Command Palette
Search for a command to run...
Liang Song Jinlu Liu Yongqiang Qin
Abstract
Many Few-Shot Learning research works have two stages: pre-training base model and adapting to novel model. In this paper, we propose to use closed-form base learner, which constrains the adapting stage with pre-trained base model to get better generalized novel model. Following theoretical analysis proves its rationality as well as indication of how to train a well-generalized base model. We then conduct experiments on four benchmarks and achieve state-of-the-art performance in all cases. Notably, we achieve the accuracy of 87.75% on 5-shot miniImageNet which approximately outperforms existing methods by 10%.
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| few-shot-image-classification-on-cifar-fs-5 | ACC + Amphibian | Accuracy: 73.1 |
| few-shot-image-classification-on-cifar-fs-5-1 | ACC + Amphibian | Accuracy: 89.3 |
| few-shot-image-classification-on-fc100-5-way | ACC + Amphibian | Accuracy: 41.6 |
| few-shot-image-classification-on-fc100-5-way-1 | ACC + Amphibian | Accuracy: 66.9 |
| few-shot-image-classification-on-mini-2 | ACC + Amphibian | Accuracy: 62.21 |
| few-shot-image-classification-on-mini-3 | ACC + Amphibian | Accuracy: 80.75 |
| few-shot-image-classification-on-tiered | ACC + Amphibian | Accuracy: 68.77 |
| few-shot-image-classification-on-tiered-1 | ACC + Amphibian | Accuracy: 86.75 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.