Research map / Label-efficient and robust learning / Recognition and long-tail
Orthogonal Projection Loss
Kanchana Ranasinghe, Muzammal Naseer, Munawar Hayat, Salman Khan, Fahad Shahbaz Khan
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 OPL to CE | ResNet-56 · CIFAR-100 | top-1 accuracy: 72.40% -> 73.52% | Improved |
| add OPL to CE | ResNet-110 · CIFAR-100 | top-1 accuracy: 73.79% -> 74.85% | Improved |
| add OPL to CE | ResNet-50 · ImageNet | top-1 accuracy: 76.15% -> 76.98% | Improved |
| add OPL to CE | ResNet-18 · ImageNet | top-1 accuracy: 69.91% -> 70.27% | Improved |
| add OPL to Truncated Loss | ResNet18 · CIFAR100 class-dependent noisy labels | accuracy: 47.66% -> 53.94% | Improved |
| add OPL to Madry adversarial training | Madry et al. · CIFAR10 AutoAttack | robust accuracy: 44.04 -> 49.15 | Improved |
| add OPL to MART | MART · CIFAR10 | clean accuracy: 84.49 -> 84.41 | No clear effect |
| add OPL to MART | MART · CIFAR100 | clean accuracy: 58.90 -> 58.01 | Hurt |
| add OPL to RSC | RSC · PACS | average accuracy: 87.47 -> 88.48 | Improved |
| add OPL to RFS | RFS · miniImageNet 1-shot | accuracy: 62.02 -> 63.10 | Improved |
| add OPL to SUR | SUR · Meta-Dataset average | accuracy: 71.4 -> 71.9 | Improved |
| add OPL to RFS under Gaussian input noise sigma=0.1 (ablation) | RFS · miniImageNet 5-shot (noisy) | accuracy: 74.46 -> 73.20 | Hurt |
Closest papers by meaning
- OLÉ: Orthogonal Low-Rank Embedding - A Plug and Play Geometric Loss for Deep Learning (CVPR 2018 · similarity 0.90)
- CASE: Exploiting Intra-class Compactness and Inter-class Separability of Feature Embeddings for Out-of-Distribution Detection (AAAI 2024 · similarity 0.87)
- Engineering the Neural Collapse Geometry of Supervised-Contrastive Loss (Student Abstract) (AAAI 2024 · similarity 0.85)
- Neural Collapse for Cross-entropy Class-Imbalanced Learning with Unconstrained ReLU Features Model (ICML 2024 · similarity 0.85)
- Preventing Dimensional Collapse in Self-Supervised Learning via Orthogonality Regularization (NeurIPS 2024 · similarity 0.84)
- Unlocking Better Closed-Set Alignment Based on Neural Collapse for Open-Set Recognition (AAAI 2025 · similarity 0.84)
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.
