Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Zero-Shot Task Transfer
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 |
|---|---|---|---|
| Increase the number of known source tasks (ablation) | TTNet · Taskonomy zero-shot tasks | Win rate | Helped |
| Use 20 known tasks instead of 6 (ablation) | TTNet · Taskonomy surface normal | Win rate vs Wang: 84 -> 91 | Helped |
| Use 20 known tasks instead of 6 (ablation) | TTNet · Taskonomy surface normal | Win rate vs Zamir: 75 -> 87 | Helped |
| Fine-tune the regressed model on 20% target-task data (ablation) | TTNet with 20 known tasks · Taskonomy surface normal | Mean angular error: 19.27 -> 15.10 | Helped |
| Vary the number of crowd annotators used to construct task correlations (ablation) | TTNet_6 · Taskonomy camera pose | Win rate vs RANSAC: 28% (3 annotators); 51% (10); 90% (20); 88% (30); 88% (35); 90% (40); 87% (45); 90% (50) | Mixed |
| Vary the choice of source tasks (ablation) | TTNet · Taskonomy zero-shot tasks | Win rate | No clear effect |
| Use task correlations from the Taskonomy graph instead of crowd votes (ablation) | TTNet_6 · Taskonomy zero-shot tasks | Qualitative task performance | No clear effect |
Closest papers by meaning
- Zero-Shot Task Adaptation with Relevant Feature Information (AAAI 2024 · similarity 0.89)
- Meta-Transfer Learning for Few-Shot Learning (CVPR 2019 · similarity 0.88)
- Task Aligned Generative Meta-learning for Zero-shot Learning (AAAI 2021 · similarity 0.87)
- A Generalization Theory for Zero-Shot Prediction (ICML 2025 · similarity 0.85)
- Explanatory Instructions: Towards Unified Vision Tasks Understanding and Zero-shot Generalization (ICML 2025 · similarity 0.85)
- Task-Adaptive Prompted Transformer for Cross-Domain Few-Shot Learning (AAAI 2024 · 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.