Abstract: Task management techniques based on user context are provided. More particularly, techniques are presented for calculating task attribute values based on user context data. Once task attributes of a user have been determined, the tasks can be prioritized and a suggestion can be made to the user to perform the tasks in the given order. In a first aspect of the invention, a computer-based technique for scheduling at least one task associated with at least one user includes obtaining context associated with the at least one user, and automatically determining a schedule for the at least one user to perform the at least one task based on at least a portion of the obtained context and based on one or more task attributes associated with the at least one task.

Abstract: The invention provides methods and apparatus for establishing a connection between a mobile client and an external computer network such that the client computer network connection provides the illusion of seamless connectivity. In one aspect of the invention data is processed in a wireless communication network and is received from a access point at a gateway to an external computer network. Access points are disposed in the vicinity of the path of the mobile client. In another aspect of the invention data is processed in a wireless communication network and is received at an onboard proxy from an access point. The data is then transmitted to a mobile client.

Abstract: A method (and system) of using a machine learning infrastructure to support the learning requirements of a plurality of goal-oriented applications includes instantiating a machine learning infrastructure that includes a plurality of learning algorithms with an interface to the algorithms, and registering a plurality of goal-oriented applications with the machine learning infrastructure

Abstract: Services associated with one or more patterns of context about users are invoked. Patterns are determined within context corresponding to the one or more users. The patterns in the context are correlated with services. The services are invoked based on current context of the users. Context is data about the environment in which a given user is located, characteristics of a given user, qualities of a given user, or some combination of these. The patterns in the context may correlate or not correlate with services. Learned patterns do correlate with services and may be used to invoke the services when current context of a user indicates the user has performed the pattern, is performing the pattern, or will perform the pattern.

Abstract: This invention provides methods, systems and apparatus for performing fast computation of metric queries. To achieve this, in an example embodiment, the present invention segments metric regions into disjoint primitive atomic shapes. It then represents these primitive atomic shapes and then performs off-line computation of their relevant properties. As a result of the off-line computation, the execution of a query requires a minimal number of on-line calculations resulting in a very fast query. Further optimization occurs via storage of query histories and prioritization of queries with respect to the access frequency of a metric space's primitive atomic shapes.