Enhancing the Effectiveness of High Schools-Level English Translation Teaching Using Transformer Models and Reinforcement Learning Algorithms
Main Article Content
Abstract
Translation teaching is changed as artificial intelligence and deep learning constructs a wonderland. The traditional forms of teaching translation are devoid of personalized feedback, adaptive methods of learning, and dynamic error correction-all components that do not allow the student to attain high translation accuracy and fluency proficiency. This study investigates the application of a Transformer-based deep learning model and reinforcement learning algorithms in English translation education for the secondary school level. The study creates a system of AI-aided translation teaching which includes elements such as in-time feedback, adjusted translation exercises, and automated error diagnosis. The transformer models are for contextually accurate translations while reinforcement learning would ensure flexibility in difficult levels dependent on student performance. Effectiveness was evaluated through a study of 342 high school students concerning measures like translation accuracy and fluency, cultural adaptation, and error reduction. The findings showed apparent progress in translation performances, where BLEU scores rose from 45.3 to 62.7 and perplexity values declined from 21.4 to 14.8, reflecting greater translation precision and coherence. In addition, grammar errors decreased by 10.9%-12.6%, spelling errors by 11.4%-12.6% and contextual logic errors by 14.7%-19.5%, thus confirming the working of a reinforcement learning-based feedback mechanism. It concludes that AI-powered teaching of translations is an approach of learning that is scalable, efficient, and above all, student-centered in imparting translation precision, fluency, and cultural bearing. The combination of Transformer models and reinforcement learning is a promising perspective for the intelligent translation teaching of the future as it would enable real-time personalized feedback and optimization of learning throughout life.