US 11,922,942 B1
Natural language processing
Emre Can Kilinc, Santa Barbara, CA (US); Thomas Reno, Santa Barbara, CA (US); John Zucchi, Santa Barbara, CA (US); and Joshua Kaplan, Santa Barbara, CA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Jun. 4, 2020, as Appl. No. 16/892,653.
Int. Cl. G10L 15/26 (2006.01); G10L 15/18 (2013.01)
CPC G10L 15/26 (2013.01) [G10L 15/18 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving text data defining a natural language input configured to invoke a first response template;
receiving a response definition defining a response associated with the first response template;
storing the first response template including the response definition in at least one memory;
receiving natural language input data representing a first question;
receiving metadata identifying a first knowledge graph associated with the natural language input data;
determining that the natural language input data corresponds to the first response template in the at least one memory based at least in part on a correspondence between the natural language input data and the natural language input configured to invoke the first response template;
identifying a first entity in the natural language input data;
determining response data for the first entity using the response definition and the first knowledge graph; and
generating natural language output data comprising the response data.