Artificial intelligence driven personalized vocabulary learning path optimization for English education

Abstract

Aiming at the problems of traditional English vocabulary learning path's lack of personalization and low matching efficiency of learning resources, this paper proposes an artificial intelligence (AI)-driven English vocabulary learning path optimization model for personalized learning. The model focuses on learning path optimization and learner portrait construction, and realizes dynamic modeling of learning state and adaptive path recommendation through deep learning (DL), graph embedding, multi-attention mechanism and knowledge tracing. In the process of processing, the model integrates time series features and lexical association features, and improves the ability of multi-source feature fusion through Gate Fusion Unit (GFU). At the same time, it combines the characteristics of multimodal resources such as text and image to improve the matching accuracy of learning content. Experimental results demonstrate the stronger performance of our model over traditional ANN model in terms of recommendation accuracy, recall, processing efficiency and user satisfaction? The top recommendation accuracy rate among them is 90.7%. The research finds that this approach is a promising way to boost the personalized degree and learning effectiveness of English word acquisition.