Research On Dialogable Video Tree Based On Natural Language Understanding

With the rapid development of artificial intelligence and computer technology,intelligence has been integrated into all aspects of human life.Using natural language to communicate with computers has been the pursuit of human beings.The improvement of natural language understanding accuracy has laid a solid foundation for the realization of intelligent life.In recent years,interactive video has received widespread attention in the field of new media.The concept of interactive video breaks the barriers between video producers and viewers,and changes the way traditional video viewers passively receive content.With the development of deep learning models,Chinese natural language processing models make the machine's semantic understanding of human natural language closer and closer to the human level.However,natural language has ambiguity,and the problem of ambiguity in the understanding of natural language semantics by machines makes it impossible to accurately analyze human needs.At the same time,interactive videos are mainly produced for storytelling videos,Due to the professionalism of the producers,interactive tutorial videos have the problem of low content quality.This article conducts a series of researches on the content quality and ambiguity of tutorial interactive videos,and discusses the importance of tutorial interactive videos in the process of learning domain expertise.A dialogueable video tree generation model based on natural language understanding in the field of interactive video tutorials is proposed,and combined with this model,the automatic generation of question-and-answer pairs of video text content is realized.The question-and-answer pair generation model includes a question generation model,a question screening model,an answer generation model,an answer screening model,and an answer recursive model.Among them,the model uses the deep learning natural language processing pre-training model Bert to automatically generate questions.The question types are mainly professional questions in the field,and the answer types are composed of text,pictures and videos.At the same time,in order to solve the ambiguity of adverbial omission in the process of automatically generating the question,a word distance formula is proposed to effectively recover the omission component in the question sentence.Define the concept of domain coverage,and propose a domain coverage formula to evaluate the quality of the produced video.The conversational video tree model based on natural language understanding can help video producers discover potential effective content in videos to solve problems and improve the professionalism and quality of video content.Combining deep learning models in the field of natural language processing and related knowledge of natural language understanding to help the production process of tutorial interactive videos,so that video producers can accurately understand the needs of viewers,explore the potential value of video content,and improve the quality of video content.At the same time,expand the coverage of video content,making tutorial videos easier to learn and understand.This not only lowers the entry threshold for professional knowledge learning,but also promotes the development of the field of interactive video production.