A study to understand the language understanding so far it often takes a difficult idea of the text. Most datasets are found in the text, newly mixed into a talk, and many models often rely on default text. This is harmful to Prosody – additional details of the talk signal across the sounds for themselves and it is difficult to recover from the text alone. In this work, we investigate the Posy role in the spoken question responds. By separating the Prosodic and Lexical information in the Inlu-SQA Data 5, which contains the natural language, we show that the Models are trained in the Prosodic Details only effective Prosodic. However, we find that when the Lexical information is available, models are tend to reduce it. Our acquisition suggests that while the Prosodic CAs provide important information on the import, efficient and necessary methods are required to ensure that the Proses contributed significantly in accordance with Lexical features.
