Collaborative Spatial Artificial Intelligence in RealtimeThe Project – „CoSAIR – Collaborative Spatial Artificial Intelligence in Realtime“ is funded by the Federal Ministry of Education and Research within the program „Research at Universities of Applied Sciences“ in order to create, consolidate and bundle attractive working and research conditions for young scientists at universities of applied sciences in the field of artificial intelligence (KI-Nachwuchs@FH).
Even in AI age, 80% of decisions have a spatial reference. This is clarified by examples like smart cities, autonomous driving, real-time video analyses or personalised marketing. Therefore, in CoSAIR, the basis for an intelligent connection of AI and geoinformation (spatial AI) will be created.
Mobile and autonomous systems, social media posts or business processes provide data with spatial reference and in real time. Research assets of the Jade University of Applied Sciences like land and water vehicles, drones or the research aircraft JADE ONE also generate such data, the intelligent processing of which is challenging. Therefore, an infrastructure will be realised in the project that enables real-time processing (AI-based Complex Event Processing) of this data.
The main goal is a broad usage of the CoSAIR platform at all three campuses of the Jade University of Applied Sciences in research, research-related teaching and transfer. The CoSAIR platform therefore offers spatial AI Infrastructure as-a-Service (on-premises, i.e. provided in the universities computing centre) with flexible GPU allocation, a GPU cluster node for high-performance requirements such as reinforcement learning and tiered storage for very large AI data sets. The already available GIS infrastructure of the Jade University will be integrated into the CoSAIR platform.
In addition, a collaborative environment for spatial AI development will be customised and provided on-premises as part of the project to offer young academics easy access to the coSAIR platform via the web browser with single sign-on. This enables collaborative AI development within Jupyter notebooks incl. chat function, access to spatial AI datasets and worked examples as well as time-travel version control.
As a result, the CoSAIR project will realise a customised AI platform that is unique in this form.