Research map / Label-efficient and robust learning / Recognition and long-tail
Unsupervised Hyperbolic 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 |
|---|---|---|---|
| use Unsupervised Hyperbolic Metric Learning with Hierarchical Similarity instead of TAC-CCL | GoogLeNet unsupervised metric learning · CARS196 | Recall@1: 46.1 -> 47.7 | Helped |
| use Unsupervised Hyperbolic Metric Learning with Hierarchical Similarity instead of TAC-CCL | GoogLeNet unsupervised metric learning · CUB-200-2011 | Recall@1: 57.5 -> 58.9 | Helped |
| use Unsupervised Hyperbolic Metric Learning with Hierarchical Similarity instead of TAC-CCL | GoogLeNet unsupervised metric learning · SOP | Recall@1: 63.9 -> 65.1 | Helped |
| add a hyperbolic layer (ablation) | Contrastive Loss unsupervised metric learning · CARS196 | Recall@1: 33.3 -> 43.7 | Helped |
| add a hyperbolic layer (ablation) | Contrastive Loss unsupervised metric learning · CUB-200-2011 | Recall@1: 48.8 -> 54.5 | Helped |
| add a hyperbolic layer (ablation) | Triplet Loss unsupervised metric learning · CARS196 | Recall@1: 37.9 -> 44.7 | Helped |
| add a hyperbolic layer (ablation) | Triplet Loss unsupervised metric learning · CUB-200-2011 | Recall@1: 50.7 -> 55.1 | Helped |
| add a hyperbolic layer (ablation) | LiftedStructure Loss unsupervised metric learning · CARS196 | Recall@1: 42.4 -> 44.3 | Helped |
| add a hyperbolic layer (ablation) | LiftedStructure Loss unsupervised metric learning · CUB-200-2011 | Recall@1: 53.1 -> 55.1 | Helped |
| add hierarchical clustering and the log-ratio loss with hierarchical similarity beyond the hyperbolic layer (ablation) | GoogLeNet unsupervised metric learning · CARS196 | Recall@1: 44.7 -> 47.7 | Helped |
| add hierarchical clustering and the log-ratio loss with hierarchical similarity beyond the hyperbolic layer (ablation) | GoogLeNet unsupervised metric learning · CUB-200-2011 | Recall@1: 55.1 -> 58.9 | Helped |
| increase the number of hierarchical similarity levels (ablation) | Unsupervised Hyperbolic Metric Learning with Hierarchical Similarity · CUB-200-2011 | Recall@1 | Mixed |
Closest papers by meaning
- Hyperbolic Vision Transformers: Combining Improvements in Metric Learning (CVPR 2022 · similarity 0.91)
- Unsupervised Feature Learning with Emergent Data-Driven Prototypicality (CVPR 2024 · similarity 0.88)
- Hyperbolic Image Embeddings (CVPR 2020 · similarity 0.88)
- Robust Hyperbolic Learning with Curvature-Aware Optimization (NeurIPS 2025 · similarity 0.86)
- HiHPQ: Hierarchical Hyperbolic Product Quantization for Unsupervised Image Retrieval (AAAI 2024 · similarity 0.86)
- Hyperbolic Embeddings of Supervised Models (NeurIPS 2024 · 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.