US 12,135,739 B2
Generating conversation topics using neural networks
Fergal Reid, Dublin (IE); Cathal Horan, Dublin (IE); and Mario Kostelac, Zagreb (HR)
Assigned to Intercom, Inc., San Francisco, CA (US)
Filed by Intercom, Inc., San Francisco, CA (US)
Filed on Feb. 22, 2023, as Appl. No. 18/112,967.
Claims priority of provisional application 63/313,065, filed on Feb. 23, 2022.
Prior Publication US 2023/0281235 A1, Sep. 7, 2023
Int. Cl. G06F 16/00 (2019.01); G06F 16/35 (2019.01)
CPC G06F 16/353 (2019.01) [G06F 16/355 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A method comprising:
providing, by a processing device, a plurality of conversations to a neural network to generate a plurality of clusters;
selecting, for each cluster of the plurality of clusters, a topic and one or more keywords from one or more n-grams;
evaluating, for each cluster of the plurality of clusters, the topic and the one or more keywords by searching historical conversations and current conversations to identify one or more conversations related to the cluster;
determining, based on one or more false positives (FPs), that a first cluster of the plurality of clusters overlaps a second cluster of the plurality of clusters, wherein a FP is a conversation that matches a query without being a part of the first cluster;
executing a remedial action to prevent processing a duplicate cluster responsive to determining that the first cluster of the plurality of clusters overlaps the second cluster of the plurality of clusters by generating an updated set of keywords for the first cluster of the plurality of clusters and augmenting a search query associated with the first cluster of the plurality of clusters;
storing topic memberships by linking the plurality of conversations and the updated set of keywords; and
configuring a real-time query data store to answer topic queries associated with the topic memberships.