| KEYWORD |
User stories generation through the use of screen captures and LLMs
Reference persons RICCARDO COPPOLA
Research Groups DAUIN - GR-16 - SOFTWARE ENGINEERING GROUP - SOFTENG
Thesis type RICERCA
Description The thesis aims to explore the application of large language models (LLMs) for generating user stories from screen captures, with the goal of aiding end-to-end (E2E) software testing using Playwright. The research will focus on leveraging LLMs' ability to process and understand visual input, such as screen captures, to automatically generate coherent and detailed user stories that accurately describe user interactions with the software under test. These user stories will then serve as the foundation for E2E tests, which will be implemented and automated using the Playwright testing framework.
By automating the creation of user stories from visual content, the thesis seeks to streamline the testing process, reduce the time and effort required for manual test case design, and enhance the accuracy of testing coverage. The work will involve a combination of natural language processing, computer vision, and software testing methodologies. Additionally, it will examine the efficacy of LLM-generated user stories in representing real-world user behaviors and their impact on the robustness of E2E tests.
The expected outcomes include a framework or tool that integrates screen capture-based user story generation with automated testing, along with an evaluation of its performance in different software development environments.
Deadline 07/10/2025
PROPONI LA TUA CANDIDATURA