关键词: 深度学习, 小样本, 改进匹配网络, 平方欧氏距离, LSTM

Abstract: The current deep learning is based on a large number of labeled data samples to automatically identify the model through a multilayer network. However, in many special scenarios, it is difficult to obtain a large amount of sample data, and the identification of fewshot learning is still a key problem in deep learning. To solve this problem, the fourlayer deep convolutional neural network (DCNN) is first used to extract the highlevel semantic features of the training samples and the test samples. Then use the bidirectional LSTM and attLSTM algorithms for further extraction and code of more critical and useful features of training samples and test samples based on the improved matching network. Finally,  the softmax nonlinear classifier is used to classify the test samples on the squared euclidean distance. The experiment tests on the proposed improved model with the Omniglot data set and achieves very good results. The improved model can achieve a 93.2% recognition rate even in the most complicated 20way 1shot case, and the original matching network model of Vinyals only achieve 88.2% recognition in the case of 20way 1shot. Compared with the original matching network model, the improved model has a better recognition effect in a complex scenario with more categories and fewer samples.

Key words: deep learning, fewshot, improved matching network, squared Euclidean distance, LSTM