Prize for contribution to AI error prediction

to HFU News

Best Paper Award for team from the WI faculty at computer science conference

The contribution of an HFU research project on the topic of artificial intelligence was awarded the prize for the best paper at the most recent conference of the International Academy, Research and Industry Association (IARIA). The award was presented to Valentin Göttisheim, Holger Ziekow, Ulf Schreier and Alexander Gerling at the Eleventh International Conference on Data Analytics.

The paper "Shapley Values based Regional Feature Importance Measures Driving Error Analysis in Manufacturing" deals with the use of AI to prevent errors in production processes. The AI learns to predict errors using measurement data from the production process. By analyzing the model, quality management can obtain information on the causes of errors and process improvements. The paper explains this process, which can draw companies' attention to particularly relevant inspection characteristics. Factors for the occurrence of errors can vary greatly depending on the production process, from fluctuating ambient temperatures to waiting times between individual work steps. One challenge is to identify the relevant values in the often very large number of inspection characteristics.  The AI-based method enables companies to assess for which characteristics an in-depth analysis is promising.

More information about the conference:

Research blog article on the AI method in production:
KI und der Blick in die Kristallkugel