US 11,657,428 B1
Enhanced goal-based audience selection
Graham Reid Scarth Ritchie, Burntisland (GB); Pawel Pomorski, Edinburgh (GB); Zhun Zhang, Edinburgh (GB); and Ravi Bhagavan, Edinburgh (GB)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Dec. 8, 2020, as Appl. No. 17/115,392.
Int. Cl. G06Q 30/02 (2012.01); G06Q 30/0251 (2023.01); G06N 20/00 (2019.01)
CPC G06Q 30/0255 (2013.01) [G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method for generating an audience to which to present an advertisement campaign based on an objective of the advertisement campaign, the method comprising:
receiving, by a device, a request to generate a target audience for an advertisement campaign, wherein the request comprises an objective associated with presentation of the advertisement campaign, and wherein the target audience and audience criteria defining the target audience are absent from the request;
generating, by the device, using a machine learning model configured to generate, responsive to the request from which the target audience and the audience criteria defining the target audience are absent, based on the machine learning model being trained with user preference data to model a sequence of user actions of a system, predictions of which of the user actions are most likely to result in satisfaction of the objective;
identifying, by the device, using the machine learning model, first user actions of a system predicted by the machine learning model to result in satisfaction of the objective;
identifying, by the device, using the machine learning model, first users of the system who performed the first user actions;
determining, by the device, using the machine learning model, second user actions performed by the first users prior to performing the first user actions, the second user actions having probabilities exceeding a threshold indicating a likelihood that the second user actions will result in subsequent performance of the first user actions;
determining, by the device, using the machine learning model, third user actions performed by the first users prior to performing the first user actions, the third user actions having probabilities below the threshold;
identifying, by the device, using the machine learning model, based on the second user actions having probabilities exceeding the threshold, second users of the system who performed the second user actions and failed to perform the first user actions;
generating, by the device, as an output of the machine learning model, the target audience to which to present the advertisement campaign, the target audience comprising the second users and excluding the first users;
causing presentation, by the device, of the advertisement campaign to the target audience;
receiving, by the device, in response to the presentation of the advertisement campaign to the target audience, data indicative of a performance of the advertisement campaign, wherein data indicative of a performance of the advertisement campaign comprises an action of viewing at least one advertisement of the advertising campaign; and
updating, by the device, the machine learning model based on the action viewing the advertisement campaign, the data indicative of the performance of the advertisement campaign, the predictions of first user actions that are most likely to result in satisfaction of the objective.