US 12,462,166 B2
Machine learning approaches for interface feature rollout across time zones or geographic regions
Michele Saad, Austin, TX (US); and Lauren Dest, Austin, TX (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Oct. 13, 2021, as Appl. No. 17/500,785.
Prior Publication US 2023/0115855 A1, Apr. 13, 2023
Int. Cl. G06N 5/022 (2023.01); G06N 5/04 (2023.01); G06F 3/04847 (2022.01)
CPC G06N 5/022 (2013.01) [G06N 5/04 (2013.01); G06F 3/04847 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
one or more memory devices comprising a feature visualization machine learning model and data corresponding to a set of digital content items; and
one or more computing devices that are configured to cause the system to:
generate, for display within one or more graphical user interfaces of a plurality of client devices located in a sample time zone, a graphical visualization of the set of digital content items in a first arrangement;
determine, for the sample time zone, client device interactions in relation to digital content items from among the set of digital content items in the first arrangement within the one or more graphical user interfaces;
determine a similarity score between the sample time zone and a target time zone according to historical network user behavior within the sample time zone and within the target time zone;
based on the client device interactions in relation to the digital content items in the first arrangement and according to the similarity score between the sample time zone and the target time zone, generate, for the target time zone and utilizing the feature visualization machine learning model, a target time zone graphical visualization of the set of digital content items in a second arrangement by relocating, removing, or rearranging one or more digital content items from among the set of digital content items by utilizing the feature visualization machine learning model to:
generate a latent time zone vector representing client device interactions from the sample time zone;
generate weights and biases associated with neurons of the feature visualization machine learning model according to a target performance metric; and
generate, from the latent time zone vector utilizing the weights and biases, the target time zone graphical visualization depicting visual modifications to the set of digital content items, wherein the weights and biases of the feature visualization machine learning model are modified by:
accessing or retrieving a ground truth graphical visualization corresponding to sample client device interactions; and
comparing the ground truth graphical visualization with the target time zone graphical visualization; and
provide, for display within a graphical user interface of a client device located within the target time zone, the target time zone graphical visualization of the one or more digital content items in the second arrangement.