Research map / Label-efficient and robust learning / Recognition and long-tail
Hardness-Aware Deep Metric Learning
Wenzhao Zheng, Zhaodong Chen, Jiwen Lu, Jie Zhou
Check 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 hardness-aware synthesis to N-pair loss | N-pair metric learning · CUB-200-2011 | NMI: 60.2 -> 62.6 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Cars196 | NMI: 62.7 -> 69.7 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Stanford Online Products | NMI: 87.9 -> 89.3 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · CUB-200-2011 | F1: 28.2 -> 31.6 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Cars196 | F1: 31.8 -> 41.6 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Stanford Online Products | F1: 27.1 -> 32.2 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · CUB-200-2011 | R@1: 51.9 -> 53.7 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Cars196 | R@1: 68.9 -> 79.1 | Improved |
| Add hardness-aware synthesis to N-pair loss | N-pair metric learning · Stanford Online Products | R@1: 66.4 -> 68.7 | Improved |
| Use HDML rather than DAML with N-pair loss | N-pair metric learning · Stanford Online Products | NMI: 89.4 -> 89.3 | Hurt |
| Use HDML rather than DAML with N-pair loss | N-pair metric learning · Stanford Online Products | F1: 32.4 -> 32.2 | Hurt |
| Remove the softmax label-preservation loss (ablation) | HDML with N-pair loss · Cars196 | clustering and retrieval performance | Hurt |
Closest papers by meaning
- LoOp: Looking for Optimal Hard Negative Embeddings for Deep Metric Learning (ICCV 2021 · similarity 0.87)
- Unsupervised Metric Learning with Synthetic Examples (AAAI 2020 · similarity 0.86)
- Smart Mining for Deep Metric Learning (ICCV 2017 · similarity 0.86)
- Improving Deep Metric Learning by Divide and Conquer (TPAMI 2021 · similarity 0.86)
- Symmetrical Synthesis for Deep Metric Learning (AAAI 2020 · similarity 0.85)
- Deep Disentangled Metric Learning (AAAI 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.
