To rescue and preserve an endangered language, this paper studied an end-to-end speech\nrecognition model based on sample transfer learning for the low-resource Tujia language. From the\nperspective of the Tujia language international phonetic alphabet (IPA) label layer, using Chinese\ncorpus as an extension of the Tujia language can effectively solve the problem of an insufficient\ncorpus in the Tujia language, constructing a cross-language corpus and an IPA dictionary that is\nunified between the Chinese and Tujia languages. The convolutional neural network (CNN) and\nbi-directional long short-term memory (BiLSTM) network were used to extract the cross-language\nacoustic features and train shared hidden layer weights for the Tujia language and Chinese phonetic\ncorpus. In addition, the automatic speech recognition function of the Tujia language was realized\nusing the end-to-end method that consists of symmetric encoding and decoding. Furthermore,\ntransfer learning was used to establish the model of the cross-language end-to-end Tujia language\nrecognition system. The experimental results showed that the recognition error rate of the proposed\nmodel is 46.19%, which is 2.11% lower than the that of the model that only used the Tujia language\ndata for training. Therefore, this approach is feasible and effective.
Loading....