چارچوب مبتنی بر یادگیری انتقالی (Transfer Learning) و مدلهای Attention-Mechanism برای شخصیسازی آموزش پزشکی با استفاده از دادههای ناهمگون و پراکنده
کد: G-1400
نویسندگان: Ghadir Pourbairamian * ℗
زمان بندی: زمان بندی نشده!
برچسب: دستیار مجازی هوشمند
دانلود: دانلود پوستر
خلاصه مقاله:
خلاصه مقاله
Background and aims: With advancements in artificial intelligence (AI)-driven educational technologies, personalized learning processes in medical education have become feasible. Educational data in this domain are often heterogeneous and fragmented, encompassing resources such as academic records, clinical performance, instructor feedback, and learner interactions with digital learning systems. Classical machine learning models struggle to handle such data effectively. This study proposes a hybrid framework integrating transfer learning and attention mechanisms to optimize personalized educational pathways by transferring knowledge from analogous domains with high precision. Method: The proposed model employs transfer learning to leverage insights from prior learners and utilizes attention mechanisms to prioritize information critical to academic progress. First, heterogeneous data from diverse educational sources—including textual, numerical, and behavioral data—were preprocessed and transformed into multi-dimensional vectors. A deep neural network model with an attention mechanism was then applied to identify salient features. Finally, a transfer learning model was implemented, utilizing historical student data to refine predictions for personalized learning trajectories. Model performance was evaluated using accuracy, recall, and F1-score metrics on real-world datasets. Results: The results demonstrated the superiority of the proposed framework over traditional machine learning models and classical neural networks in predicting individual learning needs. The transfer learning model achieved higher accuracy in personalizing education, even with limited data, while the attention mechanism played a pivotal role in identifying and focusing on critical educational information. Conclusion: This study highlights the potential of combining transfer learning and attention mechanisms to create a robust framework for enhancing personalization in medical education. The approach not only optimizes learning pathways but also enables continuous improvement of educational systems by leveraging historical learner experiences. Future research should explore the application of this methodology across diverse educational environments and analyze its impact on learning outcomes.
کلمات کلیدی
Transfer Learning, Attention Mechanism, Medical Education
