DeepIntegrate: Integrating heterogeneous data sources in deep learning: architectures, algorithms and applications in plant breeding
- Duration: 01.01.2019 - 31.12.2021
Funding Organization: BMBF
Prof. Dr. Marius Kloft
University of Kaiserslautern-LandauCombining the advantages of two methods, data integration and deep machine learning, is the central goal of the DeepIntegrate project. This project will develop, extend and test advanced computational methods that, very similar to human learning, use data from different sources to make decisions and predictions. The project will provide initial conceptual proof in plant breeding, where a variety of data and data sources are used. In DeepIntegrate, predictive models will be developed using image data, genetic data and environmental data in combination with new machine learning methods. This will lead, for example, to an automated evaluation of the performance of plant varieties in plant breeding and thus contribute to the resource-efficient production of food for a growing world population.
https://www.softwaresysteme.dlr-pt.de/media/content/Projektblatt_DeepIntegrate.pdf
AI Focus Areas of the Research Project
Basic Research
- Machine Learning (ML): Anomaly Detection, Density Estimation, Feature Engineering/Feature Extraction
- Robotics: Sensory Acquisition and Perception
- Technology Analysis: Economic Effects