Machine learning based data fusion algorithms development

University of Turku

Environmental studies involve combining diverse data sources like national topographic maps, satellite imagery, and real-time sensor networks. Each offers unique insights, but they have their limitations. Satellite images might lack detail, while sensor data might be sparse. Combining them allows of creating a richer picture. Machine learning can address this by fusing different data sources. Especially deep learning techniques have a high potential to combine data for the given objectives. This understanding of the combined data is crucial for creating accurate digital representations of the environment. Ultimately, these algorithms will inform better decision-making in areas like water resource management.

We seek a talented PhD researcher with expertise in machine learning and spatial environmental data analysis. The PhD project focuses on developing novel machine learning algorithms especially based on deep learning methodologies to fuse diverse environmental data sources, such as national topographic data (elevation and landscapes), high-resolution satellite imagery, and real-time sensor network data capturing dynamic changes. By learning from vast datasets, the algorithmical goal is to identify patterns and relationships between these data types with the given objectives to create a comprehensive view crucial for building accurate digital representations of the environment, supporting objectives in hydrology, atmosphere, and cryosphere (water, air, and frozen ground systems). The successful candidate will develop, test, and refine these machine learning algorithms.

The position requires an active approach, diligence and cooperation skills, willingness to work as a part of a team and a passion to learn new. We also expect an excellent master’s degree in, for example, computer science, information technology, mathematics or related fields. The candidate should possess a strong understanding of machine learning algorithms and techniques, particularly those relevant to data fusion. Proficiency in the Python programming language is also required. Experience with environmental and GIS data analysis is a valuable asset.

Luke

Department of computing/Faculty of Technology

Prof. Jukka Heikkonen, UTU

Prof. Tomi Westerlund, UTU

Dr. Petra Virjonen, UTU

??