Introducing Variational Autoencoders to High School Students
Project Overview
The document explores the integration of Variational Autoencoders (VAEs) in K-12 education, particularly targeting high school students, through innovative teaching methods that incorporate creative tools and philosophical metaphors. It emphasizes the use of interactive web-based games and hands-on experiences, which serve to demystify the complexities of generative AI concepts. The findings from pilot studies reveal that this approach significantly enhances student engagement and comprehension of AI principles, effectively bridging theoretical knowledge with practical application. By fostering an interactive learning environment, the document underscores the potential of generative AI not only as a subject of study but also as a means to inspire critical thinking and creativity among students in the evolving landscape of education.
Key Applications
Interactive Learning Tools for Variational Autoencoders (V AEs)
Context: High school education targeting students interested in AI and philosophy, combining technical and philosophical concepts to enhance understanding.
Implementation: A web-based game (Shadow Matching Game) and a simplified Google Colab notebook designed for hands-on learning and retraining of Variational Autoencoders using hand-written digits. Philosophical concepts, such as Plato's cave allegory, were integrated to ground AI concepts.
Outcomes: Students demonstrated a comprehensive understanding of the encoder and decoder roles, engaged with the complete cycle of machine learning development, and found philosophical ideas relatable to abstract AI concepts.
Challenges: Students with no coding experience struggled with the Google Colab interface and some found the philosophical content abstract, benefiting from more direct examples.
Implementation Barriers
Technical Barrier
Complexity of the Google Colab interface for students with little or no coding experience.
Proposed Solutions: Simplified user interface and additional instructional support could be provided.
Content Barrier
Philosophical concepts may be too abstract for some students.
Proposed Solutions: Using more direct and relatable examples to explain philosophical ideas connected to AI.
Curriculum Barrier
Lack of foundational knowledge in AI for students without prior experience, which may hinder the learning of more advanced AI concepts.
Proposed Solutions: Incorporate a broader introduction to AI concepts to better prepare students for learning V AEs.
Project Team
Zhuoyue Lyu
Researcher
Safinah Ali
Researcher
Cynthia Breazeal
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Zhuoyue Lyu, Safinah Ali, Cynthia Breazeal
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18