US 12,394,411 B2
Domain specific neural sentence generator for multi-domain virtual assistants
Pranav Singh, Sunnyvale, CA (US); Yilun Zhang, Toronto (CA); Eunjee Na, Busan (KR); and Olivia Bettaglio, Santa Clara, CA (US)
Assigned to SoundHound AI IP, LLC., Santa Clara, CA (US)
Filed by SoundHound, Inc., Santa Clara, CA (US)
Filed on Oct. 27, 2022, as Appl. No. 18/050,182.
Prior Publication US 2024/0144921 A1, May 2, 2024
Int. Cl. G10L 15/18 (2013.01); G10L 15/06 (2013.01); G10L 15/22 (2006.01)
CPC G10L 15/1815 (2013.01) [G10L 15/063 (2013.01); G10L 15/1822 (2013.01); G10L 15/22 (2013.01); G10L 2015/0631 (2013.01); G10L 2015/223 (2013.01)] 23 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
obtaining training data comprising query data samples, the query data samples including pairs of text data representing queries and responses and corresponding keywords specific to a domain;
training a domain-specific sentence generation model using the pairs of text data and the corresponding keywords;
receiving an initial query that includes a sentence;
extracting one or more keywords from the received sentence;
determining the one or more keywords are associated with a specific intent;
generating, via the trained domain-specific sentence generation model, a plurality of generated sentences based on the specific intent;
applying a classifier model to the generated plurality of sentences to map sentences to the specific intent and to determine one or more sentences with a high probability of having the same specific intent;
selecting the one or more sentences that are associated with the specific intent, wherein each selected sentence comprise one or more placeholders representing a specific type of word;
receiving a new client query corresponding to one of the one or more selected sentences; and
invoking the specific intent corresponding to the one of the one or more selected sentences.