A Robot Dialogue System for Preschool Education Based on AI Artificial Intelligence

In order to improve the dialogue system of preschool education robots and achieve more accurate understanding and execution of text instructions in preschool children’s speech, the author proposes a children’s speech enhancement method based on a two-stage attention mechanism. This method is based on a two-stage attention mechanism and has undergone a series of steps and experimental validation, achieving significant optimization results. Firstly, a detailed analysis was conducted on mainstream speech enhancement methods, and the problems of single channel time-domain speech enhancement methods were identified. Subsequently, a two-stage attention enhancement network for speech enhancement was established. This method allows the machine to focus more on processing input speech data to improve speech quality and clarity. The experimental results show that the two-stage attention speech enhancement method improves the SSNR score by 3.2dB compared to the channel attention enhancement network with higher scores, and its loss value is the lowest among the four models, only 0.87. In addition, the enhanced speech waveform is also clearer, making it easier for the machine to understand and execute instructions in children’s speech