Few-shot Image Classification: Just Use a Library of Pre-trained Feature Extractors and a Simple Classifier

Recent papers have suggested that transfer learning can outperform\nsophisticated meta-learning methods for few-shot image classification. We take\nthis hypothesis to its logical conclusion, and suggest the use of an ensemble\nof high-quality, pre-trained feature extractors for few-shot image\nclassification. We show experimentally that a library of pre-trained feature\nextractors combined with a simple feed-forward network learned with an\nL2-regularizer can be an excellent option for solving cross-domain few-shot\nimage classification. Our experimental results suggest that this simpler\nsample-efficient approach far outperforms several well-established\nmeta-learning algorithms on a variety of few-shot tasks.\n