Publication
On the Importance of Domain Knowledge for Real-Time Event Detection
Janina Schneider; Daniel Lukats; Elmar Berghöfer; Iring Paulenz; Lars Nolle; Frederic Theodor Stahl; Jochen Wollschläger
In: OCEANS 2024 - Singapore. OCEANS MTS/IEEE Conference (OCEANS-2024), April 15-18, Singapore, Singapore, Pages 1-6, IEEE Xplore, Piscataway Township, New Jersey, 11/2024.
Abstract
Due to the complexity and dynamics of marine ecosystems, real-time monitoring and possibilities for anticipatory intervention are required. This study emphasises the crucial role of domain (expert) knowledge in developing effective real-time event detection systems. Focusing on the Change Event based Sensor Sampling (ChESS) project, which utilises methods based on Artificial Intelligence (AI) for the analysis of marine data streams, the research integrates stakeholder input and domain knowledge into the development process. Using data from the Time-Series Station (TSS) Spiekeroog, domain knowl-edge influences the design of software architecture, algorithms’ goals and constraints, data retrieval, and related actions as an automated triggering of a water sampler. Lessons learned stress the importance of early domain expert involvement and highlight the value of combining technological advancements with domain insights. Participatory approaches enhance the development of monitoring systems, supporting informed decision-making for sustainable marine ecosystem management.