An Intelligent Speech Interaction Model for Mobile Teaching

Intelligent answer system is a novel information service system which integrates and applies information retrieval, natural language processing, artificial intelligence and semantic analysis to respond users' questions in intelligence and accuracy. This paper focuses on Chinese questions answering matching task and puts forward intensive learning-based questions-answering matching method. First, we study words vector and the model from sequence to sequence, to design and generate questions answering-typed model. Then, we introduces neural network-based sequence mapping model, uses lots of questions answering-paired corpus training model and learns mapping relationship between questions sequence and answer sequence. After the model training is over, the model can generate corresponding answers of questions according to users' inputting questions. Finally, the influence of different comparison experiment tests on system performance and experiment data designs by us show that the model is feasible in practical application.