Research map / Label-efficient and robust learning / Recognition and long-tail
Proxy Anchor Loss for Deep Metric Learning
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 |
|---|---|---|---|
| replace MS with Proxy-Anchor | Inception-BN metric learning model · CUB-200-2011 | Recall@1: 65.7 -> 68.4 | Helped |
| replace SoftTriple with Proxy-Anchor | Inception-BN metric learning model · Cars-196 | Recall@1: 84.5 -> 86.1 | Helped |
| replace SoftTriple with Proxy-Anchor | Inception-BN metric learning model · SOP | Recall@1: 78.3 -> 79.1 | Helped |
| replace MS with Proxy-Anchor | Inception-BN metric learning model · In-Shop | Recall@1: 89.7 -> 91.5 | Helped |
| increase batch size from 30 to 180 (ablation) | Proxy-Anchor · CUB-200-2011 | Recall@1: 65.9 -> 69.0 | Helped |
| increase batch size from 30 to 150 (ablation) | Proxy-Anchor · Cars-196 | Recall@1: 84.6 -> 86.4 | Helped |
| increase batch size from 30 to 300 (ablation) | Proxy-Anchor · SOP | Recall@1: 76.0 -> 79.3 | Helped |
| increase batch size from 30 to 300 (ablation) | Proxy-Anchor · In-Shop | Recall@1: 91.3 -> 92.0 | Helped |
| increase batch size from 300 to 600 (ablation) | Proxy-Anchor · SOP | Recall@1: 79.3 -> 79.3 | No clear effect |
| increase batch size from 300 to 600 (ablation) | Proxy-Anchor · In-Shop | Recall@1: 92.0 -> 91.7 | Hurt |
| increase image size from 224 × 224 to 448 × 448 (ablation) | Proxy-Anchor with Inception-BN · CUB-200-2011 | Recall@1: 68.4 -> 77.3 | Helped |
| increase image size from 224 × 224 to 448 × 448 (ablation) | Proxy-Anchor with Inception-BN · Cars-196 | Recall@1: 86.1 -> 92.9 | Helped |
Closest papers by meaning
- Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation (AAAI 2024 · similarity 0.88)
- Deep Disentangled Metric Learning (AAAI 2025 · similarity 0.87)
- Mean Field Theory in Deep Metric Learning (ICLR 2024 · similarity 0.84)
- Learning Semantic Proxies from Visual Prompts for Parameter-Efficient Fine-Tuning in Deep Metric Learning (ICLR 2024 · similarity 0.84)
- Threshold-Consistent Margin Loss for Open-World Deep Metric Learning (ICLR 2024 · similarity 0.84)
- Potential Field Based Deep Metric Learning (CVPR 2025 · 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.