Scribd
Upload
2 views13 pages

AB PPT

This seminar explores the impact of ChatGPT on the quality of Java code written by novice programmers, focusing on adherence to coding standards and complexity metrics. The study finds that ChatGPT can improve code quality and reduce complexity, but challenges remain in complex topics. Future work will involve larger studies and strategies to prevent over-reliance on AI tools in programming education.

Uploaded by

vaibhav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views13 pages

AB PPT

This seminar explores the impact of ChatGPT on the quality of Java code written by novice programmers, focusing on adherence to coding standards and complexity metrics. The study finds that ChatGPT can improve code quality and reduce complexity, but challenges remain in complex topics. Future work will involve larger studies and strategies to prevent over-reliance on AI tools in programming education.

Uploaded by

vaibhav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

TECHNICAL SEMINAR (21CS81) ON


Popular Roommates In Simply Exponential Time
2024-2025

Arjun Bharadwaj Guided By Dr. Sunita Chalageri


1KS21CS119
Table of contents
1. Introduction
2. Literature Survey
3. Problem Statement
4. Proposed Solution
5. Implementation
6. Results
7. Future Work
8. Conclusion
9. References
Introduction

•ChatGPT, a Large Language Model (LLM), has been increasingly used in education
to assist with coding tasks.

•This study focuses on assessing how ChatGPT impacts the quality of Java code
written by novice programmers.

•By analyzing static code quality through rule adherence and complexity metrics, the
research seeks to determine ChatGPT's potential benefits in programming education.

•The findings aim to guide educators in integrating AI tools to improve learning


outcomes.
Literature Survey
TITLE AUTHORS CONTRIBUTION ADVANTAGES DRAWBACKS

Explored ChatGPT’s role in code


Limited effectiveness in domain-
refinement, found it outperformed Improved code quality by
specific tasks.
"Refining ChatGPT- traditional tools but struggled with adhering to coding conventions
Generated Code" Guo et al. (2024) complex, domain-specific code. and reducing errors.
Analyzed Codex's impact on
coding ability, noting
"Impact of OpenAI improvements but also the need Enhances novice programmers'
Codex on Novice Kazemitabaar et al. for supervision to prevent over- ability to write functional code Requires substantial supervision to
Programmers" (2023) reliance on AI. with less effort. avoid dependency on the tool.
Studied ChatGPT's integration in
software engineering education,
"How ChatGPT is noting its benefits in assisting Supports student learning by Over-reliance on AI tools could hinder
Transforming Software Daun and Brings students but warning against offering personalized feedback critical thinking and problem-solving
Engineering Education" (2023) unsupervised use. and instant assistance. skills.
Surveyed student experiences
using ChatGPT in introductory
"ChatGPT in programming courses, Increases student engagement Could lead to a reliance on AI for
Programming highlighting improvements in by offering multiple solutions, problem-solving, potentially reducing
Education: A Survey of engagement and understanding enhances understanding of critical thinking and independent
Student Experiences" Jacques (2023) coding concepts. coding concepts. coding skills.
Problem Statement
•Novice programmers often produce code that is error-prone, hard to maintain,
and non-compliant with coding standards.

•Without proper support, these issues can lead to frustration and hinder learning
outcomes.

•There is a gap in understanding how AI tools like ChatGPT can specifically


address the challenges of adhering to coding conventions and reducing
complexity.

•The study seeks to determine if ChatGPT can meaningfully assist in improving


code quality without undermining learning.
Problem Statement

•RQ1: What violations to coding conventions can be observed if students use or


do not use ChatGPT for programming exercises?

•RQ2: What are the most violation-dense topics in programming exercises when
students use or do not use ChatGPT?

•RQ3: How does ChatGPT usage influence the cyclomatic and cognitive
complexity of students' code?
Proposed Solution

•Leverage ChatGPT as an assistive tool for programming assignments, allowing


students to use it as needed for guidance.

•Static code analysis tools like Checkstyle will evaluate adherence to coding
conventions, while SonarQube will measure complexity.

•Students code will be analyzed to see if ChatGPT helps reduce errors and
improve readability and maintainability.

•The study will also explore whether ChatGPT influences students understanding
of good coding practices.
Implementation

• Two student groups were studied: the control group (without ChatGPT) and the
treatment group (with ChatGPT).

• Programming exercises included topics like loops, OOP, file handling,


collections, and multithreading.

• The lecturer provided personalized feedback on weekly submissions to help


students improve their coding skills.

• The analysis included statistical comparisons of rule violations and complexity


metrics between the two groups.
Results
Future Work

•Larger, more diverse student populations should be studied to validate the


findings and enhance generalizability.

•Investigating the use of ChatGPT in other programming languages or advanced


concepts like algorithms could offer broader insights.

•Further research should explore how to design prompts for ChatGPT to maximize
its effectiveness for learning.

•Educators should develop strategies to prevent over-reliance on AI tools while


encouraging critical thinking and independent problem-solving.
Conclusion
•Findings for RQ1:
Students using ChatGPT had fewer coding rule violations, particularly for rules like LineLength,
FinalParameters, and DesignForExtension, indicating better adherence to coding standards.

•Findings for RQ2:


Despite overall improvements, areas like Collections, File IO and Streams, and Object-Oriented
Programming remained the most challenging, with high violation rates in both groups. ChatGPT
showed some benefits but was less effective in these complex topics.

•Findings for RQ3:


ChatGPT usage reduced both cyclomatic and cognitive complexity, making code simpler and easier
to understand. However, the differences were modest and highlighted the need for further research
with larger populations.

•Overall Conclusion:
ChatGPT shows potential to enhance programming education by improving code quality and
reducing complexity. While promising, thoughtful integration is necessary to maximize its benefits
and address its limitations.
References

•Haindl, P., & Weinberger, G. (2024): Does ChatGPT Help Novice Programmers
Write Better Code? IEEE Access.

•Liu, Y., et al. (2024): Refining ChatGPT-generated code. ACM Trans. Softw. Eng.
Methodol.

•Guo, Q., et al. (2024): Exploring the potential of ChatGPT in automated code
refinement.

•Oracle’s Code Conventions for Java: (https://www.oracle.com/java/technologies/).

•Checkstyle: Code quality and static analysis tool (https://checkstyle.sourceforge.io/).


References
5. Brandl, F., Kavitha, T.: Two problems in max-size popular matchings.
Algorithmica 81(7), 2738–2764 (2019)

6. Chung, K.: On the existence of stable roommate matchings. Games Econ.


Behav. 33(2), 206–230 (2000)

7. Condorcet method. https://en.wikipedia.org/wiki/Condorcet_method

8. Cseh, Á.: Popular matchings. Trends in Computational Social Choice, Ulle


Endriss (ed.) (2017)

9. Cseh, Á., Huang, C.-C., Kavitha, T.: Popular matchings with two-sided
preferences and one-sided ties. SIAM J. Discret. Math. 31(4), 2348–2377 (2017)

You might also like