Identification of English Teaching Effectiveness and Students’ English Proficiency Development in Colleges and Universities Combining Refined Bilinear Attention Networks

Abstract

Abstract With the current application of artificial intelligence in education, English teaching methods in colleges and universities are developing rapidly. In this paper, we explore the application of the streamlined bilinear attention network to enhance students’ English proficiency and evaluate teaching effectiveness to improve teaching efficiency and improve students’ English proficiency. The study applies the bilinear attention mechanism to effectively process visual information during college English teaching using computerized visual learning methods to improve the accuracy of classroom teaching quality assessment. The methodology includes implementing a bilinear attention mechanism and applying a deformable convolution module to extract more accurate image features and optimize model performance. The model achieves 98.89% accuracy in facial expression recognition on the CK+ dataset, and 88.5% and 78.9% on the RAF-DB and Fer2013 datasets, respectively. Through online emotion evaluation experiments, the model performed well in recognizing emotional activity in teaching videos, which is highly consistent with the manual evaluation results. In analyzing students’ English proficiency development, the model-assisted teaching method can significantly improve students’ listening, grammar, phonetics, and many other abilities, with a mean value of self-evaluation reaching 4.09576 points. It shows the effectiveness of the streamlined bilinear attention network in assessing the teaching effectiveness of English teaching in colleges and universities and improving students’ abilities. Not only does the model accurately identify emotional activity in the classroom, but it also enhances students’ English learning.