POETRY AUTOMATIC GENERATION SYSTEM BASED ON NATURAL LANGUAGE PROCESSING

Abstract

In order to increase comprehension on abstract theory and decrease the fear of writing intelligent algorithms, engineering training based on poetry automatic generation system of natural speech recognition technology is proposed. Firstly, the background of artificial intelligence(AI) is introduced, then the comprehensive experimental teaching was separated into the simple operation training model designing and exploring teaching model, due to the characteristics and problems of complex artificial intelli-gence engineering simulation education for undergraduates. Finally, Combining the advantages of some algorithms, Displaying the automatically generated poetry to collect samples, which can enable students to learn the application of intelligent algorithm intuitively. The test on the training platform shows that the automatic generation of five characters or seven characters poems has a high probability of poetic features, and the strong correlation is between the poems generated by LSTM model and keywords.