FutureArctic aims to pave the way for generalized permanently connected data acquisition systems for key environmental variables and processes. We will initiate a new machine-learning approach to analyse large high-throughput environmental data-streams, through installing a pioneer "ecosystem-of-things" at the ForHot site.
FutureArctic will thus channel, building on a timely project in the ForHot area, an important evolution to machine-assisted environmental fundamental research. This is achieved through the dedicated training of researchers with profiles at the intersectoral edge of computer science, artificial intelligence, environmental science (both experimental and modelling), social sciences and sensor engineering and communication.

One of the 15 PhDs trained in this project will be working in the field of Science and Technology Studies and will specifically look how these new data intensive methods will change the way environmental research will be performed. It will address the question: What does it mean to abandon the classical order of designing a hypothesis and test it experimentally, and work with correlations to make predictions based on huge datasets? How does the use of algorithms impact questions posed, problems identified and answers produced? Close empirical scrutiny of the challenges of applying machine learning in environmental research projects is key to understand the nuances of change in research practices, as is close collaboration with the respective domain researchers. 

Methodological approach: qualitative social science methods such as interviews, participant observation and group discussions.

Cooperating Partners: 

At the University of Vienna: Andreas Richter, Christina Kaiser

For the other partners in the training network see: http://www.futurearctic.be/supervisors/