Abstract: Methods and systems of training a neural network include training a feature extractor and a classifier using a first set of training data that includes one or more base cases. The classifier is trained with few-shot adaptation using a second set of training data, smaller than the first set of training data, while keeping parameters of the feature extractor constant.

Abstract: A processor collecting advertisement (ad) events of a user using a mobile device, such as a mobile phone or eyewear, parsing the ad events from the mobile device, and generating an ad receptivity profile on the granularity of a user identification (ID) and an hour of day. In one example, the processor computes the percentage of ad time watched by the individual user, such as on an hourly basis, and by monitoring a click-through rate (CTR) of the respective user as a measure for user ad receptivity. The processor adjusts an ad allocation/ad load on a per user basis according the user level ad receptivity profile, resulting in dynamically providing ads on the mobile device display when a user is active and receptive viewing the ads.

Abstract: A processor having a performance engine tracking user engagement of advertisement (ad) events using a mobile device, such as a mobile phone or eyewear, to generate a user level ad receptivity profile. In one example, the processor tracks both the percentage of ad time watched by the individual user, such as on an hourly basis, and ad engagement such as by monitoring a click-through rate (CTR) of the respective user as a measure for user ad receptivity. The processor downloads data from the performance engine to a server processor, and the server processor adjusts an ad allocation/ad load on a per user basis according the user level ad receptivity profile. The server processor dynamically provides ads on the mobile device display when a user is active and receptive viewing the ads.