Large-Scale Scientific Information Systems Research Group

 
home publications projects openings for students team contact & imprint

AgriCube

Scientific goal of the collaboration is to advance datacube concepts and technology for large-scale use on Earth data, from satellite to drones to climate data. Particular emphasis is on datacube analytics on time-series and on combining Machine Learning (ML) with datacubes, on the basis of Linear Algebra (Tensor Algebra) as the common ground for both disciplines. Such a combination has a high potential for a quantum leap in Earth data ex¬ploitation – ML adds intelligence, datacubes add scalability. Re¬quire¬ments and evaluation will be done through an agri¬cult¬ural decision making scenario.

This work will be conducted by two young researchers from research groups which both have a high international reputation in their science field as well as in the field of international standardization. The main scientific gals are:

  • Breakthrough: Determine how to perform Machine Learning using Array DBMSs, establish a demonstrator showing feasibility in an agricultural decision making scenario (crop yield determination in Taiwan).
  • 4 joint scientific publications: Collaborative work in every exchange visit and the period after, until the next quarterly meeting, will be consolidated into one paper every 6 months submitted to an international conference or journal.
  • Deepened collaboration: Establish strong, day-to-day proven collaboration between Constructor University and Feng-Chia University, preparing for a full project (which can include SMEs from Germany and Taiwan).
  • Sustainability: A joint follow-up research project proposal has been submitted. The Taiwanese datacube is member of the EarthServer datacube federation.

Project Partners

  • Constructor University, DE
  • Feng-Chia University, Taiwan

This project is supported by DAAD under project no. 57561009.