AI-based digital twin (KIDZ)

Insights gained through AI methods are often only available in isolation for partial aspects of a production process (e.g. the probability of failure of a single machine). The recognition of overarching patterns for the entire production process and product life cycle usually fails due to a lack of an overall model. In order to develop such an overall model, semantic annotation of the existing data is required, i.e. the enrichment of data sets with meta and context data. Insights gained by means of AI are brought into an overall context here. This improves the interpretability and explainability of the AI models and enables complex analyses and forecasts, especially through various simulation techniques. Methods that contribute to the understanding of AI (eXplainable AI) enable the description of AI models and their findings.