US 12,299,027 B2
Emoji recommendation system
Cameron Alston, Seattle, WA (US); Tristan Cossio, San Francisco, CA (US); Shuyu Luo, Belmont, CA (US); Arin Schwartz, San Francisco, CA (US); and Ken Soong, Redwood City, CA (US)
Assigned to DISCORD INC., San Francisco, CA (US)
Filed by DISCORD INC., San Francisco, CA (US)
Filed on Mar. 28, 2023, as Appl. No. 18/191,806.
Prior Publication US 2024/0330354 A1, Oct. 3, 2024
Int. Cl. G06F 16/00 (2019.01); G06F 16/2457 (2019.01); G06F 16/438 (2019.01)
CPC G06F 16/438 (2019.01) [G06F 16/24578 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
creating sets of community emoji embeddings in respective embedding spaces, the sets of community emoji embeddings comprising respective vectors representing respective past emoji usage by respective users in respective online communities of a community hosting service;
creating different sets of user embeddings for a same user in the respective online communities based on transforming the respective set of community emoji embeddings into the respective set of user embeddings by shifting the respective vectors of the respective set of community emoji embeddings based on respective past emoji usage by the same user in the respective online communities, wherein the different sets of user embeddings represent the respective different past emoji usage by the same user and different past emoji usage in the respective online communities;
measuring similarity scores between the respective sets of community emoji embeddings and the different sets of user embeddings; and
generating different rankings of top emojis based on the measured similarity scores for the same user for the respective online communities.