Year: 2025
Type: article
Abstract: This study presents an artificial intelligence-driven intelligent system framework for real-time assessment, instant correction, and adaptive learning feedback in digital and remote education environments. By integrating computer vision, deep learning, and reinforcement learning, the system provides a data-driven approach to monitoring learner performance and customizing instructional support. A hybrid residual network-bidirectional long short-term memory architecture extracts spatial and temporal features to classify learner actions and predict proficiency levels with high accuracy. Simultaneously, the Soft Actor-Critic algorithm dynamically adjusts feedback timing and difficulty based on ongoing performance metrics, ensuring a personalized and optimized learning trajectory. Validation using 10-fold cross-validation yielded 97.3% classification accuracy and low prediction errors (mean squared error and mean absolute error), while comparative evaluations with conventional teaching methods showed notable gains in learning outcomes and feedback effectiveness.
Authors: XiuYan Su, Ran Huo, Chang Liu, Xiling Liu
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W4414508778