Introduction of an industrial transfer learning use case systematization for machine tools

Traditional intelligent systems such as fault detection are mainly developed for isolated use on single machines and are mostly trained individually. A cross-machine knowledge transfer of intelligent systems holds an immense potential and can reduce implementation effort of intelligent systems. To enable a structured analysis of real-world transfer problems, an industry-relevant, far-reaching systematization of typical machine tool use cases is developed. This provides the overview over different typical use case classes from the industrial use case perspective. Based on the derived transfer criteria of each use case class, a conceptual approach of concrete transfer methods is proposed.