Machine Learning Engineering: Infusing Software Engineering through the Semantic Web

The Semantic Web provides a wealth of high-quality, structured, and contextual data, which can be used to train machine learning models. This can lead to the creation of models, i.e., the engineering of Machine Learning, that adhere to non-functional requirements, which include considerations such as safety, security, and reliability, which are key elements of Software Engineering. These requirements do not concern a system's functionality, but rather its quality attributes. By incorporating these concepts into the engineering of machine learning models, one can strive to create models that are secure, reliable, and exhibit the desired quality attributes. Furthermore, Verification and Validation, or V&V, is integral to successful software engineering, by ensuring that a system is implemented correctly and meets specified requirements. In engineering Machine Learning, it's equally important to define processes and methods to thoroughly test and validate models to ensure they're performing as expected and providing accurate results. Together, the fusion of Software Engineering principles into Machine Learning Engineering, aided by the Semantic Web's capabilities, can bolster trustworthiness in machine learning systems. This trustworthiness ensures that the systems can be relied upon to behave as expected. In essence, by combining these fields, one can develop machine learning systems that are reliable, secure, interpretable, and trustworthy, upholding the core principles of Software Engineering. Our research group focuses on the most recent developments in these areas, i.e., Knowledge Graphs and Large Language Models, to accomplish these goals.