US 12,086,211 B2
Granular cluster generation for real-time processing
Xuerui Wang, Mountain View, CA (US); Daniel Li, Mountain View, CA (US); Xiaodan Song, Cupertino, CA (US); Jie Han, Mountain View, CA (US); and Rahul Sharma, Mountain View, CA (US)
Assigned to GOOGLE LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Feb. 10, 2023, as Appl. No. 18/167,297.
Application 18/167,297 is a continuation of application No. 16/670,809, filed on Oct. 31, 2019, granted, now 11,580,170.
Claims priority of provisional application 62/754,372, filed on Nov. 1, 2018.
Prior Publication US 2023/0267176 A1, Aug. 24, 2023
Int. Cl. G06F 18/23211 (2023.01); G06F 16/215 (2019.01); G06F 16/906 (2019.01); G06F 18/23213 (2023.01)
CPC G06F 18/23211 (2023.01) [G06F 16/215 (2019.01); G06F 16/906 (2019.01); G06F 18/23213 (2023.01)] 18 Claims
OG exemplary drawing
 
1. A system comprising:
a data processing system comprising one or more processors configured to:
access content associated with a third-party content provider;
retrieve, based on the accessed content, tokens that represent an intent of the third-party content provider;
identify, based on the retrieved tokens, a segment from a plurality of segments for recommending to the third-party content provider, wherein the identified segment includes a token cluster,
wherein the token cluster is generated by removing, using a filtering technique, a token from a plurality of tokens associated with the identified segment, the filtering technique includes a query hits metric technique, and
wherein the plurality of segments is generated by using a clustering model that clusters input data into the plurality of segments, the clustering model being generated using unsupervised machine learning; and
recommend a token in the identified segment to the third-party content provider.