Conversational AI platforms with closed domain and open domain dialog integration

Patent metadata

Field	Value
Publication number	US-11769495-B2
Application number	US-202218067217-A
Country	US
Kind code	B2
Filing date	Dec 16, 2022
Priority date	Mar 31, 2021
Publication date	Sep 26, 2023
Grant date	Sep 26, 2023

Abstract

In various examples, systems and methods of the present disclosure combine open and closed dialog systems into an intelligent dialog management system. A text query may be processed by a natural language understanding model trained to associate the text query with a domain tag, intent classification, and/or input slots. Using the domain tag, the natural language understanding model may identify information in the text query corresponding to input slots needed for answering the text query. The text query and related information may then be passed to a dialog manager to direct the text query to the proper domain dialog system. Responses retrieved from the domain dialog system may be provided to the user via text output and/or via a text to speech component of the dialog management system.

First claim

What is claimed is:

1. A processor comprising processing circuitry to:
generate a determination of whether at least one user query corresponds to an open domain dialog system or a closed domain dialog system;
based at least on the determination, send first data corresponding to the at least one user query to one of the open domain dialog system or the closed domain dialog system;
receive, based at least on the first data corresponding to the at least one user query, second data indicative of at least one first response from the open domain dialog system or data indicative of at least one second response from the closed domain dialog system; and
generate at least one output based at least on one or more of the at least one first response or the at least one second response.

2. The processor of claim 1 , wherein the processing circuitry is further to access one or more request policies corresponding to the closed domain dialog system, and the first data corresponding to the at least one user query is determined based at least on the one or more request policies.

3. The processor of claim 2 , wherein the one or more request policies are based at least on one or more of:
at least one domain tag associated with the at least one user query,
at least one intent classification associated with the at least one user query, or
at least one input slot associated with the at least one user query.

4. The processor of claim 1 , wherein the processing circuitry is further to:
analyze the at least one user query to determine one or more indications of one or more of:
at least one intent classification associated with the at least one user query,
at least one domain tag associated with the at least one user query,
or at least one input slot associated with the at least one user query, and

the sending the first data corresponding to the at least one user query is to the open domain dialog system when the one or more indications correspond to the open domain dialog system.

5. The processor of claim 4 , wherein the sending the first data corresponding to the at least one user query is to the closed domain dialog system when the one or more indications correspond to the closed domain dialog system.

6. The processor of claim 1 , wherein the output includes at least one of:
one or more displays of text corresponding to the second data, or
one or more presentations of audio corresponding to audio data generated based at least on the second data using one or more text-to-speech (TTS) algorithms.

7. The processor of claim 1 , wherein the sending the first data to the closed domain dialog system includes generating at least one request based at least on the first data, the generating the at least on request including populating one or more input slots corresponding to the at least one request using one or more portions of the at least one user query.

8. The processor of claim 1 , wherein at least a portion of the second data indicates the at least one second response from the closed domain dialog system and the portion of the second data is used to transmit one or more indications of the at least one second response to the open domain dialog system to cause the open domain dialog system to maintain a context of an ongoing conversation corresponding to the at least one user query and the at least one output.

9. The processor of claim 1 , wherein the open domain dialog system is trained to bias responses toward domains associated with the closed domain dialog system or one or more other closed domain dialog systems.

10. The processor of claim 1 , wherein the processor is comprised in at least one of:
a system of an autonomous or semi-autonomous machine;
a system implemented using an edge device;
a system implemented using a robot;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.

11. A system comprising:
one or more processing units to execute operations comprising:
receiving data indicating at least one response from at least one closed domain dialog system to at least one request corresponding to at least one user query;
providing, using the data, at least one open domain dialog system with at least one indication of the at least one response; and
generating at least one output based at least on the data.

12. The system of claim 11 , wherein the operations further include:
accessing one or more request policies corresponding to the at least one closed domain dialog system; and
generating the at least one request based at least on the one or more request policies.

13. The system of claim 12 , where the one or more request policies are based at least on one or more of:
at least one domain tag associated with the at least one user query, at least one intent classification associated with the at least one user query, or
at least one input slot associated with the at least one user query.

14. The system of claim 11 , wherein the at least one open domain dialog system uses the at least one indication to update one or more dialog state trackers maintained using the at least one open domain dialog system.

15. The system of claim 14 , wherein the operations further include:
sending one or more indications of the at least one user query to the at least one open domain dialog system, and the at least one open domain dialog system uses the one or more indications to update the one or more dialog state trackers.

16. The system of claim 11 , wherein the system is comprised in at least one of:
a system of an autonomous or semi-autonomous machine;
a system implemented using an edge device;
a system implemented using a robot;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.

17. A method comprising:
receiving first data indicating at least one first response from at least one open domain dialog system to at least one user query;
generating at least one request for at least one closed domain dialog system using the data indicating the at least one first response;
receiving second data indicating at least one second response, from the at least one closed domain dialog system, responsive to the at least one request; and
generating at least one output based at least on the second data.

18. The method of claim 17 , further comprising updating one or more dialog states using the first data indicating the at least one first response to the at least one user query, wherein the generating the at least one request is based at least on the one or more dialog states updated using the first data.

19. The method of claim 17 , wherein the at least one output corresponds to at least one of the first data indicating the at least one response or the second data indicating the at least one second response.

20. The method of claim 17 , wherein the method further includes:
analyzing the at least one user query to determine one or more indications of one or more of:
at least one intent classification associated with the at least one user query, at least one domain tag associated with the at least one user query, or at least one input slot associated with the at least one user query;
determining the one or more indications do not correspond to a specific dialog system; and

based at least on the determining, transmitting the at least o

Assignees

• Nvidia Corp

Inventors

• Das Shubhadeep
• Bhattacharya Sumit
• Kumar Ratin

Classifications

• G10L15/1815 Primary

  Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning · CPC title

• G10L13/02

  Methods for producing synthetic speech; Speech synthesisers · CPC title

• G10L15/22

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title

• G10L15/30

  Distributed recognition, e.g. in client-server systems, for mobile phones or network applications · CPC title

• G06F16/3329 Primary

  Natural language query formulation · CPC title