Abstract: Systems and methods are described for providing content recommendations by maintaining preference values indicative of a user's preference (e.g., degree of like or dislike) for different assets and updating these preference values based on observed events and known relationships between events and content elements. Also disclosed herein are systems and methods for providing content recommendations using polarity scores, which represent how much user preferences for certain content differ from the average user preference for that content.

Abstract: Methods and systems are described for generating media playlists, or selecting a media asset, according to a “shape” selected by a user. Specifically, a user may “shape” the playlist by designating specific sub-categories of media assets that should be presented at selected times in the playlist. The media application then interpolates the sub-categories for a media asset between the selected times such that adjacent media assets have smooth categorical transitions (e.g., feature incremental changes in the range of sub-categories).

Abstract: Methods and systems are described for generating media playlists, or selecting a media asset, according to a “shape” selected by a user. Specifically, a user may “shape” the playlist by designating specific sub-categories of media assets that should be presented at selected times in the playlist. The media application then interpolates the sub-categories for a media asset between the selected times such that adjacent media assets have smooth categorical transitions (e.g., feature incremental changes in the range of sub-categories).

Abstract: Systems and methods are provided for cataloging content metadata from a variety of sources and providing metadata to client devices. A processing device receives inconsistent data records representative of a common content element, with different values for a metadata field descriptive of a common attribute of the content element. The processor assign confidence scores metadata fields from each data record, and use these confidence scores to select the metadata that is transmitted to the client device.

Abstract: Systems and methods are provided for identifying a content domain associated with seed information, and searching metadata based on seed information. A processor receives seed information and accesses a database of tokens, each token associated with a content domain and each content domain associated with two or more tokens. For example, the content domain may be “Music” and the token may include “feat,” “with,” or “duet.” The processor determines whether a portion of the seed information matches a token in the database of tokens, and in response to determining that a portion of the seed information matches a token, the processor identifies the content domain associated with the matching token and determines whether at least some of the seed information, including the matching token, corresponds to some of the metadata in a database of data records.

Abstract: Systems and methods are described for providing content recommendations by maintaining preference values indicative of a user's preference (e.g., degree of like or dislike) for different assets and updating these preference values based on observed events and known relationships between events and content elements. Also disclosed herein are systems and methods for providing content recommendations using polarity scores, which represent how much user preferences for certain content differ from the average user preference for that content.