SYSTEMS AND METHODS FOR E-COMMERCE API ORCHESTRATION USING NATURAL LANGUAGE INTERFACES

• US 20200135180A1
• Filed: 10/31/2018
• Published: 04/30/2020
• Est. Priority Date: 10/31/2018
• Status: Active Grant

• Current Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Original Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Inventors: Mukherjee, Snehasish, Subramanya, Shankara Bhargava
• Granted Patent: US 11,238,850 B2
• Time in Patent Office: Days
• CPC Class Codes: G06Q 30/0641 Shopping interfaces
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• 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 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: G10L 2015/228 of application context

Abstract

Systems and methods for e-commerce systems using natural language understanding are described. A computing device is configured receive a user utterance and identify at least one semantic element within the user utterance. An intent associated with the at least one semantic element is identified and an intent flow associated with the identified intent is executed. The intent flow includes a set of tasks executed in a predetermined order. A system utterance is generated by instantiating a response template selected from a plurality of response templates associated with the executed intent.

Claims

1. A system, comprising:

a computing device configured to;

receive a user utterance and identify at least one semantic element within the first user utterance;

identify an intent associated with the at least one semantic element;

execute an intent flow associated with the identified intent, wherein the intent flow includes a set of tasks executed in a predetermined order; and

generate a system utterance by instantiating a response template selected from a plurality of response templates associated with the executed intent.

2. The system of claim 1, wherein the computing device is configured:

identify a missing semantic element required by the intent flow associated with the identified intent;

review at least one prior utterance to locate the missing semantic element; and

execute the intent flow associated with the identified intent including the at least one semantic element from the user utterance and the missing semantic element identified from the at least one prior utterance.

3. They system of claim 2, wherein the missing semantic element is identified by a context mining component configured to:

maintain a context stack including a set of context entries, wherein each of the context entries comprises a root intent element, an entity list element, and a dialogue stack element;

review at least one context entry in the set of context entries to locate the missing semantic element within the dialogue stack element; and

generate an intent flow execution including the at least one semantic element from the user utterance and the missing semantic element.

4. The system of claim 3, wherein the dialogue stack element comprises a sequentially-stored structured representation of each utterance received from a user associated with the user utterance.

5. The system of claim 1, wherein the intent flow is executed by a task stitcher element configured to:

receive a request to execute the intent flow associated with the identified intent;

select a domain-specific intent flow from a set of intent flows associated with the identified intent;

execute a first task of the domain-specific intent flow, wherein the first task generates a first output;

execute a second task of the domain-specific intent flow, wherein the second task is selected based on the first output, and wherein the second task generates a second output.

6. The system of claim 5, wherein the task stitcher is configured to receive one or more domain-specific intent flows from a designer system associated with a domain.

7. The system of claim 5, wherein the first task comprises an entity check configured to determine if a first required entity is identified.

8. The system of claim 1, wherein the system utterance is generated by a response generator component configured to implement a response generator task.

9. They system of claim 8, wherein the response generator task is configured to select the response template from a plurality of response templates associated with the identified intent.

10. A method comprising:

receiving a user utterance and identify at least one semantic element within the user utterance;

identifying an intent associated with the at least one semantic element;

executing an intent flow associated with the identified intent, wherein the intent flow includes a set of tasks executed in a predetermined order; and

generating a system utterance by instantiating a response template selected from a plurality of response templates associated with the executed intent.

11. The method of claim 10, comprising

identifying a missing semantic element required by the intent flow associated with the identified intent;

reviewing at least one prior utterance to locate the missing semantic element; and

executing the intent flow associated with the identified intent including the at least one semantic element from the user utterance and the missing semantic element identified from the at least one prior utterance.

12. The method of claim 11, wherein the missing semantic element is identified by a context mining component configured to:

maintain a context stack including a set of context entries, wherein each of the context entries comprises a root intent element, an entity list element, and a dialogue stack element;

review at least one context entry in the set of context entries to locate the missing semantic element within the dialogue stack element; and

generate an intent flow execution requires including the at least one semantic element from the user utterance and the missing semantic element.

13. The method of claim 10, wherein the intent flow is executed by a task stitcher element configured to:

receive a request to execute the intent flow associated with the identified intent;

select a domain-specific intent flow from a set of intent flows associated with the identified intent;

execute a first task of the domain-specific intent flow, wherein the first task generates a first output;

execute a second task of the domain-specific intent flow, wherein the second task is selected based on the first output, and wherein the second task generates a second output.

14. The method of claim 13, comprising determining if a first required entity associated with the domain-specific intent flow is identified.

15. The system of claim 13, wherein the task stitcher is configured to receive one or more domain-specific intent flows from a designer system associated with a domain.

16. The method of claim 10, comprising selecting the response template from a plurality of response templates associated with the identified intent.

17. A non-transitory computer readable medium having instructions stored thereon, wherein the instructions, when executed by a processor cause a device to perform operations comprising:

receiving a user utterance and identify at least one semantic element within the user utterance;

identifying an intent associated with the at least one semantic element;

executing an intent flow associated with the identified intent, wherein the intent flow includes a set of tasks executed in a predetermined order; and

generating a system utterance by instantiating a response template selected from a plurality of response templates associated with the executed intent.

18. The non-transitory computer readable medium of claim 17, wherein the instructions, when executed by the processor cause the device to perform operations comprising:

identifying a missing semantic element required by the intent flow associated with the identified intent;

reviewing at least one prior utterance to locate the missing semantic element; and

executing the intent flow associated with the identified intent including the at least one semantic element from the user utterance and the missing semantic element identified from the at least one prior utterance.

19. The non-transitory computer readable medium of claim 17, wherein the instructions, when executed by the processor cause the device to perform operations comprising:

receiving a request to execute the intent flow associated with the identified intent;

selecting a domain-specific intent flow from a set of intent flows associated with the identified intent;

executing a first task of the domain-specific intent flow, wherein the first task generates a first output; and

executing a second task of the domain-specific intent flow, wherein the second task is selected based on the first output, and wherein the second task generates a second output

20. The non-transitory computer readable medium of claim 17, wherein the instructions, when executed by the processor cause the device to perform operations comprising selecting the response template from a plurality of response templates associated with the identified intent.