Patent attributes
Disclosed embodiments pertain to filtering application notifications from one or more applications on an electronic device. A request can be received from an application installed on the electronic device to present a notification. In response, a machine learning model is invoked that is trained to generate a relevancy score for a notification based on historical behavior of the user with respect to previous notifications from the application as well as the active context of the user. The relevancy score can be compared with a threshold to determine whether the threshold is satisfied. Notification filtering can be performed based on whether or not the threshold is satisfied. The method can filter out the notification from presentation if the threshold is satisfied and present the notification if the threshold is unsatisfied, or vice versa.
