Research map / Label-efficient and robust learning / Recognition and long-tail
Hier-COS: Making Deep Features Hierarchy-aware via Composition of Orthogonal Subspaces
Depanshu Sani, Saket Anand
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 |
|---|---|---|---|
| use Hier-COS instead of HAFrame | ResNet-50 · FGVC-Aircraft | top-1 accuracy: 80.55 -> 81.75 | Improved |
| use Hier-COS instead of HAFrame | ResNet-50 · CIFAR-100 | top-1 accuracy: 77.53 -> 77.79 | No clear effect |
| use Hier-COS instead of HAFrame | ResNet-50 · iNaturalist-19 | HOPS: 0.89 -> 0.96 | Improved |
| use Hier-COS instead of HAFrame | ResNet-50 · FGVC-Aircraft | HOPS: 0.86 -> 0.89 | Improved |
| use Hier-COS instead of cross-entropy with ViT backbone | ViT · iNaturalist-19 | top-1 accuracy: 78.39 -> 80.81 | Improved |
| use Hier-COS instead of cross-entropy with ViT backbone | ViT · iNaturalist-19 | HOPS: 0.53 -> 0.98 | Improved |
| use Hier-COS instead of HAFrame | ResNet-50 · FGVC-Aircraft | mistake severity: 2.00 -> 2.09 | Hurt |
| use Hier-COS instead of Flamingo-I | ResNet-50 · FGVC-Aircraft | FPA: 77.92 -> 81.56 | Improved |
| use Hier-COS instead of HAFeat-I | ResNet-50 · CIFAR-100 | FPA: 75.52 -> 77.37 | Improved |
| use Hier-COS instead of Flamingo-II | ResNet-50 · iNaturalist-19 | FPA: 68.75 -> 69.93 | Improved |
| use Hier-COS instead of Cross Entropy (super-class probabilities summed) | ResNet-50 · CIFAR-100 | FPA: 77.11 -> 77.37 | Improved |
Closest papers by meaning
- Self-Paced Unified Representation Learning for Hierarchical Multi-Label Classification (AAAI 2024 · similarity 0.87)
- Deep Taxonomic Networks for Unsupervised Hierarchical Prototype Discovery (NeurIPS 2025 · similarity 0.84)
- Learning Structured Representations with Hyperbolic Embeddings (NeurIPS 2024 · similarity 0.84)
- Hierarchical Selective Classification (NeurIPS 2024 · similarity 0.84)
- From Logits to Hierarchies: Hierarchical Clustering made Simple (ICML 2025 · similarity 0.84)
- HiGDA: Hierarchical Graph of Nodes to Learn Local-to-Global Topology for Semi-Supervised Domain Adaptation (AAAI 2025 · similarity 0.83)
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.
