Digitalisation of Natural Complexity to Solve Socially Relevant Ecological Problems
The State Excellence Program MV "Digitization in Research" supports the consortium "DIG-IT! Digitalisation of Natural Complexity to Solve Socially Relevant Ecological Problems" under the direction of Prof. Martin Wilmking. Inbetween July of 2019 and June of 2022, researchers will have 2 million euros at their disposal to develop a methodological toolbox that can independently capture and categorize ecological image and audio data using machine learning techniques (deep convolutional neural networks). Partners are the Fraunhofer Institut für Grafische Datenverarbeitung Rostock (Prof. Uwe von Lukas), Biomathematik (Prof. Mareike Fischer) and working groups of the Institut für Botany und Landscape Ecology (Profs. Joosten, Kreyling, Wilmking) and the Zoological Institute and Museum (Prof. Gerald Kerth).
By exploring the opportunities of digitalization for the ecological sciences, DIG-IT! will meet pressing ecological questions of high societal relevance with a future-oriented arsenal of methods and thereby qualify digitally competent ecologists and ecologically experienced biomathematicians and computer scientists. Dig-It! will address a broad array of questions including (but not limited to) service functions and stability of ecosystems under climate and land use change, species protection and innovative environmental monitoring. The overarching goal is to facilitate a "quantum leap" for the field of ecology through the development of universally applicable methods using self-learning algorithms ("Deep Convolutional Neural Networks"), because in the digital age the challenge no longer lies in the amount of available primary data, but in its evaluation. For this purpose, DIG-IT! will combine the developmental expertise for the automated analysis of image data (Fraunhofer Institute for Computer Graphics, Rostock and Biomathematics University Greifswald) with the application to urgent ecological questions (Botany / Landscape Ecology / Zoology University Greifswald).
Extract from the jury vote:
"The DIG-IT! project aims to address urgent ecological issues through the use of digital technologies, with the evaluation of primary data posing a particular challenge. (...) The jury is convinced of the scientific excellence of the project. The idea of developing an ecological toolbox is particularly impressive. It also appreciates the approach of docking onto already existing excellent research alliances such as WETSCAPES and RESPONSE [DFG Research Training Group "Biological RESPONSEs to Novel and Changing Environments"] and using their findings above all to promote young scientists."