#programming #technology #merging #education

The rapid evolution of AI and data-driven technologies presents both opportunities and challenges in education. In this research talk, I will explore the synergy between AI and human elements in education. This bi-directional bridge involves using AI models enriched with human knowledge for educational purposes while harnessing AI to enhance education and knowledge. Dr. Yang Shi will begin by discussing how domain features in students' programming data can uncover common errors without labels through clustering code embeddings. This technology offers a novel approach to error discovery. Dr. Shi will then delve into deep models for knowledge tracing tasks, providing predictive support to students while offering interpretability on code structures leading to specific predictions. These projects demonstrate how incorporating domain features enhances data-driven models in education. Dr. Shi will then introduce the KC-Finder model, which autonomously uncovers students' practiced knowledge from programming logs, thanks to the incorporation of knowledge component properties. This work highlights the potential of merging educational theory with AI models for education. Dr. Shi's research has focused on infusing human knowledge into AI models, and in the future, he plans to develop tools and methods to further enhance human knowledge through improved AI models. Dr. Shi will conclude by briefly outlining his future research plans in this direction, offering a glimpse of the exciting possibilities ahead.