Research map / Label-efficient and robust learning / Recognition and long-tail
The Group Loss++: A Deeper Look Into Group Loss for Deep Metric Learning
Ismail Elezi, Jenny Seidenschwarz, Laurin Wagner, Sebastiano Vascon, Alessandro Torcinovich, Marcello Pelillo, Laura Leal-Taixé
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 mixed pooling instead of average pooling (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 67.5 | Improved |
| add beta-normalization (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 66.8 | Improved |
| replace ReLU with LeakyReLU (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 61.3 | Hurt |
| replace ReLU with LeakyReLU (ablation) | Group Loss · CARS196 | Recall@1: 85.6 -> 85.6 | No clear effect |
| add re-ranking (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 68.3 | Improved |
| add flip inference (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 68.0 | Improved |
| add all inference strategies cumulatively (Group Loss++) (ablation) | Group Loss · CUB-200-2011 | Recall@1: 65.5 -> 72.6 | Improved |
| add all inference strategies cumulatively (Group Loss++) (ablation) | Group Loss · CARS196 | Recall@1: 85.6 -> 90.4 | Improved |
| add all inference strategies cumulatively (Group Loss++) (ablation) | Group Loss · In-Shop Clothes | Recall@1: 86.8 -> 90.9 | Improved |
| add flip inference to mixed + beta + L + R (ablation) | Group Loss++ · CUB-200-2011 | Recall@1: 70.2 -> 72.6 | Improved |
| add refinement steps (one step) over cross-entropy only (ablation) | Group Loss · CUB-200-2011 | Recall@1: +0.6pp | Improved |
| increase refinement steps to five (ablation) | Group Loss · CARS196 | Recall@1: best at three or four steps | Hurt |
Closest papers by meaning
- Multi-Similarity Loss With General Pair Weighting for Deep Metric Learning (CVPR 2019 · similarity 0.89)
- Embedding Expansion: Augmentation in Embedding Space for Deep Metric Learning (CVPR 2020 · similarity 0.88)
- The Group Loss for Deep Metric Learning (ECCV 2020 · similarity 0.88)
- Improving Deep Metric Learning by Divide and Conquer (TPAMI 2021 · similarity 0.88)
- Rethinking Preventing Class-Collapsing in Metric Learning With Margin-Based Losses (ICCV 2021 · similarity 0.88)
- Ranked List Loss for Deep Metric Learning (CVPR 2019 · similarity 0.88)
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.
