Methods and apparatus for determining a domain based on the content and context of a natural language utterance

• US 9,734,825 B2
• Filed: 05/15/2014
• Issued: 08/15/2017
• Est. Priority Date: 06/03/2002
• Status: Expired due to Term

• Current Assignee: Dialect, LLC
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Kennewick, Robert A., Locke, David, Kennewick, Sr., Michael R., Kennewick, Jr., Michael R., Freeman, Tom, Kennewick, Richard
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US14/278,627
• Publication Number:
  US 20140249821A1
• Time in Patent Office: 1,188 Days
• Field of Search: 704231, 704257, 704270, 7042701, 704275, 704240, 707755
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

Systems and methods for receiving user generated natural language utterances such as queries and/or commands and executing the queries and/or commands. A parser may be used to determine a domain for a user generated natural language utterance based on the content and context of the natural language utterance. One or more of a plurality of domain agents associated with particular domains are configured to receive and respond to queries and/or commands based on the natural language utterance.

Claims

1. A system responsive to a user generated natural language speech utterance, comprising:

a plurality of autonomous executable domain agents, each of which is configured to respond to queries and/or commands within a particular domain, wherein the particular domain indicates an area of expertise within which the domain agent is configured to respond to the queries and/or commands;

a speech recognition engine configured to recognize one or more words in the user generated natural language speech utterance;

a parser configured to;

receive from a system agent or an active domain agent of the plurality of autonomous executable domain agents, keyword and associated prior probabilities or fuzzy possibilities;

determine, for the natural language speech utterance, a score for each of at least two possible contexts, wherein the scores are determined based on the received keyword and associated prior probabilities or fuzzy possibilities;

determine a domain for the user generated natural language utterance based on the recognized one or more words of the natural language utterance and the determined scores for each of the at least two possible contexts;

select at least one of the plurality of domain agents based, at least in part, on the determined domain; and

provide at least one query and/or command based on the natural language utterance to the selected at least one of the plurality of domain agents,

wherein each of the selected at least one of the plurality of domain agents is configured to create one or more queries based on the at least one query and/or command and send the one or more queries in an asynchronous manner to one or more local or external information sources.

2. The system of claim 1, wherein the parser is further configured to generate the at least one query and/or command based on the natural language utterance.

3. The system of claim 2, wherein

the at least one query and/or command based on the natural language utterance comprises a plurality of queries and/or commands.

4. The system of claim 1, wherein one or more of the selected at least one of the plurality of domain agents is further configured to create duplicate queries and send the duplicate queries in an asynchronous manner to different information sources.

5. A method for responding to a user generated natural language speech utterance, the method comprising:

recognizing, by a speech recognition engine, one or more words in the user generated natural language speech utterance;

receiving, at a parser, keyword and associated prior probabilities or fuzzy possibilities from a system agent or an active domain agent of a plurality of autonomous executable domain agents;

determining, for the natural language speech utterance, a score for each of at least two possible contexts, wherein the scores are determined based on the received keyword and associated prior probabilities or fuzzy possibilities;

determining by the parser, a domain for the user generated natural language utterance based on the recognized one or more words of the natural language utterance and the determined scores for each of the at least two possible contexts;

selecting at least one of the plurality of autonomous executable domain agents based, at least in part, on the determined domain, wherein each of the plurality of domain agents is configured to respond to queries and/or commands within a particular domain, wherein the particular domain indicates an area of expertise within which the domain agent is capable of responding to the queries and/or commands;

providing at least one query and/or command based on the natural language utterance to the selected at least one of the plurality of domain agents;

creating, by the selected at least one of the plurality of domain agents, one or more queries based on the at least one query and/or command;

sending, by the selected at least one of the plurality of domain agents, the one or more queries in an asynchronous manner to one or more local or external information sources.

6. The method of claim 5, further comprising generating, by the parser, the at least one query and/or command based on the natural language utterance.

7. The method of claim 6, wherein the at least one query and/or command based on the natural language utterance comprises a plurality of queries and/or commands.

8. The method of claim 5, wherein one or more of the selected at least one of the plurality of domain agents is further configured to create duplicate queries and send the duplicate queries in an asynchronous manner to different information sources.