Research on the Application of Deep Learning-Based Intelligent Universal Language Listening and Speaking Training Systems in Minority Language Education
Year: 2024
Type: conference-paper
Abstract: As digitization technology develops by leaps and bounds, deep learning models have shown significant potential in language and text services. In this paper, we design and implement an intelligent universal language listening and speaking training system, it is intended to improve the efficiency and effectiveness of minority groups in learning the national common language through deep learning and natural language processing technologies. It is discovered that although Xinjiang and other minority regions have made significant progress in national common language education, they still face many challenges. For this reason, this paper presents a question generation model (GTCNQA) based on temporal convolutional networks (TCNs). The model provides an accurate context for question generation by keeping the correct answer and labeling related keywords; generates semantically relevant insertions by capturing the contextual information at the core of the answer using Graph Convolutional Networks (GCNs); and creates question sequences by splicing the features of word embeddings, keyword annotations, and GCN embeddings, and inputting them into the TCN encoder, which makes use of its time-sequence modeling capability. In the meantime, the model adopts a gated self-attention mechanism to combine contextual keywords and answers when generating questions, ensuring that the generated questions are fluent and semantically clear. The GTCNQA model combines the strengths of GCN and TCN to accurately create questions closely related to the content of the conversation, enhance the system's intelligence, and offer learners interactive and challenging learning environments to improve their listening and speaking comprehension and expression skills. Through natural language processing and deep learning technologies, the system realizes interactive scenario-based dialogue practice and learning situation analysis, which prominently improves the effect of listening and speaking training, and enhances the language skills and social integration ability of ethnic minority learners through personalized learning programs. The research presented in this paper provides new solutions and fresh perspectives in the field of language services.
Authors: Kexuan Zhu, Lifen Jiang, Dongbin Zhang, Xueying Zhang, Keyi Lu
Institution: Yili Normal University
Language: English
Topic: Speech and dialogue systems
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409965480