Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Task Aligned Generative Meta-learning for Zero-shot Learning
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 |
|---|---|---|---|
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · AWA1 | average per-class Top-1 accuracy: 3.0% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · AWA2 | average per-class Top-1 accuracy: 2.1% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · CUB | average per-class Top-1 accuracy: 2.5% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · aPY | average per-class Top-1 accuracy: 7.6% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · GZSL | harmonic mean (H): 3.6% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · AWA1&aPY | average per-class Top-1 accuracy: 5.5% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · AWA2&aPY | average per-class Top-1 accuracy: 7.9% | Helped |
| Use task-wise distribution alignment with attribute-conditioned TAE and meta-generative learning | TGMZ · AWA&aPY | average per-class Top-1 accuracy: 3.6% | Helped |
| Set noise parameter sigma to 1 (ablation) | TGMZ-Softmax · ZSL | average per-class Top-1 accuracy | Helped |
| Increase the number of synthetic instances at test time (ablation) | TGMZ-Softmax · ZSL | average per-class Top-1 accuracy | Helped |
| Use 50 synthetic instances at test time (ablation) | TGMZ-Softmax on AWA2 · AWA2 | average per-class Top-1 accuracy | Hurt |
| Vary sigma and synthetic sample number (ablation) | TGMZ-SVM · four ZSL datasets | average per-class Top-1 accuracy | No clear effect |
Closest papers by meaning
- GenZSL: Generative Zero-Shot Learning Via Inductive Variational Autoencoder (ICML 2025 · similarity 0.91)
- Data Distribution Distilled Generative Model for Generalized Zero-Shot Recognition (AAAI 2024 · similarity 0.90)
- Data-Free Generalized Zero-Shot Learning (AAAI 2024 · similarity 0.90)
- ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning (ICLR 2025 · similarity 0.89)
- Improving Generalized Zero-Shot Learning by Exploring the Diverse Semantics from External Class Names (CVPR 2024 · similarity 0.89)
- Zero-Shot Task Adaptation with Relevant Feature Information (AAAI 2024 · similarity 0.88)
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.