Synthetic Model Generation from Metamodel Using Conditional Tabular Generative Adversarial Network

Building a dataset of models that are metamodel instances for a Model-Driven Engineering (MDE) process using Artificial Intelligence (AI) techniques is now gaining traction. These datasets are rare. For this reason, the availability of such datasets is a challenge in the context of MDE and AI. The central idea of the paper is to use deep learning techniques to build instances of a metamodel that constitute a dataset. Thus, our contribution focuses on creating synthetic instances of a metamodel using a Conditional Tabular Generative Adversarial Network (CTGAN).