Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
POODLE: Improving Few-shot Learning via Penalizing Out-of-Distribution Samples
Duong Hoang Le, Khoi Duc Nguyen, Khoi Nguyen, Quoc-Huy Tran, Rang Nguyen, Binh-Son Hua
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 POODLE-R to Rot + KD | ResNet-12 · mini-ImageNet 1-shot, inductive | accuracy: 65.91 -> 67.80 | Improved |
| Add POODLE-B to Rot + KD | ResNet-12 · mini-ImageNet 1-shot, transductive | accuracy: 65.91 -> 77.56 | Improved |
| Add POODLE-R to Rot + KD | ResNet-12 · tiered-ImageNet 1-shot, inductive | accuracy: 69.43 -> 70.47 | Mixed |
| Add POODLE-B to Rot + KD | ResNet-12 · CUB 1-shot, transductive | accuracy: 79.70 -> 89.88 | Improved |
| Use target-domain noisy negatives with POODLE-B | Rot + KD ResNet-12 · iNat 1-shot, cross-domain | accuracy: 48.90 -> 50.62 | Improved |
| Use target-domain noisy negatives with POODLE-B | Rot + KD ResNet-12 · ISIC 1-shot, cross-domain | accuracy: 32.35 -> 33.56 | Improved |
| Use uniform random features instead of base-class negatives (ablation) | POODLE, mini-ImageNet 1-shot · in-domain inductive FSL | accuracy: $66.32\pm0.20$ -> $68.20\pm0.20$ | Improved |
| Add the push term using base-class negatives (ablation) | CE + L_pull, Simple ResNet-12 · mini-ImageNet 1-shot, transductive | accuracy: 70.91 -> 74.21 | Improved |
| Replace base-class negatives with uniform random negatives (ablation) | CE + L_pull, Simple ResNet-12 · mini-ImageNet 1-shot, transductive | accuracy: 70.91 -> 74.82 | Improved |
| Remove the pull term and use only the push term (ablation) | Rot + KD ResNet-12 · mini-ImageNet 1-shot, inductive | accuracy: $66.32\pm0.20$ -> $66.32\pm0.20$ | No clear effect |
| Fine-tune a (k+1)-way classifier with OOD as an extra class (ablation) | Rot + KD ResNet-12 · mini-ImageNet 1-shot, inductive | accuracy: $66.32\pm0.20$ -> $66.36\pm0.20$ | No clear effect |
| Apply label smoothing (ablation) | Rot + KD ResNet-12 · mini-ImageNet 5-shot, inductive | accuracy: $82.99\pm0.13$ -> $80.81\pm0.13$ | Hurt |
Closest papers by meaning
- OOD-MAML: Meta-Learning for Few-Shot Out-of-Distribution Detection and Classification (NeurIPS 2020 · similarity 0.87)
- Prototypical Networks for Few-shot Learning (NeurIPS 2017 · similarity 0.86)
- PTN: A Poisson Transfer Network for Semi-supervised Few-shot Learning (AAAI 2021 · similarity 0.86)
- FSL-Rectifier: Rectify Outliers in Few-Shot Learning via Test-Time Augmentation (AAAI 2025 · similarity 0.85)
- Adaptive Multi-prompt Contrastive Network for Few-shot Out-of-distribution Detection (ICML 2025 · similarity 0.85)
- Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learning (CVPR 2024 · 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.
