A modeling assistant for cognifying MBSE tools

The steep learning curve of tools for Model-Based Software Engineering still slows down its widespread adoption. Users indeed encounter a variety of technical difficulties when modeling, including using modeling tools that are often considered too complex. Because they usually require too much advanced or diverse skills, modeling tools push users to focus on how to model with the tool instead of on what to model with the tool, and hence curb users’ creativity and productivity. Our goal is to address this issue by refocusing the modeling activity on the content of the model, and by facilitating the creation of new model elements. In this tool demonstration, we present our modeling assistant, which gathers knowledge from model repositories and provides in-editor modeling recommendations for users to complete their models. Our system cognifies the Papyrus modeling tool by augmenting its users with a recommender system fed with more than 90,000 models to assist UML class diagrams authoring. It features an architecture scalable to multiple recommendation clients, and can be deployed locally to ensure the confidentiality of the source models. We showcase our modeling assistant with a use case and discuss preliminary measures.