LAPIS (Learner-Adaptive, Pedagogical, Interactive Solutions)
LAPIS advances and aligns with learning and CS education research by applying feedback, motivation, and self-regulated learning theories while pioneering generative AI technology for just-in-time adaptive interventions.
Technology
Despite considerable research contributions, Introductory Computer Science (CS1) failure rates have only dropped from 33% to 25% over the last fifteen years. Some students struggle, causing immense student frustration and hindering positive learning outcomes. Personalized learning shows promise, but under-resourced schools, growing student numbers, and other obstacles hamper implementation. Additionally, diverse student backgrounds add complexities to making education personalized. Although Generative AI (GenAI) offers potential solutions, its application in education still needs to be explored, and many barriers to including GenAI and other educational technologies persist. UD inventors have developed LAPIS, a generative AI-integrated “smart teaching” augmentation system for CS1 courses that can be expanded to other mediums and course levels. Using an open-source Python curriculum, although not limited to it, LAPIS employs an AI-driven engine (Alchemy), leveraging a model representative of a student for real-time feedback, intervention, and logging. The novelty of the invention includes addressing student struggle with a just-in-time intervention utilizing a personalized feedback approach, determining necessary student information for personalized feedback, enhancing the generated AI-integrated model for improved results, and incorporating “low-AI” features (more traditional machine learning models) for reduced costs and adoptability. Additionally, instructors benefit from workload reduction by identifying general student struggles, making personalized intervention suggestions, and assisting in responding to student questions, reducing the administrative burden. LAPIS advances and aligns with learning and CS education research by applying feedback, motivation, and self-regulated learning theories while pioneering generative AI technology for just-in-time adaptive interventions. This invention is compatible with Learning Management systems and visualization tools, aiming to reduce instructor workload and enhance student learning in undergraduate CS1 courses.
Benefits
- Instructor workload reduction
- Offers students just-in-time assistance during coding
- Offer students real-time feedback
Applications
- Computer Science Professors