Efficient AI-supported quality assessment for electronics production

In the BMBF project AdaPEdge, process and sensor data is collected, condensed, evaluated using AI and compared with a defined optimum, independently of the existing digital production infrastructure, in order to initiate measures for production optimization at an early stage. The development process used in the project for practical AI models and their optimization for use on edge devices will be presented. Existing human expertise is incorporated in order to increase the efficiency and acceptance of AI.
As team leader of the "Machine Learning" department in the "Cognitronics and Sensor Technology" specialist group at Bielefeld University, the speaker Dr. Michael Thies focuses on industry-related AI applications for monitoring and predicting product quality. The focus here is on specialized solutions that can be explored and put into operation with scarce and unbalanced training data.