US 10,504,512 B1Granted

Natural language speech processing application selection

AssigneeAmazon Technologies, Inc.

InventorsRuhi Sarikaya, Rohit Prasad, Kerry Hammil, Spyridon Matsoukas, Nikko Strom, Frédéric Johan Georges Deramat, Stephen Frederick Potter, Young-Bum Kim

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

Techniques for limiting natural language processing performed on input data are described. A system receives input data from a device. The input data corresponds to a command to be executed by the system. The system determines applications likely configured to execute the command. The system performs named entity recognition and intent classification with respect to only the applications likely configured to execute the command.

Claims20

Dependent claims

A system comprising:

at least one processor;
and
at least one memory including instructions that, when executed by the at least one processor, cause the system to:
receive input text data associated with a source device;
determine a first confidence score representing a first likelihood that a first application is configured to execute a command represented in the input text data;
determine a second confidence score representing a second likelihood that a second application is configured to execute the command;
determine a third confidence score representing a third likelihood that a third application is configured to execute the command;
perform natural language understanding (NLU) processing on the input text data using at least one first NLU component associated with the first application to generate first NLU data corresponding to at least one first NLU score;
perform NLU processing on the input text data using at least one second NLU component associated with the second application to generate second NLU data corresponding to at least one second NLU score;
select the first NLU data based at least in part on the at least one first NLU score and the at least one second NLU score;
and
cause, based at least in part on the first NLU data, the first application to execute the command.

9 more elements

A computer-implemented method comprising:

receiving input text data associated with a source device;
determining a first confidence score representing a first likelihood that a first application is configured to execute a command represented in the input text data;
determining a second confidence score representing a second likelihood that a second application is configured to execute the command;
determining a third confidence score representing a third likelihood that a third application is configured to execute the command;
performing natural language understanding (NLU) processing on the input text data using at least one first NLU component associated with the first application to generate first NLU data corresponding to at least one first NLU score;
performing NLU processing on the input text data using at least one second NLU component associated with the second application to generate second NLU data corresponding to at least one second NLU score;
selecting the first NLU data based at least in part on the at least one first NLU score and the at least one second NLU score;
and
causing, based at least in part on the first NLU data, the first application to execute the command.

6 more elements

Dependencies are inferred from the claim text.