Towards context-aware mobile services through the use of Hierarchical Temporal Memory

With the emergence of pervasive computing and the Internet of Things, the numerous sensors of the intelligent surrounding environment will provide more and more context parameters at any time. These parameters will identify the context of an entity (user, object, etc.) more precisely than ever, but in most cases there will be only a few parameters relevant for a given situation or service. The goal of this paper is to prove that the Hierarchical Temporal Memory framework provides an efficient and scalable solution to detect these relevant parameters, if a large enough training set is provided. This opens the way to the adaptive, context-aware services that will emerge in the coming years. As a proof of concept example we analyze here the case of a mobile application that monitors different environmental parameters and learns the situations in which the owner of the mobile phone decides to upload to the Internet a photo immediately after taking it.