TipsC: Tips and Corrections for programming MOOCs
Project Overview
The document discusses the implementation of TipsC, a generative AI tool specifically developed for enhancing programming courses in massive open online courses (MOOCs). TipsC offers personalized feedback on code submissions by analyzing students' errors and suggesting corrections to logical runtime issues without providing full solutions, thereby fostering independent problem-solving skills. Additionally, it aids instructors by visualizing submission patterns and improving grading efficiency through clustering similar submissions, which streamlines the assessment process. As an open-source and scalable tool, TipsC not only aims to enhance student learning outcomes but also reduces reliance on teaching assistants, making it a valuable resource in the educational landscape. The findings indicate that generative AI applications like TipsC can significantly improve the learning experience and operational efficiency in online education environments.
Key Applications
TipsC - a tool for analyzing and clustering programming submissions.
Context: Massively Open Online Courses (MOOCs) teaching programming.
Implementation: Implemented in Scala, integrated into MOOC platforms to analyze student submissions and provide feedback.
Outcomes: Reduced grading variance by 47% in clusters; improved student performance and engagement; provided personalized hints without disclosing full solutions.
Challenges: Ensuring the balance between helpfulness and avoiding spoon-feeding; preventing solution leakage.
Implementation Barriers
Technical challenge
The computational complexity of clustering and comparing programs can be high, especially with large submissions.
Proposed Solutions: Utilizing efficient algorithms and parallel processing to manage performance.
Educational challenge
Potential for students to become overly reliant on automated hints instead of learning to debug their code.
Proposed Solutions: Designing TipsC to provide hints that guide students without giving away complete solutions.
Project Team
Saksham Sharma
Researcher
Pallav Agarwal
Researcher
Parv Mor
Researcher
Amey Karkare
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Saksham Sharma, Pallav Agarwal, Parv Mor, Amey Karkare
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