AI for Chemical Risk Prediction in Aquatic Environments

The interdisciplinary AI4ChemRisk research team, coordinated by Prof. Kloft, combines expertise from environmental sciences, computer science and chemical process engineering to identify and predict chemical risks in freshwater ecosystems on a global scale. The starting point is the increasing pollution of water bodies by chemicals from agriculture and wastewater. A well-founded assessment of these risks has hardly been possible to date due to a lack of measurement data and the complex interaction of environmental factors such as topography, weather and hydrology. The project is therefore developing innovative AI models that automatically recognize chemical pollution, realistically simulate missing measurements and take physical and ecological relationships into account. Using deep learning methods for anomaly detection and generative models, new approaches are being created to close data gaps and predict developments more precisely. In addition, user-friendly tools such as voice interfaces are being developed to facilitate access to analyses and support data-based decisions.

The aim is to identify high-risk areas at an early stage, create stress maps and thus contribute to the protection and sustainable management of aquatic ecosystems. The approaches are transferable to areas such as healthcare, agriculture and security and have transformative potential.