Research map / Label-efficient and robust learning / Self-supervised and contrastive
ESPT: A Self-Supervised Episodic Spatial Pretext Task for Improving Few-Shot Learning
Yi Rong, Xiongbo Lu, Zhaoyang Sun, Yaxiong Chen, Shengwu Xiong
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 the ESPT spatial-consistency objective (ablation) | ESPT · miniImageNet 5-way 1-shot | accuracy: +1.63$ \% $ | Improved |
| Add the ESPT spatial-consistency objective (ablation) | ESPT · miniImageNet 5-way 5-shot | accuracy: +0.90$ \% $ | Improved |
| Add the ESPT spatial-consistency objective (ablation) | ESPT · CUB-200-2011 5-way 1-shot | accuracy: +1.29$ \% $ | Improved |
| Add the ESPT spatial-consistency objective (ablation) | ESPT · CUB-200-2011 5-way 5-shot | accuracy: +0.76$ \% $ | Improved |
| Use multiple rotation transformations instead of a single rotation (ablation) | ESPT · miniImageNet and CUB-200-2011 5-way few-shot | accuracy | Improved |
| Use U={180°,270°} (ablation) | ESPT · miniImageNet 5-way 1-shot and 5-shot | accuracy: 68.36; 84.11 | Improved |
| Use U={90°,270°} (ablation) | ESPT · CUB-200-2011 5-way 1-shot and 5-shot | accuracy: 85.45; 94.02 | Improved |
| Use all three rotations instead of the best two-rotation set (ablation) | ESPT · miniImageNet 5-way 5-shot | accuracy: 84.11 -> 84.07 | Hurt |
| Use all three rotations instead of U={90°,270°} (ablation) | ESPT · CUB-200-2011 5-way 1-shot | accuracy: 85.45 -> 85.41 | Hurt |
Closest papers by meaning
- IEPT: Instance-Level and Episode-Level Pretext Tasks for Few-Shot Learning (ICLR 2021 · similarity 0.91)
- Boosting Few-Shot Visual Learning With Self-Supervision (ICCV 2019 · similarity 0.87)
- Self-training For Few-shot Transfer Across Extreme Task Differences (ICLR 2021 · similarity 0.86)
- Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference (CVPR 2022 · similarity 0.86)
- TAFSSL: Task-Adaptive Feature Sub-Space Learning for few-shot classification (ECCV 2020 · similarity 0.86)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (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.
