Representation of Machine Learning Models to Enhance Simulation Capabilities Within Digital Twins in Personalized Healthcare

Abstract

Healthcare has always been a strategic area where innovative technologies can be applied to increase the effectiveness of services and the quality of patient care. Recent progress has been made in the adoption of machine-learning models within digital twins and knowledge graphs. Nevertheless, their deployment needs to address the complex nature of the framework itself, which entails numerous technical, organizational, legal, and ethical challenges. In this paper, we propose an evolution of the CONNECTED conceptual framework, a multi-layered system in which heterogeneous data sources are integrated, standardized, and used to realize digital twins supported by knowledge graphs accessible through dedicated APIs. The extension involves the integration of machine learning models into digital twins, thereby enabling simulation capabilities. The inclusion of a formal and machine-readable self-description with these models serves as a foundation for semantic reasoning. This pivotal feature empowers our architecture with the capability for automatic indexing, aggregation, and querying of the models.

Ähnliche Arbeiten