Workshopbeitrag
Heterogeneous Data in Health Applications: An Algorithmic Approach Leveraging Continual Learning
Lade...
Volltext URI
Dokumententyp
Text/Workshop Paper
Zusatzinformation
Datum
2024
Autor:innen
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Gesellschaft für Informatik e.V.
Zusammenfassung
Over the past decades, data-driven AI-based methods have been widely used in human activity recognition, leading to the successful fostering of healthy behavior through health applications. Apart from advanced neural network model algorithms, researchers also focused on novel and reliable sensing modalities to monitor more complex activities, resulting in even more diverse and heterogeneous data over time. Inference and drawing implications from those heterogeneous data stream is still a great challenge. Besides, the heterogeneous data stream's diverse dimension is another issue in making one neural model for continual learning. In this abstract, we present a recent algorithmic solution based on the novel Kolmogorov Arnold Network to address these issues simultaneously, leading to the efficient utilization of heterogeneous data in health applications.