Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Contrastive Prototypical Network with Wasserstein Confidence Penalty
Haoqing Wang, Zhi-Hong Deng
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 pairwise contrast (ablation) | CPN (Conv4) · miniImageNet 5-way 1-shot | accuracy (%): 46.08 -> 46.96 | Improved |
| add pairwise contrast (ablation) | CPN (Conv4) · miniImageNet 5-way 50-shot | accuracy (%): 74.81 -> 75.63 | Improved |
| add pairwise contrast (ablation) | CPN (ResNet12) · miniImageNet 5-way 1-shot | accuracy (%): 48.80 -> 50.01 | Improved |
| add pairwise contrast (ablation) | CPN (ResNet12) · miniImageNet 5-way 50-shot | accuracy (%): 80.83 -> 82.74 | Improved |
| add Wasserstein Confidence Penalty (ablation) | CPN (Conv4) · miniImageNet 5-way 1-shot | accuracy (%): 46.96 -> 47.93 | Improved |
| add Wasserstein Confidence Penalty (ablation) | CPN (Conv4) · miniImageNet 5-way 50-shot | accuracy (%): 75.63 -> 78.20 | Improved |
| add Wasserstein Confidence Penalty (ablation) | CPN (ResNet12) · miniImageNet 5-way 1-shot | accuracy (%): 50.01 -> 53.56 | Improved |
| add Jensen-Shannon Confidence Penalty (ablation) | CPN (Conv4) · miniImageNet 5-way 1-shot | accuracy (%): 46.96 -> 46.82 | No clear effect |
| add Jensen-Shannon Confidence Penalty (ablation) | CPN (ResNet12) · miniImageNet 5-way 1-shot | accuracy (%): 50.01 -> 49.87 | Hurt |
| add Consistency Regularization (ablation) | CPN (Conv4) · miniImageNet 5-way 50-shot | accuracy (%): 75.63 -> 75.50 | Hurt |
| use Wasserstein penalty instead of KL Confidence Penalty (ablation) | CPN (Conv4) · miniImageNet 5-way 1-shot | accuracy (%): 47.22 -> 47.93 | Improved |
| use Wasserstein penalty instead of Label Smoothing (ablation) | CPN (ResNet12) · miniImageNet 5-way 1-shot | accuracy (%): 50.41 -> 53.56 | Improved |
Closest papers by meaning
- BECLR: Batch Enhanced Contrastive Few-Shot Learning (ICLR 2024 · similarity 0.88)
- Prototype Rectification for Few-Shot Learning (ECCV 2020 · similarity 0.87)
- Prototypical Networks for Few-shot Learning (NeurIPS 2017 · similarity 0.87)
- Self-Supervised Contrastive Learning is Approximately Supervised Contrastive Learning (NeurIPS 2025 · similarity 0.85)
- Simplicity Bias and Optimization Threshold in Two-Layer ReLU Networks (ICML 2025 · similarity 0.85)
- On the Similarities of Embeddings in Contrastive Learning (ICML 2025 · 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.
