Toward a Symbiotic Approach Leveraging Generative AI for Model Driven Engineering
Vinay Kulkarni, Sreedhar Reddy, Souvik Barat, Jaya Dutta
Abstract
Model Driven Engineering (MDE) proposes models as primary artefacts for analysis, simulation, software development etc. While MDE has delivered on the promise of enhanced productivity through automation, it continues to pose a significant entry barrier for domain experts who are typically not well-versed with MDE technology. With modelling gaining traction for analysis-heavy use cases like decision-making and regulatory compliance where domain experts play a central role, this barrier is beginning to hurt even more. We posit that Generative AI techniques can significantly lower this barrier by enabling domain experts to construct purposive models by operating at natural language level. This requires domain experts to interact with Generative AI tools using the right purpose-specific contextual prompts. We propose a model-driven approach where purposive meta models guide the interactions between domain expert and Generative AI to generate such prompts. The proposed approach helps in overcoming some of the limitations of Generative AI such as missing local context, limited context window size, attention fading etc. Industry scale models are typically large, necessitating a team of experts to work in a coordinated manner which requires sharing of outputs and persistence across sessions. Our approach brings together MDE and Generative AI in a symbiotic relationship complimenting respective strengths and overcoming limitations. We have validated this approach for development of digital twin based applications and early results are encouraging.