Paper Title
Deep Learning Framework for Real-Time Student and Instructor Performance Prediction and Optimization

Artificial Intelligence (AI) and Machine Learning (ML) have been increasingly and quite successfully implemented in many fields over the last decade, but they still need to establish pedagogical advantages on a broader scale in educational technology. This study aims to establish a framework based on deep learning (DL) techniques that assess and optimizes student and instructor performance in classes using digital learning management systems (LMS) regardless of their delivery modality. We have developed methods to map selected student and instructor class activities and their respective performance parameters. The DL system,trained with historical data,determinesthe optimal impact factors for each performance parameter at various temporal checkpoints in the class. During the running of theclass, the system measures the activity parameters andassesses students standing at each checkpoint. After predicting student performance, it proposes the optimal action needed to maximize it. We apply a similar procedureto the instructors and their performance parameters. After completion of the class, we measure the model's accuracyand includeit in the future training set.The proposed framework represents a comprehensive approach that combines historical and real-time class data. We foresee a Big Data Learning Analytics Platform that can identify the optimal teaching and learning activitiesresulting in self-improving student and instructor performance tools. Keywords - Deep Learning, Learning Analytics, Learning Management System, Student Performance, Instructor Performance, Learning Activities