Two-stage Training for Chinese Dialect Recognition
This is an incremental improvement for Chinese dialect recognition systems, potentially benefiting speech technology applications in multilingual regions.
The authors tackled Chinese dialect recognition by developing a two-stage system combining ResNet14 and RNN, which won first place in a competition among 110 teams. Their approach achieved high accuracy with less training time compared to a three-stage alternative.
In this paper, we present a two-stage language identification (LID) system based on a shallow ResNet14 followed by a simple 2-layer recurrent neural network (RNN) architecture, which was used for Xunfei (iFlyTek) Chinese Dialect Recognition Challenge and won the first place among 110 teams. The system trains an acoustic model (AM) firstly with connectionist temporal classification (CTC) to recognize the given phonetic sequence annotation and then train another RNN to classify dialect category by utilizing the intermediate features as inputs from the AM. Compared with a three-stage system we further explore, our results show that the two-stage system can achieve high accuracy for Chinese dialects recognition under both short utterance and long utterance conditions with less training time.