US 11,948,594 B2
Automated conversation content items from natural language
Maheen Sohail, Brooklyn, NY (US); Hyunbin Park, Redwood City, CA (US); Ruoni Wang, Sunnyvale, CA (US); and Vincent Charles Cheung, San Carlos, CA (US)
Assigned to Meta Platforms Technologies, LLC, Menlo Park, CA (US)
Filed by Meta Platforms Technologies, LLC, Menlo Park, CA (US)
Filed on Oct. 20, 2022, as Appl. No. 18/048,246.
Application 18/048,246 is a continuation of application No. 16/894,690, filed on Jun. 5, 2020, granted, now 11,508,392.
Prior Publication US 2023/0063052 A1, Mar. 2, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G10L 21/10 (2013.01); G06N 20/00 (2019.01); G10L 15/197 (2013.01); G10L 15/26 (2006.01); H04M 11/10 (2006.01); G06Q 50/00 (2012.01)
CPC G10L 21/10 (2013.01) [G06N 20/00 (2019.01); G10L 15/197 (2013.01); G10L 15/26 (2013.01); H04M 11/10 (2013.01); G06Q 50/01 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for adding one or more content items to a natural language interface, the method comprising:
generating one or more intents for a conversation by applying natural-language processing to A) a first portion of the conversation and to B) one or more boosts that identify one or more phrases associated with a second portion of the conversation;
obtaining the one or more content items by causing a matching between the one or more intents and multiple content items,
wherein causing the matching comprises:
ranking the multiple content items based on levels of correlation between A) an intent and B) a content item; and
selecting one or more of the highest ranked content items to include in an augmentation of the conversation; and
augmenting the conversation with the selected one or more content items.