23 -> 64.07 | Improved |
| add QCL fine-grained alignment to GCL baseline (ablation) | DP-VLPA · VOC-LT | mAP: 82.84 -> 83.62 | Improved |
| add DQR on top of STAG+QCL (ablation) | DP-VLPA · VOC-LT | mAP: 84.65 -> 88.19 | Improved |
| add DQR on top of STAG+QCL (ablation) | DP-VLPA · COCO-LT | mAP: 70.45 -> 72.21 | Improved |
| add COA loss on top of STAG+QCL+DQR (ablation) | DP-VLPA · VOC-LT | mAP: 88.19 -> 90.72 | Improved |
| add COA loss on top of STAG+QCL+DQR (ablation) | DP-VLPA · COCO-LT | mAP: 72.21 -> 74.42 | Improved |
| use 16 queries instead of 32 (ablation) | DP-VLPA · COCO-LT | mAP: 74.42 -> 69.30 | Hurt |
| use 16 queries instead of 32 (ablation) | DP-VLPA · VOC-LT | mAP: 90.72 -> 90.92 | No clear effect |
| use 64 queries instead of 32 (ablation) | DP-VLPA · COCO-LT | mAP: 74.42 -> 65.44 | Hurt |
| use 64 queries instead of 32 (ablation) | DP-VLPA · VOC-LT | mAP: 90.72 -> 89.48 | Hurt |
Closest papers by meaning
- DualCoOp++: Fast and Effective Adaptation to Multi-Label Recognition With Limited Annotations (TPAMI 2023 · similarity 0.90)
- DualCoOp: Fast Adaptation to Multi-Label Recognition with Limited Annotations (NeurIPS 2022 · similarity 0.89)
- LTGC: Long-tail Recognition via Leveraging LLMs-driven Generated Content (CVPR 2024 · similarity 0.88)
- Recover and Match: Open-Vocabulary Multi-Label Recognition through Knowledge-Constrained Optimal Transport (CVPR 2025 · similarity 0.87)
- From Head to Tail: Towards Balanced Representation in Large Vision-Language Models through Adaptive Data Calibration (CVPR 2025 · similarity 0.87)
- Continuous Contrastive Learning for Long-Tailed Semi-Supervised Recognition (NeurIPS 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.
