Class Representation Networks for Few-Shot Learning

摘要

In this paper, we proposed a novel network referred as Class Representation Networks (CRNs) to solve the few-shot learning problems. In the proposed CRNs, a high-quality class representation is learned by training a set-based neural network. In addition, a network with fully connected layers was constructed for learning distance metric instead of using a predefined distance metric. Compared with recent methods for few-shot learning, our network achieves state-of-the-art performance for few-shot learning. Extensive experiments on three benchmark datasets validate the effectiveness of our proposed model.