Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

On the Utility of Domain Modeling Assistance with Large Language Models

Meriem Ben Chaaben, Lola Burgueño, Istvan David, Houari Sahraoui

TL;DR

The aim of this approach is to overcome the need for extensive training of traditional AI-based completion algorithms on domain-specific datasets and to offer versatile support for various modeling activities, providing valuable recommendations to software modelers.

Abstract

Model-driven engineering (MDE) simplifies software development through abstraction, yet challenges such as time constraints, incomplete domain understanding, and adherence to syntactic constraints hinder the design process. This paper presents a study to evaluate the usefulness of a novel approach utilizing large language models (LLMs) and few-shot prompt learning to assist in domain modeling. The aim of this approach is to overcome the need for extensive training of AI-based completion models on scarce domain-specific datasets and to offer versatile support for various modeling activities, providing valuable recommendations to software modelers. To support this approach, we developed MAGDA, a user-friendly tool, through which we conduct a user study and assess the real-world applicability of our approach in the context of domain modeling, offering valuable insights into its usability and effectiveness.

On the Utility of Domain Modeling Assistance with Large Language Models

TL;DR

The aim of this approach is to overcome the need for extensive training of traditional AI-based completion algorithms on domain-specific datasets and to offer versatile support for various modeling activities, providing valuable recommendations to software modelers.

Abstract

Model-driven engineering (MDE) simplifies software development through abstraction, yet challenges such as time constraints, incomplete domain understanding, and adherence to syntactic constraints hinder the design process. This paper presents a study to evaluate the usefulness of a novel approach utilizing large language models (LLMs) and few-shot prompt learning to assist in domain modeling. The aim of this approach is to overcome the need for extensive training of AI-based completion models on scarce domain-specific datasets and to offer versatile support for various modeling activities, providing valuable recommendations to software modelers. To support this approach, we developed MAGDA, a user-friendly tool, through which we conduct a user study and assess the real-world applicability of our approach in the context of domain modeling, offering valuable insights into its usability and effectiveness.

Paper Structure

This paper contains 47 sections, 3 equations, 12 figures, 10 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background and Related work
  3. Assistance in Software Modeling
  4. Large Language Models and Few-shot Learning
  5. Prompting LLMs for Domain Model Completion
  6. Suggesting model elements
  7. Suggesting classes
  8. Suggesting attributes
  9. Suggesting associations
  10. Suggestion modes
  11. Suggestions On request
  12. Automatic suggestions
  13. Suggestions at the end
  14. Tool support
  15. Editor
  16. ...and 32 more sections

Figures (12)

  • Figure 1: Hospital domain model (in construction)
  • Figure 2: Completion Suggestions for Hospital Domain Model (classes in red, attributes in green, association in yellow).
  • Figure 3: High-level architectural view of MAGDA
  • Figure 4: User interface of the model editor
  • Figure 5: Simplified excerpt from the metamodel of the modeling framework of MAGDA
  • ...and 7 more figures