rethinking few shot image classification a good embedding is all you need

• DOI: 10.48550/arxiv.2003.11539
• Publication Date: 2020-01-01
• Authors: Joshua B. Tenenbaum; Yonglong Tian; Phillip Isola; Dilip Krishnan; Yue Wang
• Keywords: FOS: Computer and information sciences; Computer Science - Machine Learning; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition; 0202 electrical engineering, electronic engineering, information engineering; 006; 02 engineering and technology; Machine Learning (cs.LG)

The focus of recent meta-learning research has been on the development of learning algorithms that can quickly adapt to test time tasks with limited data and low computational cost. Few-shot learning is widely used as one of the standard benchmarks in meta-learning. In this work, we show that a simple baseline: learning a supervised or self-supervised representation on the meta-training set, followed by training a linear classifier on top of this representation, outperforms state-of-the-art few-shot learning methods. An additional boost can be achieved through the use of self-distillation. This demonstrates that using a good learned embedding model can be more effective than sophisticated meta-learning algorithms. We believe that our findings motivate a rethinking of few-shot image classification benchmarks and the associated role of meta-learning algorithms. Code is available at: http://github.com/WangYueFt/rfs/.<br/>First two authors contributed equally. Project Page: https://people.csail.mit.edu/yuewang/projects/rfs/ Code: http://github.com/WangYueFt/rfs/<br/>