From ChatGPT to Sora:How Generative Artificial Intelligence Is Reshaping Deep Learning Scenarios
The rapid advancement of generative artificial intelligence(GAI)is reshaping production,daily life,and learning.Exemplified by ChatGPT and Sora,GAI has achieved transformative breakthroughs in text generation and video synthesis,respective-ly.These advancements not only accelerate the development of visualized,immersive,and highly interactive learning environments but also establish a robust technological foundation for deep learning through human-AI knowledge co-construction.The proposed deep learning pathway comprises four progressive and iterative cognitive stages,including perception activation,knowledge construction,knowledge transfer,and knowledge innovation,which collectively encompass twelve specific processes.At the knowledge innovation stage,the SECI model of knowledge creation is adopted to integrate GAI seamlessly into every phase of the learning process.This in-tegration results in a GAI-empowered deep learning model that supports the design of diverse and enriched learning scenarios.To en-hance the effectiveness of GAI integration,three strategic approaches are proposed:independent enhancement,integrated synthesis,and customized modeling.These approaches aim to facilitate deep conceptual understanding,enable knowledge internalization and transfer,foster creative problem-solving,and cultivate students'higher-order thinking skills.