Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Asymmetric Metric Learning via Rich Relationship Mining
Xinyi Xu, Yanhua Yang, Cheng Deng, Feng Zheng
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 |
|---|---|---|---|
| Use DAMLRRM instead of AML | deep metric learning · CUB-200-2011 test | R@1: 52.7 -> 55.1 | Improved |
| Use DAMLRRM instead of AML | deep metric learning · Cars196 test | R@1: 72.5 -> 73.5 | Improved |
| Use DAMLRRM instead of AML | deep metric learning · Stanford Online Products test | R@1: 66.3 -> 69.7 | Improved |
| Add asymmetric data streams and their shuffled loss (ablation) | DAMLRRM · CUB-200-2011 test | R@1: 52.3 -> 55.1 | Improved |
| Add graph loss with minimum-cost tree positive pairs (ablation) | DAMLRRM · CUB-200-2011 test | R@1: 52.3 -> 55.1 | Improved |
| Use minimum-cost spanning tree rather than fully combined positive pairs (ablation) | graph-loss metric learning · CUB-200-2011 test | R@1: 48.2 -> 51.2 | Improved |
| Set boundary α to 30 (ablation) | DAMLRRM · CUB-200-2011 test | R@1: 52.9 -> 55.1 | Improved |
| Set margin β to 0.5 (ablation) | DAMLRRM · CUB-200-2011 test | R@1: 51.9 -> 55.1 | Improved |
| Use DAMLRRM rather than one-stream Lifted or N-pair sampling (ablation) | metric learning with limited sampling · CUB-200-2011 test | R@1: 46.9 -> 52.3; 51.0 -> 52.3 | Improved |
| Use DAMLRRM instead of AML | deep metric learning · Stanford Online Products test | NMI: 89.1 -> 88.2 | Hurt |
| Use DAMLRRM instead of AML | deep metric learning · Stanford Online Products test | F1: 31.7 -> 30.5 | Hurt |
Closest papers by meaning
- Asymmetric Metric Learning for Knowledge Transfer (CVPR 2021 · similarity 0.86)
- No Fuss Distance Metric Learning Using Proxies (ICCV 2017 · similarity 0.86)
- Deep Metric Learning by Online Soft Mining and Class-Aware Attention (AAAI 2019 · similarity 0.86)
- Multi-Similarity Loss With General Pair Weighting for Deep Metric Learning (CVPR 2019 · similarity 0.85)
- Improving Deep Metric Learning by Divide and Conquer (TPAMI 2021 · similarity 0.85)
- Deep Disentangled Metric Learning (AAAI 2025 · similarity 0.85)
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.
