Probeable Problems for Beginner-level Programming-with-AI Contests
Project Overview
The document explores the use of generative AI in education, particularly through the implementation of 'Probeable Problems' in competitive programming contests tailored for beginner-level computer science students. These problems are designed to lack certain details, compelling participants to engage in critical thinking by identifying omissions and asking clarifying questions. This strategy aims to counteract the advantages that AI coding tools, such as GitHub Copilot, may confer on students, ultimately promoting essential problem-solving skills. A study conducted during a multi-institutional contest assessed the effectiveness of this approach, revealing that while many students found it challenging to pinpoint the missing information, the experience significantly enhanced their learning. The findings suggest that incorporating such problem-solving frameworks can be beneficial in educational settings, as they encourage deeper cognitive engagement and a more nuanced understanding of programming concepts, despite the initial difficulties faced by students. Overall, the document highlights the potential of generative AI to transform educational practices by fostering critical skills necessary for the digital age.
Key Applications
Probeable Problems
Context: Competitive programming contests for undergraduate computer science students
Implementation: Conducted a 2-hour contest where students worked individually or in teams to solve Probeable Problems using AI tools.
Outcomes: Encouraged students to engage critically with programming specifications, although overall performance in identifying missing details was low.
Challenges: Participants struggled to determine missing details and the problems were deemed too challenging given the time constraints.
Implementation Barriers
Educational
Students may develop an over-reliance on AI code-generation tools, potentially hindering their coding skills. Additionally, AI tools can easily solve beginner-level programming problems, making it difficult to gauge students' true abilities.
Proposed Solutions: Implement Probeable Problems to encourage critical thinking and the ability to recognize ambiguities in specifications. Design problems that specifically require human intervention to resolve ambiguities, as seen with Probeable Problems.
Logistical
Time constraints during the contest limited participants' ability to fully engage with the problems.
Proposed Solutions: Future contests could offer fewer problems or provide more time for completion.
Project Team
Mrigank Pawagi
Researcher
Viraj Kumar
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Mrigank Pawagi, Viraj Kumar
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