Selection and Optimization of University English Teaching Path Based on Knowledge Distillation and Transfer Learning

The wide application of artificial intelligence in the field of education makes how to effectively integrate new technologies in order to improve the quality and efficiency of education an urgent problem to be solved. This paper takes English teaching in colleges and universities as the research object, combines knowledge distillation and transfer learning, and constructs a path selection and optimization model for English teaching in colleges and universities. Among them, DenseNet and ANN are used as teacher model and student model respectively. At the same time, a teaching control experiment was designed to assess the actual teaching effect of the model-optimized college English teaching paths using independent sample t-tests and paired sample t-tests. The experimental group adopts the teaching path optimized by the model of this paper, while the control group adopts the traditional teaching method, and both of them are homogeneous in the dimensions of background variables, such as gender, ethnicity, place of origin, and English scores in the college entrance examination, as well as in the dimensions of English proficiency, such as listening, reading comprehension, oral expression, and written expression (P>0.05). In the posttest, there was a statistically significant difference between the scores of the experimental group and the control group (P<0.05). And comparing the experimental pre-test and post-test, the scores of all dimensions in the experimental class have significant differences (P<0.05), while there is no significant difference in the control group (P>0.05). This paper confirms that the model proposed can effectively optimize the English teaching process in colleges and universities, and the optimized path has a significant impact on the improvement of students’ English proficiency.