Recent trends towards data-driven methods may require a substantial rethinking regarding the practice of Modeling &Simulation (M&S). Machine Learning (ML) is now becoming an instrumental artefact for developing new insights, or improving already established knowledge. Reflecting this broad scope, the paper presents a conceptual framework to guide the integration of simulation models with ML. At its core, our approach is based on the premise that system knowledge can be (partially) captured and learned from data in an automated manner aided by ML. We conceive that the approach can help realise adaptive simulation models that learn to change their behaviour in response to behavioural changes in the actual system of interest. Broadly, the study is conceived to foster new ideas and speculative directions towards integrating the practice of M&S with data-driven knowledge learned by ML.