Design and Research of an Assisted Vocal Training System Based on Bi-LSTM and Softmax
To further improve the auxiliary effect of English oral pronunciation training,a bidirectional multi round information sharing network IN-SN based oral comprehension model is proposed to assist learners in English oral pronunciation training.Among them,the semantic information extraction method based on Bi LSTM combines bidirectional multi round information sharing network and attention mechanism to achieve better oral comprehension effects.The experimental results show that compared with other seman-tic extraction models,the designed IN-SN based oral comprehension model has better semantic extraction ability,achieving accuracy and F1 scores of 96.06％,95.98％,97.47％,and 94.03％on the dataset ATIS and dataset SNPIS,respectively;The designed Eng-lish speaking auxiliary training system can perform accurate intention recognition and semantic slot filling.In summary,the designed English speaking auxiliary training system has good performance and can be applied to practical English speaking vocalization learning scenarios,with high feasibility.
万方数据
维普
