Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Metric Learning With Tuplet Margin Loss
Baosheng Yu, Dacheng Tao
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 slack margin beta=0.1 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 89.4 -> 91.5 | Improved |
| increase slack margin from 0.1 to 0.2 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 91.5 -> 85.2 | Hurt |
| add intra-pair variation minimization (lambda=0.5) (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 91.5 -> 93.7 | Improved |
| raise lambda from 0.5 to 1.5 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 92.6 | Hurt |
| raise scale factor s from 1 to 64 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 69.1 -> 93.7 | Improved |
| raise scale factor s from 64 to 128 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 92.1 | Hurt |
| reduce feature dimension from 512 to 128 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 92.3 | No clear effect |
| use ResNet-101 instead of ResNet-50 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 94.3 | Improved |
| increase batch from k=32,n=8 to k=64,n=8 (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 92.3 | Hurt |
| reduce n from 8 to 4 (k=32) (ablation) | tuplet margin loss · CARS196 (cropped) | Recall@1: 93.7 -> 93.2 | No clear effect |
Closest papers by meaning
- Improved Deep Metric Learning with Multi-class N-pair Loss Objective (NeurIPS 2016 · similarity 0.90)
- Rethinking Preventing Class-Collapsing in Metric Learning With Margin-Based Losses (ICCV 2021 · similarity 0.89)
- Learning Towards The Largest Margins (ICLR 2022 · similarity 0.88)
- A Theoretically Sound Upper Bound on the Triplet Loss for Improving the Efficiency of Deep Distance Metric Learning (CVPR 2019 · similarity 0.87)
- Mean Field Theory in Deep Metric Learning (ICLR 2024 · similarity 0.86)
- Threshold-Consistent Margin Loss for Open-World Deep Metric Learning (ICLR 2024 · similarity 0.86)
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.
