Towards Timely and Efficient Semantic Reasoning for the Networked Society

This paper presents our work in progress on enabling computerized reasoning capability in machine-to-machine communication scenarios for the Networked Society (or Internet of Things). Such reasoning capability is about drawing high-level conclusions on the situation in real time based on raw data streams generated by various sources. There are challenges posed by the dynamic and heterogenous availability of raw data coming from different sources, as well as the stringent time constraints for conclusions to be made. Our goal hence is to make machine-based reasoning processes time-efficient, resource-efficient, and scalable. We present an approach that addresses the challenges by decomposing a reasoning process into two stages: “shallow reasoning” and “deep reasoning”. The former deals with the dynamic and heterogenous availability of raw data from different sources, while the latter executes semantic reasoning with a lightweight workload that has been reduced by the former. We present our prototype implementation of a reasoning system that adopts the proposed approach in a proactive healthcare use case. Performance evaluation is currently ongoing to verify the effectiveness of our approach.