US 12,288,237 B2
Online inference and learning for nonsymmetric determinantal point processes
Ryan A. Rossi, San Jose, CA (US); Aravind Reddy Talla, Evanston, IL (US); Zhao Song, San Jose, CA (US); Anup Rao, San Jose, CA (US); Tung Mai, San Jose, CA (US); Nedim Lipka, Campbell, CA (US); Gang Wu, San Jose, CA (US); and Eunyee Koh, Sunnyvale, CA (US)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on May 12, 2022, as Appl. No. 17/743,360.
Prior Publication US 2023/0368265 A1, Nov. 16, 2023
Int. Cl. G06Q 30/0601 (2023.01)
CPC G06Q 30/0631 (2013.01) [G06Q 30/0629 (2013.01); G06Q 30/0643 (2013.01)] 20 Claims
OG exemplary drawing
 
1. One or more computer storage media storing computer-useable instructions that, when used by one or more computing devices, cause the one or more computing devices to perform operations comprising:
receiving a stream representing compatible sets of entities from a pool of entities in a database;
extracting latent low-dimensional representations of the entities of the compatible sets, wherein the latent low-dimensional representations of the entities are low compared to a size of the stream of compatible sets;
generating a compatibility distribution for the compatible sets using the extracted latent low-dimensional representations; and
for each compatible set represented in the stream, and in response to a new entity arriving in the stream, (i) updating the latent low-dimensional representation of the entities in the compatible set, and (ii) updating a compatibility distribution, that represents likelihood of compatibility of different subsets of the entities, to maximize a probability function that quantifies likelihood of compatibility based on a comparison between latent low dimensional representations of the entities in the compatible set and of the entities in the pool.