Intelligent Dialogue Method for Educational Applications Based on Improved Large Models

This paper designs an improved large model by incorporating a fully connected layer into the pre-trained large model. Furthermore, an intelligent dialogue method for educational applications is devised, leveraging the improved large model to transform user dialogue into structured data demanded by the applications. In the example of the complaints handling application in universities, user complaints are automatically categorized into appropriate consulting units, the effectiveness of this method is validated with an accuracy rate of 93.2%.