Research map / Label-efficient and robust learning / Noisy labels
Few-Shot Learning With Noisy Labels
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 TraNFS-3 instead of Vanilla ProtoNet | Few-shot classification · MiniImageNet, 5-way 5-shot, symmetric noise 40% | accuracy: 51.41 ± 0.19 -> 56.65 ± 0.21 | Helped |
| Use TraNFS-3 instead of RNNP | Few-shot classification · MiniImageNet, 5-way 5-shot, paired noise 40% | accuracy: 47.88 ± 0.19 -> 53.96 ± 0.23 | Helped |
| Replace mean prototypes with median aggregation | ProtoNet · MiniImageNet, 5-way 5-shot, symmetric noise 40% | accuracy: 51.41 ± 0.19 -> 51.86 ± 0.19 | Helped |
| Train with varied noise levels {0%, 20%, 40%} rather than 0% only (ablation) | TraNFS-3 · MiniImageNet, 5-way 5-shot, symmetric noise 40%, meta-test | accuracy: 52.85 ± 0.19 -> 56.73 ± 0.21 | Helped |
| Train with 60% noise only instead of varied 0%, 20%, and 40% noise (ablation) | TraNFS-3 · MiniImageNet, 5-way 5-shot, symmetric noise 40%, meta-test | accuracy: 56.73 ± 0.21 -> 43.11 ± 0.21 | Hurt |
| Expose TraNFS to noisy support sets during meta-training (ablation) | TraNFS · Noisy few-shot support sets | noise rejection | Helped |
Closest papers by meaning
- FSL-Rectifier: Rectify Outliers in Few-Shot Learning via Test-Time Augmentation (AAAI 2025 · similarity 0.89)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.88)
- MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlation (ICLR 2024 · similarity 0.87)
- Enhanced Sample Selection with Confidence Tracking: Identifying Correctly Labeled Yet Hard-to-Learn Samples in Noisy Data (AAAI 2025 · similarity 0.86)
- BECLR: Batch Enhanced Contrastive Few-Shot Learning (ICLR 2024 · similarity 0.86)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.86)
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.