METHOD AND SYSTEM FOR CONDUCTING AN AUTOMATED CONVERSATION WITH A VIRTUAL AGENT SYSTEM

• US 20190068527A1
• Filed: 08/28/2017
• Published: 02/28/2019
• Est. Priority Date: 08/28/2017
• Status: Abandoned Application

• Current Assignee: Moveworks, Inc.
• Original Assignee: Moveworks, Inc.
• Inventors: CHEN, Jiang, WODEYAR, Vachan, NIVARGI, Vaibhav, SINGH, Varun
• Application Number:
  US15/688,190
• Publication Number:
  US 20190068527A1
• Time in Patent Office: Days
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/006 based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 10/00 Administration; Management
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads

Abstract

A computer-implemented method for conducting a conversation with a virtual agent system is provided. The method includes receiving a conversation input from a user during a conversation of the user with a virtual agent system. The method includes probabilistically matching the conversation input with a stack of earlier conversations between the user and the virtual agent system. The probabilistic matching determines a context of the conversation input from one or more contexts associated with the stack of earlier conversations. The method includes interpreting the conversation input to identify a user intent from among a plurality of user intents. The method further includes determining an action to be performed by the virtual agent system based on the context determined by the probabilistic matching and the user intent.

Claims

1. A computer-implemented method for conducting a conversation with a virtual agent system, the method comprising:

receiving a conversation input from a user during a conversation of the user with a virtual agent system;

probabilistically matching the conversation input with a stack of earlier conversations between the user and the virtual agent system to determine a context of the conversation input, the context determined from one or more contexts associated with the stack of earlier conversations;

interpreting the conversation input to identify a user intent from among a plurality of user intents; and

determining an action to be performed by the virtual agent system based on the context determined by the probabilistic matching and the user intent.

2. The method as claimed in claim 1, wherein determining the action comprises determining a response of the virtual agent system to the conversation input received from the user during the conversation.

3. The method as claimed in claim 2, further comprising sending the response of the virtual agent system to a user device associated with the user.

4. The method as claimed in claim 1, wherein determining the action comprises determining an activity to be performed by the virtual agent system in response to the conversation input received from the user during the conversation.

5. The method as claimed in claim 1, wherein determining the action further comprises determining that the conversation input corresponds to an ongoing context if the conversation input matches at least one context among the one or more contexts associated with the stack of earlier conversations and perform at least one of:

sending a response to a user device associated with the user; and

performing an activity.

6. The method as claimed in claim 1, wherein determining the action further comprises determining that the conversation input relates to a new context if the conversation input does not match the one or more contexts associated with the stack of conversations and perform at least one of:

sending a response to a user device associated with the user; and

performing an activity.

7. The method as claimed in claim 1, wherein the conversation input is a voice input.

8. The method as claimed in claim 1, wherein the conversation input is a text input.

9. The method as claimed in claim 1, wherein interpreting the conversation input to identify the user intent comprises parsing the conversation input to extract one or more data strings defined in at least one intent model storing the plurality of user intents.

10. The method as claimed in claim 1, wherein probabilistically matching the conversation input with the stack of earlier conversations comprises enabling the virtual agent system to lookup the stack of earlier conversations.

11. The method as claimed in claim 1, further comprising enabling the virtual agent system to request the user to modify the conversation input.

12. A system for conducting an automated conversation with a virtual agent system, the system comprising:

a memory configured to store instructions and at least one intent model; and

a processor configured to execute the stored instructions to cause the system to at least perform;

receiving a conversation input from a user during a conversation of the user with a virtual agent system;

probabilistically matching the conversation input with a stack of earlier conversations between the user and the virtual agent system to determine a context of the conversation input, the context determined from one or more contexts associated with the stack of earlier conversations;

interpreting the conversation input to identify a user intent from among a plurality of user intents; and

determining an action to be performed by the virtual agent system based on the context determined by the probabilistic matching and the user intent.

13. The system as claimed in claim 12, wherein for determining the action, the system is further caused at least in part to determine response of the virtual agent system to the conversation input received from the user during the conversation.

14. The system as claimed in claim 12, wherein for determining the action, the system is further caused at least in part to determine an activity to be performed by the virtual agent system in response to the conversation input received from the user during the conversation.

15. The system as claimed in claim 12, wherein for interpreting the conversation input to identify the user intent, the system is further caused at least in part to, parse the conversation input to extract one or more data strings defined in the at least one intent model.

16. The system as claimed in claim 12, wherein for probabilistically matching the conversation input with the stack of earlier conversations, the system is further caused at least in part to:

determine that the conversation input relates to an ongoing context if the conversation input matches at least one context among the one or more contexts associated with the stack of earlier conversations; and

perform at last one of;

sending a response to a user device associated with the user; and

performing an activity.

17. The system as claimed in claim 12, wherein for probabilistically matching the conversation input with the stack of earlier conversations, the system is caused at least in part to:

determine that the conversation input relates to a new context if the conversation input does not match the one or more contexts associated with the stack of earlier conversations; and

perform at last one of;

sending a response to a user device associated with the user; and

performing an activity.

18. The system as claimed in claim 12, wherein the system is further caused to enable the virtual agent system to request the user to modify the conversation input.

19. The system as claimed in claim 12, further comprising a machine learning module configured to train the intent model on a plurality of conversation inputs received from a plurality of users through a plurality of user devices.

20. A computer program product including a computer readable storage medium and including computer executable code for a virtual agent system, the computer executable code being executed by a processor, the computer program product comprising one or more instructions, the instructions causing the virtual agent system to:

receive a conversation input from a user during a conversation of the user with a virtual agent system;

probabilistically match the conversation input with a stack of earlier conversations between the user and the virtual agent system to determine a context of the conversation input, the context determined from one or more contexts associated with the stack of earlier conversations;

interpret the conversation input to identify a user intent from among a plurality of user intents; and

determine an action to be performed by the virtual agent system based on the context determined by the probabilistic matching and the user intent.