STSM WG3 (Ulrich Schaechtle)

Start Date: 2012-08-12

End Date: 2012-09-03

Host Institution: U App. Sc. West Switzerland

Host Country: Switzerland

Home Institution: Royal Holloway University

Home Country: UK


Widespread diseases are a global problem. For example, 346 million people worldwide have diabetes. WHO projects that diabetes deaths will double between 2005 and 2030. These numbers show, how important it is to develop personal health systems with devices that support artificially intelligent software to monitor widespread diseases like diabetes. As such software could provide medical diagnosis and doctoral advice, it has the potential to help doctors and patients alike.
Computational agents are autonomous reasoning entities that can serve as the artificially intelligent software in personal health devices. Such agents should be able to monitor the disease for a patient, diagnose problems and offer the required advice. Therefore, it is crucial how to specify and implement medical knowledge for these agents to use when dealing with a patient. We envisage to use computational logic to represent medical knowledge in a useful way. Furthermore, the knowledge representation for the agent should be able to incorporate and interpret real world biomedical and physiological data.
The research will focus on the application of a specialised version of the KGP agent model on a specific diabetic domain based on existing medical guidelines and knowledge from doctors. We will study the required knowledge representation to perform monitoring and model-based diagnosis in this domain, we will implement the required knowledge to support medical decision-making for doctors and we will study how to relate the approach to existing patient data. This research will be empirically evaluated in heath-care scenarios for diabetes type 1 and type 2 patients.