中文
HOME
HOME / LATEST NEWS / Content

Advancing AI Education in Computer Science Classrooms

December 31, 2025

As AI education deepens, teachers of the Computer Science Department are actively leveraging interactive platforms to develop lightweight, user-friendly teaching tools that enhance classroom interaction and effectiveness. During the professional develpment session held at the end of December, two teachers shared their experiences integrating AI teaching aids into classroom instruction. Their cases centered on the advanced concept of reinforcement learning, designing both engaging and inquiry-based pedagogical approaches.


B4E6


Mr. Yuan independently developed an "Interactive Reinforcement Learning Maze Tool" using the Cursor platform. This tool transforms the abstract algorithm into a visual game environment. Through group collaboration, students place obstacles in a maze, design paths, adjust reward/punishment parameters, and personally "train" AI agents. With real-time data visualization, learners directly observe how strategy adjustments impact scores, thereby grasping core relationships among states, actions, rewards, and optimal policies through learning by doing. 



Mr. Xu designed an "Intelligent Recommendation System Inquiry" case. He pre-deployed an interactive web-based teaching aid simulating a "Xiaohongshu" (Chinese social media platform) recommendation interface, enabling students to compare outcomes between AI-generated recommendations and self-defined strategies. By testing approaches like "balanced exploration" and "pure exploitation," learners clearly observed changes in recommendation scores, gaining deep understanding of the core exploration-exploitation tradeoff mechanism in reinforcement learning. This case ingeniously bridges real-life scenarios, effectively lowering comprehension barriers for complex concepts like supervised learning and reinforcement learning.


During the workshop discussion, teachers engaged in lively exchanges on expanding tool functionalities and optimizing instructional design. Participants unanimously agreed that platforms like Cursor and Trae empower educators to rapidly build visual, operable AI teaching tools. These not only significantly reduce technical barriers but also transform teaching aids into learning tools, equipping students with hands-on AI capabilities—from training and demonstration to analysis and iterative optimization—achieving authentic learning by doing.


(Written/Pictures by Computer Science Department    Reviewed by Qian Zuo)