Using machine learning and artificial intelligence for source-code related tasks has been done for a while now. There are tools that can “understand” software code without compiling it and without the need to parse it.
However, modelling and architectural design decisions are different as they operate on models and other, non-code artifacts.
In this project, we explore the possibilities of extracting features from models for the purpose of using AI to make architectural decisions and to check the conformance of the source code to these decisions.
Funding agency: CHAIR (Chalmers AI Research Center)