Shared Autonomy for Proximal Teaching
Project Overview
The document explores the implementation of Z-COACH, a generative AI framework designed to enhance motor skill learning in high-performance racing through shared autonomy. It underscores the significance of personalized instruction that aligns with a learner's Zone of Proximal Development (ZPD), showcasing a user study where Z-COACH notably improved driving performance compared to traditional self-practice methods. The findings reveal the promising advantages of integrating AI into educational contexts, particularly in skill acquisition and performance enhancement. However, the study also raises awareness of the potential risks associated with over-reliance on AI, suggesting a need for a balanced approach that leverages AI's capabilities while maintaining learner independence. Overall, the document highlights the transformative potential of generative AI in education, particularly in personalized learning experiences, while cautioning against excessive dependence on technological assistance.
Key Applications
Z-COACH
Context: High-performance racing training using a driving simulator
Implementation: Shared autonomy framework combining user inputs with autonomous agent assistance to identify and coach specific skills within a learner's ZPD.
Outcomes: Improved driving time, behavior, and smoothness; participants showed significant performance improvements compared to self-practice.
Challenges: Risk of over-reliance on AI assistance leading to loss of control skills.
Implementation Barriers
Technical
Complexity of integrating shared autonomy in a way that effectively supports skill development without causing over-reliance. This includes challenges in modeling the interaction between AI assistance and human learning processes.
Proposed Solutions: Design shared autonomy systems that optimize assistance based on student performance and ZPD. Incorporate educational psychology principles such as scaffolding and ZPD into the design of AI teaching systems.
Project Team
Megha Srivastava
Researcher
Reihaneh Iranmanesh
Researcher
Yuchen Cui
Researcher
Deepak Gopinath
Researcher
Emily Sumner
Researcher
Andrew Silva
Researcher
Laporsha Dees
Researcher
Guy Rosman
Researcher
Dorsa Sadigh
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Megha Srivastava, Reihaneh Iranmanesh, Yuchen Cui, Deepak Gopinath, Emily Sumner, Andrew Silva, Laporsha Dees, Guy Rosman, Dorsa Sadigh
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai