AI4ChemRisk: AI for Chemical Risk Prediction in Aquatic Environments
- Duration: 01.05.2026 - 30.04.2032
Funding Organization: CZS Carl-Zeiss-Stiftung
Prof. Dr. Marius Kloft
University of Kaiserslautern-Landau
AI4ChemRisk analyzes the risks of chemical pollution from wastewater and agriculture in our freshwater ecosystems.
AI models are used to predict contamination processes in order to improve their management.Objectives
The interdisciplinary AI4ChemRisk research team, coordinated by Prof. Kloft, combines expertise from environmental sciences, computer science, and chemical 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 sound assessment of these risks has been virtually impossible 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 detect chemical pollution, realistically simulate missing measurements, and take physical and ecological relationships into account. Deep learning methods for anomaly detection and generative models are being used to create new approaches to close data gaps and predict developments more accurately. 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.
https://rptu.de/en/newsroom/press-releases/detail?tx_news_pi1%5Baction%5D=detail&tx_news_pi1%5Bcontroller%5D=News&tx_news_pi1%5Bnews%5D=18227&cHash=4b472d12d56e7301d4cd7e48d03e4837
AI Focus Areas of the Research Project
Basic Research
- Machine Learning (ML): Self-Supervised Learning, (Semi) Supervised Learning, One-Shot Learning (OSL)/ Few-Shot Learning (FSL)
- Robotics: Sensory Acquisition and Perception
- Technology Analysis