Semantic Perception: Converting Sensory Observations to Abstractions

摘要: An abstraction is a representation of an environment derived from sensor observation data. Generating requires inferring explanations incomplete set observations (often the Web) and updating these on basis new information. This process must be fast efficient. The authors' approach overcomes challenges to systematically derive abstractions observations. models perception through integration abductive logic framework called Parsimonious Covering Theory with Semantic Web technologies. authors demonstrate this approach's utility scalability use cases in healthcare weather domains.