Large Language Models as Recommender Systems in Uncertainty-Aware Requirement Engineering
|
Typ | Masterarbeit |
|---|---|---|
| Aushang | Req2RELAX M Sc Thesis Proposal final 2.pdf | |
| Betreuer | Wenden Sie sich bei Interesse oder Fragen bitte an: Vincenzo Scotti (E-Mail: vincenzo.scotti@kit.edu), Tobias Hey (E-Mail: hey@kit.edu, Telefon: +49-721-608-44765) |
Motivation
In current requirement engineering approaches there is limited automation support to recognize and manage uncertainties in the creation and specification of system requirements. The main idea of this master thesis is to use the generative capability of Large Language Models (LLMs) to integrate cross-cutting concerns related to uncertainty and flexibility into requirements specification of a system.
Tasks
The project involves surveying the state of the art on how uncertainty is addressed during requirements elicitation, followed by an in-depth study of the RELAX language for handling uncertainty in self-adaptive systems requirements. It also includes exploring large language models and prompt-engineering techniques, and defining and developing an LLM-based pipeline capable of transforming a conventional requirements specification document into one that conforms to RELAX. Finally, the work requires an experimental evaluation through a quantitative evaluation and a qualitative expert study. The expert study is aimed at assessing the quality of the generated outputs against established acceptance and quality criteria.
Tools/Technology
RELAX, Python, Transformers, Prompting, In-Context Learning