Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Boosting Few-Shot Text Classification via Distribution Estimation
Read the paperCheck 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 |
|---|---|---|---|
| Estimate novel-class distributions from support samples and nearest unlabeled queries, then augment with generated samples | Way-DE and Shot-DE · News/review datasets average, 5-way 1-shot | accuracy: 71.2 -> 72.4 | Helped |
| Estimate novel-class distributions from support samples and nearest unlabeled queries, then augment with generated samples | Way-DE · News/review datasets average, 5-way 5-shot | accuracy: 81.6 -> 84.4 | Helped |
| Estimate novel-class distributions from support samples and nearest unlabeled queries, then augment with generated samples | Shot-DE · News/review datasets average, 5-way 5-shot | accuracy: 81.6 -> 84.2 | Helped |
| Estimate distributions using query samples instead of transferring statistics from seen classes | Way-DE and Shot-DE · Intent detection datasets average, 5-way 1-shot | accuracy: 90.3 -> 91.5 | Helped |
| Estimate distributions using query samples instead of transferring statistics from seen classes | Way-DE and Shot-DE · Intent detection datasets average, 5-way 5-shot | accuracy: 95.2 -> 95.9 | Helped |
| Combine statistics from seen classes with query data rather than using seen-class statistics alone | DC-DE relative to DC · Intent detection datasets average, 5-way 1-shot | accuracy: 86.6 -> 90.3 | Helped |
| Combine statistics from seen classes with query data rather than using seen-class statistics alone | DC-DE relative to DC · Intent detection datasets average, 5-way 5-shot | accuracy: 94.3 -> 95.2 | Helped |
| Use way-based rather than shot-based distribution estimation | Way-DE versus Shot-DE · News/review datasets average, 5-way 1-shot | accuracy: 72.4 -> 72.4 | No clear effect |
| Use way-based rather than shot-based distribution estimation | Way-DE versus Shot-DE · News/review datasets average, 5-way 5-shot | accuracy: 84.2 -> 84.4 | No clear effect |
| Use way-based rather than shot-based distribution estimation | Way-DE versus Shot-DE · Intent detection datasets average, 5-way 1-shot | accuracy: 91.5 -> 91.5 | No clear effect |
| Use way-based rather than shot-based distribution estimation | Way-DE versus Shot-DE · Intent detection datasets average, 5-way 5-shot | accuracy: 95.9 -> 95.9 | No clear effect |
Closest papers by meaning
- Liberating Seen Classes: Boosting Few-Shot and Zero-Shot Text Classification via Anchor Generation and Classification Reframing (AAAI 2024 · similarity 0.87)
- Multi-Label Few-Shot Image Classification via Pairwise Feature Augmentation and Flexible Prompt Learning (AAAI 2025 · similarity 0.85)
- Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learning (CVPR 2024 · similarity 0.85)
- Transductive Zero-Shot and Few-Shot CLIP (CVPR 2024 · similarity 0.85)
- Text Augmented Correlation Transformer For Few-shot Classification & Segmentation (CVPR 2025 · similarity 0.85)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 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.