The system leverages a Retrieval-Augmented Generation (RAG) Large Language Model to deliver automated, domain-aware feedback on UML class diagrams. By grounding responses in a curated knowledge base, the model provides context-sensitive guidance that supports students in identifying omissions, improving structural accuracy, and refining their design choices.

Feedback can be obtained via predefined or free-form queries, enabling both structured support for novices and flexible exploration for advanced users. This mechanism not only enhances modeling outcomes but also fosters self-regulated learning through iterative reflection and revision.

It exports all the Logs generated in the XES format. The XES standard defines a grammar for a tag-based language whose aim is to provide designers of information systems with a unified and extensible methodology for capturing systems behaviors by means of event logs and event streams is defined in the XES standard. An example of event log is available here.

It finds out the actual modelling process which is happening inside a software project. To this aim, it allows to execute the Declare Miner and MINERful techniques to process discovery.

It figures out if there are any deviations between the actual modelling process and the ideal modelling process (from the model). To this aim, it allows to execute the Declare Analyzer and the Declare Replayer methods to conformance checking. The conformance can be explored either by trace or by constraint.

It serves as a powerful tool for refining searches within log files containing student modeling actions, enabling users to extract valuable insights and support effective instructional practices in UML diagram development.

It provides a capability to generate human-readable descriptions of UML diagrams and models. Providing the description in natural language of the solution given by the teacher enriches LLM-RAG’s ability to offer detailed and contextually relevant feedback.