Multimodal Enriched Knowledge Graphs for Organizing and Understanding Complex Data in Dataspaces

What is it about?

Data spaces enable the management and analysis of ever-increasing and complex data from diverse domains like smart cities, healthcare, e-commerce, and social media. Multimodal knowledge graphs integrate data produced by different sources, providing a comprehensive view of data from diverse sources. Leveraging common sense knowledge, data spaces can better comprehend the patterns in data for more intuitive decision-making. For example, during a traffic accident, structured data from traffic sensors and GPS indicate congestion, while unstructured data from cameras provides details about vehicles and casualties involved. Integrating this data into a unified representation and enriching it with external knowledge, such as traffic patterns or nearby hospitals, can significantly improve response decisions like rerouting traffic, dispatching emergency services, or alerting hospitals for potential patients.

Featured Image

Photo by ev on Unsplash

Photo by ev on Unsplash

Why is it important?

Data is at the heart of our society. As our world is increasingly becoming interconnected and data-driven, there is a pressing need for better data management solutions to handle complex, diverse data sources. Data spaces involving multimodal knowledge graphs with enrichment have the potential to significantly impact the way complex heterogeneous data is managed and utilized, ultimately enhancing the quality of life for citizens. Our paper establishes a foundation for this research direction, identifying major challenges and requirements for developing multimodal knowledge graphs in data spaces and leveraging common sense knowledge from rich knowledge bases. Addressing these challenges in developing multimodal knowledge graphs and incorporating external knowledge within data spaces will pave the way for more efficient and effective decision-making across various applications. We present a use case of data management in smart cities and propose an ontology for constructing multimodal knowledge graphs in city data spaces.