The Application Of Artificial Intelligence In Language Training Guidance And Correction

In recent years,"Internet + education" has become a development trend,and artificial intelligence has risen to the height of national strategy."Artificial intelligence + education" is also facing new challenges.College English language skills training plays an important role in student learning and future development.Machine learning,deep learning and other technologies in artificial intelligence are used in education,especially in language training guidance and correction,which can solve many pain points in the process of instructional correction.Beijing University of Posts and Telecommunications undergraduate students will use the College English Language Skills Training System every semester for pretest,midtest and posttest.The objective questions can directly give answers.The subjective questions can only be corrected manually by the teachers.Efforts have raised higher demands.Based on the data of College English Skills Training System,the paper establishes a smart scoring model for spoken language expressions and implements intelligent marking of spoken language expressions.This article first analyzes the research background,research significance and research status at home and abroad,and elaborates the concepts related to artificial intelligence and language training.Then the audio data is recognized as text,then the text is processed in natural language,and the score feature is extracted.Then,using linear regression algorithm and neural network algorithm,according to the system data in the existing college English skills training,three kinds of intelligent scoring models were established for the spoken language questions.The first model establishes the model through the linear regression algorithm in machine learning.The second model establishes the model through the deep neural network in deep learning.The third model integrates the linear regression model and the deep neural network model to establish the model,and compare the effects of the three models.The conclusion is that the model established after the fusion of the linear regression model and the deep neural network model works best.After continuous training,the accuracy of the spoken language questions is as high as 90%.This article makes use of artificial intelligence to correct the subjective questions of spoken language expressions.This is of great significance to the instruction and correction of college English skills training.It is also very helpful for reducing the pressure of teacher correction,evaluating students' academic performance and guiding students' learning.