Sustainable use of AI-based electronic noses in the life science field in changing environments

Researcher

Julius Wörner

Publications

• Julius Wörner, Helene Dörksen, Miriam Pein-Hackelbusch (2023): Key Indicators for the Discrimination of Wines by Electronic Noses Full Publication
• Julius Wörner, Maurice Moelleken, Joachim Dissemon, Miriam Pein-Hackelbusch (2023): Supporting wound infection diagnosis: advancements and challenges with electronic noses Full Publication
• Julia Kruse, Julius Wörner, Jan Schneider, Helene Dörksen, Miriam Pein-Hackelbusch (2024): Methods for Estimating the Detection and Quantification Limits of Key Substances in Beer Maturation with Electronic Noses Full Publication

Collaboration

Alaa Othman

Research Theme

R2 Prosilience & Robustness

The project integrates an electronic nose (e-nose) in various life science use cases in order to advance the sustainable and intentional use of artificial intelligence in this research field. Methodically, relevant volatile substances found in life science applications are analyzed in a simplified environment in the laboratory via an e-nose and processed with machine learning algorithms to assess the applicability in real processes. In addition to the focus on discrimination, the regression and prediction of concentrations of certain key volatile markers also play a role. For example, the project is also looking at the production of beer and the monitoring of various key substances. The lowest concentration at which several volatile aromas in beer can be detected is analyzed to determine whether an e-nose would be suitable for process control and quality assessment. Central problems such as data drift, low concentrations of the target substance or small data sets are major problems of all use cases that need to be solved using suitable algorithms. This methodical approach of solving sub-problems of different applications provides valuable input to address our long-term goal of diagnosing wound infections with e-noses.