| KEYWORD |
Critical Appraisal of LLM Code Generation Capabilities in Educational Context
keywords AI, CODE GENERATION, EDUCATION, LLM
Reference persons MARCO TORCHIANO
Research Groups DAUIN - GR-16 - SOFTWARE ENGINEERING GROUP - SOFTENG
Thesis type EMPIRICAL RESEARCH
Description Recently, several automatic code generation tools based on AI techniques have been made available to users. These are specific products (e.g., CodePilot) or those with broader purposes (such as ChatGPT) that are capable of generating high-quality code from requirements expressed in natural language. These tools will soon enter the common toolset for software developers and therefore must be integrated into university training projects (and not only). The goal of this thesis is to experimentally evaluate the quality level of the code generated in the context of typical university course exercises, identify the limits, define strategies to mitigate them and improve the result. Additionally, it will be necessary to identify the skills required for students and future developers to effectively interact with this type of tools.
Required skills Java development, Software Engineering, possibly Machine Learning
Deadline 11/01/2025
PROPONI LA TUA CANDIDATURA