Abstract: Personalization with on-device privacy preservation is provided by receiving, at a user device, a generalized recommendation from a remote provider backend, which include a plurality of candidate items responsive to a request from an application running on the user device; generating a plurality of candidate vectors corresponding to the plurality of candidate items representing a corresponding candidate item in a binary format across a plurality of metrics associated with the plurality of candidate items; generating a user vector based on personal data stored on the user device, the user vector representing a user in a binary format across one or more metrics associated with the personal data; determining interaction propensities based on the user vector and the plurality of candidate vectors according to a global model run on the user device; and displaying, on the user device, the plurality of candidate items according to the interaction propensities.

Abstract: Personalization with on-device privacy preservation is provided by receiving, at a user device, a generalized recommendation from a remote provider backend, which include a plurality of candidate items responsive to a request from an application running on the user device; generating a plurality of candidate vectors corresponding to the plurality of candidate items representing a corresponding candidate item in a binary format across a plurality of metrics associated with the plurality of candidate items; generating a user vector based on personal data stored on the user device, the user vector representing a user in a binary format across one or more metrics associated with the personal data; determining interaction propensities based on the user vector and the plurality of candidate vectors according to a global model run on the user device; and displaying, on the user device, the plurality of candidate items according to the interaction propensities.

Abstract: An extensible method and system for dynamically transforming and routing the WWW content is provided that can enable efficient delivery and presentation of the content for a variety of end-user devices. Processing of HTTP request and response messages is combined with event-driven machinery that can perform transformations and routing of data or content. The machinery can include a logic-based inference engine adapted to determine which transformations are applicable in any given context. A knowledge base for the content adaptation and routing may also be provided. The knowledge base is segmented so that search space for the inference engine can be significantly reduced in any given context. Since the search space is much narrower, searching can be performed much more efficiently, which can result in improved performance.