| CPC G06F 16/9536 (2019.01) | 20 Claims |
|
1. A computer-implemented method comprising:
receiving a request to generate collaboration recommendations for an artist, the request including a specific collaboration objective, the specific collaboration objective comprising at least one of growing an audience of the artist or engaging the audience of the artist;
processing, by one or more machine learning models of a collaboration system, artist data for the artist with additional artist data for other artists in an artist population in order to identify similar collaborators with audiences which overlap the audience of the artist and to identify dissimilar collaborators with audiences which do not overlap the audience of the artist;
filtering the other artists in the artist population to a subset of one or more artists satisfying the specific collaboration objective based on contextual attributes included in the artist data for the artist and additional artist data for other artists;
generating, by one or more machine learning models of the collaboration system, the collaboration recommendations for the artist to include the subset of one or more artists, the one or more collaboration recommendations generated based on the specific collaboration objective to grow the audience of the artist by recommending collaborations with dissimilar collaborators, wherein the one or more machine learning models leverage a social graph that is generated by the collaboration system from the artist data, the social graph including artist nodes that represent artists and content consumer nodes that represent content consumers, the social graph including edges that connect the content consumer nodes to one or more of the artist nodes that the content consumers have listened to or interacted with, and wherein the edges are weighted based on one or more of a number of listens, amount of listening, or amount of digital interaction;
exposing, via a user interface of the collaboration system, the collaboration recommendations, the collaboration recommendations recommending at least one of the other artists that satisfy the collaboration objective as a collaborator for the artist, wherein the one or more machine learning models are trained using training data that includes historical collaboration matches between artists, and wherein the one or more machine learning models are trained to optimize for the specific collaboration objective by generating outputs corresponding to collaboration matches and rewarding acceptable collaboration matches by adjusting internal weights of the one or more machine learning models to encourage similar acceptable outputs and discouraging unacceptable collaboration matches by adjusting internal weights of the one or more machine learning models to discourage similar unacceptable outputs;
receiving, via the user interface, a selection of the collaborator;
forming, by the collaboration system, a communication channel between the artist and the collaborator; and
communicating, via the communication channel, an electronic message from the artist to the collaborator.
|