In order to prevent damage to machines, the analysis of sensor data is often helpful today. As soon as a machine behaves suspiciously, this is reflected in unusual data patterns. So-called machine learning (ML) algorithms, currently only provide clues as to "what" should be exchanged, but not "why". However, the service technician responsible for replacing a piece of machinery must also know why the algorithm suggests replacing a ventilation system, for example. The reason provided, e.g. insufficient air flow or the life of the component has been exceeded, helps the service technician to react adequately.
In order to make the deployment of the service technician even more efficient through self-explanatory predictions and to optimize the maintenance and production process, the DAIKIRI research project is an important milestone in the use of AI in the IoT environment. USU Software AG is acting as consortium leader and project coordinator for this joint project, which is funded by the German Federal Ministry of Education and Research. Other partners are the Data Science working group at the University of Paderborn, AI4BD Deutschland GmbH and pmOne AG. The 24-month research project started in January 2020.
"The goal is to enrich and train the ML algorithms with exactly the related data using new process technologies, in order to ultimately be able to automatically predict the "what" and "why" in a way that enables the service technician to understand the outcome. This creates trust and is the basis for making decisions," says Henrik Oppermann, Head of Research at USU.
Self-explanatory condition monitoring of machines for safe decisions
Already today, especially large corporations in the industry are intensively using AI technologies to optimize their production processes, i.e. to produce more and higher quality with the same effort. Medium-sized companies are thus under massive pressure and must react. As machines are increasingly equipped with sensor technology, every production step can now be precisely analysed and data extracted. Often, however, there is a lack of know-how to use the valuable industrial mass data through intelligent data processing. The aim is to make evaluations or forecasts transparent so that the data gained can be used even better to reduce production costs or optimize the production line.
Currently, the results of the algorithms can only be understood and interpreted by machine learning experts. This means that service technicians cannot understand how a result is generated. This traceability of data-driven predictions is, however, especially necessary for mechanical engineering in order to be able to make reliable decisions and avoid cost-intensive wrong decisions. DAIKIRI will therefore develop for the first time AI procedures which are self-explanatory and make the results of AI automatically "verbalized" and thus transparent.
Research Results Provide Added Value for USU AI Service Innovations
As an innovation leader for software solutions in the area of service management, USU is contributing comprehensive technical expertise and experience in setting up service platforms, AI technologies and machine data analysis to the research project. Explanatory AI procedures and the verbalization of results provide high added value for service in mechanical and plant engineering and ideally complement existing USU solutions. It also opens up new business areas for the use of machine learning, for example in medical technology.
In order to make the deployment of the service technician even more efficient through self-explanatory predictions and to optimize the maintenance and production process, the DAIKIRI research project is an important milestone in the use of AI in the IoT environment. USU Software AG is acting as consortium leader and project coordinator for this joint project, which is funded by the German Federal Ministry of Education and Research. Other partners are the Data Science working group at the University of Paderborn, AI4BD Deutschland GmbH and pmOne AG. The 24-month research project started in January 2020.
"The goal is to enrich and train the ML algorithms with exactly the related data using new process technologies, in order to ultimately be able to automatically predict the "what" and "why" in a way that enables the service technician to understand the outcome. This creates trust and is the basis for making decisions," says Henrik Oppermann, Head of Research at USU.
Self-explanatory condition monitoring of machines for safe decisions
Already today, especially large corporations in the industry are intensively using AI technologies to optimize their production processes, i.e. to produce more and higher quality with the same effort. Medium-sized companies are thus under massive pressure and must react. As machines are increasingly equipped with sensor technology, every production step can now be precisely analysed and data extracted. Often, however, there is a lack of know-how to use the valuable industrial mass data through intelligent data processing. The aim is to make evaluations or forecasts transparent so that the data gained can be used even better to reduce production costs or optimize the production line.
Currently, the results of the algorithms can only be understood and interpreted by machine learning experts. This means that service technicians cannot understand how a result is generated. This traceability of data-driven predictions is, however, especially necessary for mechanical engineering in order to be able to make reliable decisions and avoid cost-intensive wrong decisions. DAIKIRI will therefore develop for the first time AI procedures which are self-explanatory and make the results of AI automatically "verbalized" and thus transparent.
Research Results Provide Added Value for USU AI Service Innovations
As an innovation leader for software solutions in the area of service management, USU is contributing comprehensive technical expertise and experience in setting up service platforms, AI technologies and machine data analysis to the research project. Explanatory AI procedures and the verbalization of results provide high added value for service in mechanical and plant engineering and ideally complement existing USU solutions. It also opens up new business areas for the use of machine learning, for example in medical technology.
