ADAPTIVE DIALOG STRATEGY FOR MULTI TURN CONVERSATION SYSTEMS USING INTERACTION SEQUENCES

• US 20200137014A1
• Filed: 10/26/2018
• Published: 04/30/2020
• Est. Priority Date: 10/26/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: KASIREDDY, VINAY KUMAR REDDY, MUKHI, NIRMAL K., TEJWANI, RAVI
• Granted Patent: US 11,140,110 B2
• Time in Patent Office: Days
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: H04L 51/04 Real-time or near real-time messaging, e.g. instant messaging [IM]
• CPC Class Codes: H04L 51/216 Handling conversation history, e.g. grouping of messages in sessions or threads

Abstract

A method, computer program product and computer system for recovering from a likely exit condition in an electronic conversation with a user and interactive agent is provided. A processor retrieves one or more messages from a conversation between a user and an interactive agent. A processor determines at least one label for the one or more messages between a user and an interactive agent. A processor compares the at least one label for the one or more messages to a sequence model. A processor identifies the one or more messages from the conversation indicate an exit condition for the user based on the comparison of the at least one label for the one or more messages to the sequence model. A processor sends a recovery message to the user to redirect the conversation.

Claims

1. A method comprising:

retrieving, by one or more processors, one or more messages from a conversation between a user and an interactive agent;

determining, by the one or more processors, at least one label for the one or more messages between a user and an interactive agent;

comparing, by the one or more processors, the at least one label for the one or more messages to a sequence model;

identifying, by the one or more processors, the one or more messages from the conversation indicate an exit condition for the user based, at least in part, on the comparison of the at least one label for the one or more messages to the sequence model; and

sending, by the one or more processors, a recovery message to the user to redirect the conversation.

2. The method of claim 1, wherein identifying the one or more messages from the conversation indicate an exit condition further comprises:

determining, by the one or more processors, at least one likelihood of a next state in the sequence model.

3. The method of claim 2, wherein the at least one likelihood of a next state in the sequence model is based, at least in part, on a matching sequence between (i) the at least on label for the one or more messages between a user and an interactive agent and (ii) at least one sequence in the sequence model.

4. The method of claim 3, wherein the at least one likelihood of a next state in the sequence model is further based, at least in part, on at least one of (i) a response time between two messages of the one or more messages, or (ii) a mastery of at least one message.

5. The method of claim 1, wherein determining at least on label for the one or more messages between a user and an interactive agent is based on one or more of the following:

(i) content of the one or more messages or (ii) intent of the one or more messages.

6. The method of claim 1, the method further comprising:

updating, by the one or more processors, a first label of the at least one label for the one or more messages between a user and an interactive agent based, at least in part, on a response to the recovery message.

7. The method of claim 1, wherein the sequence model is a Markov sequence model.

8. A computer program product comprising:

one or more computer-readable storage media and program instructions stored on the one or more computer-readable storage media, the program instructions comprising;

program instructions to retrieve one or more messages from a conversation between a user and an interactive agent;

program instructions to determine at least one label for the one or more messages between a user and an interactive agent;

program instructions to compare the at least one label for the one or more messages to a sequence model;

program instructions to identify the one or more messages from the conversation indicate an exit condition for the user based, at least in part, on the comparison of the at least one label for the one or more messages to the sequence model; and

program instructions to send a recovery message to the user to redirect the conversation.

9. The computer program product of claim 8, wherein program instructions to identify the one or more messages from the conversation indicate an exit condition further comprises:

program instructions to determine at least one likelihood of a next state in the sequence model.

10. The computer program product of claim 9, wherein the at least one likelihood of a next state in the sequence model is based, at least in part, on a matching sequence between (i) the at least on label for the one or more messages between a user and an interactive agent and (ii) at least one sequence in the sequence model.

11. The computer program product of claim 10, wherein the at least one likelihood of a next state in the sequence model is further based, at least in part, on at least one of (i) a response time between two messages of the one or more messages, or (ii) a mastery of at least one message.

12. The computer program product of claim 8, wherein program instructions to determine at least on label for the one or more messages between a user and an interactive agent is based on one or more of the following:

(i) content of the one or more messages or (ii) intent of the one or more messages.

13. The computer program product of claim 8, the program instructions further comprising:

program instructions to update a first label of the at least one label for the one or more messages between a user and an interactive agent based, at least in part, on a response to the recovery message.

14. The computer program product of claim 8, wherein the sequence model is a Markov sequence model.

15. A computer system for comprising:

one or more computer processors;

one or more computer readable storage media; and

program instructions stored on the computer readable storage media for execution by at least one of the one or more processors, the program instructions comprising;

program instructions to retrieve one or more messages from a conversation between a user and an interactive agent;

program instructions to determine at least one label for the one or more messages between a user and an interactive agent;

program instructions to compare the at least one label for the one or more messages to a sequence model;

program instructions to identify the one or more messages from the conversation indicate an exit condition for the user based, at least in part, on the comparison of the at least one label for the one or more messages to the sequence model; and

program instructions to send a recovery message to the user to redirect the conversation.

16. The computer system of claim 15, wherein program instructions to identify the one or more messages from the conversation indicate an exit condition further comprises:

program instructions to determine at least one likelihood of a next state in the sequence model.

17. The computer system of claim 16, wherein the at least one likelihood of a next state in the sequence model is based, at least in part, on a matching sequence between (i) the at least on label for the one or more messages between a user and an interactive agent and (ii) at least one sequence in the sequence model.

18. The computer system of claim 17, wherein the at least one likelihood of a next state in the sequence model is further based, at least in part, on at least one of (i) a response time between two messages of the one or more messages, or (ii) a mastery of at least one message.

19. The computer system of claim 15, wherein program instructions to determine at least on label for the one or more messages between a user and an interactive agent is based on one or more of the following:

(i) content of the one or more messages or (ii) intent of the one or more messages.

20. The computer system of claim 15, the program instructions further comprising:

program instructions to update a first label of the at least one label for the one or more messages between a user and an interactive agent based, at least in part, on a response to the recovery message.