Systems and methods for a personality consistent chat bot

• US 10,250,532 B2
• Filed: 04/28/2017
• Issued: 04/02/2019
• Est. Priority Date: 04/28/2017
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Koukoumidis, Emmanouil, Johnson, Jr., Joseph, Abramson, Dustin I., Rangel, Maria A., Wang, Ying, Li, Di, Zittrower, Steven G., Mu, Hailong, Stuckart, Kelli A., Wang, Hudong, Hao, Pengfei, Brinkman, Donald F., London, Kati A., Chen, Zhan
• Primary Examiner(s): Boutah, Alina A
• Application Number:
  US15/581,116
• Publication Number:
  US 20180316631A1
• Time in Patent Office: 704 Days
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• 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 3/045 Combinations of networks
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• 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]

Abstract

Systems and methods for personality consistent automated chatting are provided. The systems and methods track the emotional state of the chat bot, track entity related dispositions of the chat bot and/or track the closeness of the relationship between the chat bot and a user. Accordingly, the systems and methods are able to provide consistent personality and/or points of view by generating replies based on the tracked chat bot emotional state, chat bot entity dispositions, and/or the relationship closeness between the chat bot and the user.

Claims

1. A system for a personality consistent chat bot, the system comprising:

at least one processor; and

a memory for storing and encoding computer executable instructions that, when executed by the at least one processor is operative to;

identify that a first reply is warranted from the chat bot regarding a first entity to a user;

search an entity-based disposition memory graph for the chat bot based on the first entity in response to identifying that the first reply is warranted;

identify that the entity-based disposition memory graph does not include the first entity;

extract a user disposition for the first entity from user information in response to identifying that the entity-based disposition memory graph does not include the first entity;

select a chat bot disposition that is similar to the user disposition for the first entity;

generate the first reply about the first entity based on the chat bot disposition for the first entity;

provide the first reply to the user;

update the entity-based disposition memory graph to include the first entity and the chat bot disposition for the first entity to form an updated entity-based disposition memory graph for the chat bot;

identify that a second reply is warranted from the chat bot regarding the first entity to the user;

search the updated entity-based disposition memory graph for the chat bot based on the first entity in response to identifying that the second reply is warranted;

identify the chat bot disposition for the first entity that was previously provided by the chat bot based on the search of updated entity-based disposition memory graph;

generate the second reply about the first entity based on the chat bot disposition for the first entity; and

provide the second reply to the user.

2. The system of claim 1, wherein the at least one processor is further operative to:

save the user disposition for the first entity to a user entity-based disposition memory graph.

3. The system of claim 2, wherein the at least one processor is further operative to:

extract the user disposition for the first entity from updated user information in response to identifying that the second reply is warranted from the chat bot regarding the first entity to the user;

determining that the user disposition for the first entity has changed to form a changed user disposition for the first entity;

updating the user entity-based disposition memory graph with the changed user disposition for the first entity to form an updated user entity-based disposition memory graph; and

generate a third reply that asks about the changed user disposition for the first entity.

4. The system of claim 3, wherein the updated entity-based disposition memory graph for the chat bot is maintained and is not changed based on the changed user disposition for the first entity.

5. The system of claim 3, wherein the at least one processor is further operative to:

determine a difference between the updated user entity-based disposition memory graph and the updated entity-based disposition memory graph for the chat bot for the first entity;

compare the difference to a threshold;

determine that the difference meets the threshold;

modify the updated entity-based disposition memory graph for the chat bot for the first entity to be closer to, but not the same as, the changed user disposition for the first entity after providing at least a fourth reply, a fifth reply, a sixth reply, a seventh reply, and an eighth reply to the user regarding the first entity to form a modified entity-based disposition memory graph for the chat bot.

6. The system of claim 5, wherein the at least one processor is further operative to:

identify that a ninth reply is warranted from the chat bot regarding the first entity to the user;

search the modified entity-based disposition memory graph for the chat bot based on the first entity;

identify a modified chat bot disposition for the first entity based on the searching of the modified entity-based disposition memory graph for the chat bot;

generate the ninth reply about the first entity based on the modified chat bot disposition for the first entity; and

provide the ninth reply to the user.

7. The system of claim 1, wherein the user disposition for the first entity includes at least one of a user sentiment, a user emotion, and an associated user strength for the user sentiment or the user emotion;

and

wherein the chat bot disposition for the first entity includes at least one of a chat bot sentiment, a chat bot emotion, and an associated chat bot strength for the chat bot sentiment or the chat bot emotion.

8. The system of claim 1, wherein the at least one processor is further operative to:

identify that a third reply is warranted from the chat bot regarding a second entity to the user;

search the updated entity-based disposition memory graph for the chat bot based on the second entity in response to identifying that the third reply is warranted;

identify a chat bot disposition for the second entity based on the searching of the updated entity-based disposition memory graph;

generate the third reply about the second entity based on the chat bot disposition for the second entity; and

provide the third reply to the user.

9. A method for automated chatting with personality consistency, the method comprising:

identifying that a first reply is warranted from a chat bot regarding a first entity to a user;

searching an entity-based disposition memory graph for the chat bot based on the first entity;

identifying a chat bot disposition for the first entity that was previously provided by the chat bot based on the searching of the entity-based disposition memory graph for the chat bot;

generating the first reply about the first entity based on the chat bot disposition for the first entity; and

providing the first reply to the user.

10. The method of claim 9, further comprising:

collecting at least one of user input and user information for a second entity;

extracting a user disposition for the second entity from at least one of the user input and the user information for the second entity;

selecting a similar chat bot disposition for the second entity for the chat bot;

saving the similar chat bot disposition for the second entity for the chat bot to the entity-based disposition memory graph for the chat bot;

identifying that a second reply is warranted from the chat bot regarding the second entity to the user;

searching the entity-based disposition memory graph for the chat bot based on the second entity;

identifying the similar chat bot disposition for the second entity based on the searching of the entity-based disposition memory graph for the chat bot;

generating the second reply about the second entity based on the similar chat bot disposition for the second entity; and

providing the second reply to the user.

11. The method of claim 9, further comprising:

searching world knowledge based on the first entity;

determining a related entity to the first entity that is categorically at least one level up from the first entity, wherein the related entity is a second entity;

inferring a similar chat bot disposition for the second entity based on the chat bot disposition for the first entity;

saving the similar chat bot disposition for the second entity to the entity-based disposition memory graph for the chat bot;

identifying that a second reply is warranted from the chat bot regarding the second entity to the user;

searching the entity-based disposition memory graph for the chat bot based on the second entity;

identifying the similar chat bot disposition for the second entity based on the searching of the entity-based disposition memory graph for the chat bot;

generating the second reply about the second entity based on the similar chat bot disposition for the second entity; and

providing the second reply to the user.

12. The method of claim 9, wherein the chat bot disposition for the first entity includes at least one of a chat bot sentiment, a chat bot emotion, and an associated chat bot strength for the chat bot sentiment or the chat bot emotion.

13. The method of claim 9, wherein the entity-based disposition memory graph for the chat bot includes entities labeled with at least one of sentiments, emotions, and strengths that are predetermined during creation of the chat bot.

14. The method of claim 9, further comprising:

tracking a relationship closeness between the chat bot and the user;

wherein the generating the first reply about the first entity is further based on the relationship closeness between the chat bot and the user.

15. The method of claim 9, further comprising:

tracking an emotional state of the chat bot;

wherein the generating the first reply about the first entity is further based on the emotional state of the chat bot.

16. A system for a personality consistent chat bot, the system comprising:

at least one processor; and

a memory for storing and encoding computer executable instructions that, when executed by the at least one processor is operative to;

track an emotional state of the chat bot;

form an emotional state memory graph for the chat bot based on the tracking of the emotional state of the chat bot;

identify that a reply is warranted from a chat bot;

search the emotional state memory graph for the chat bot in response to identifying that the reply is warranted;

identify a current emotional state of the chat bot based on the searching of the emotional state memory graph;

generate the reply based on the current emotional state of the chat bot; and

provide the reply to a client computing device.

17. The system of claim 16, wherein the at least one processor is further operative to:

track a user emotional state;

wherein generating the reply is further based on the user emotional state.

18. The system of claim 17, wherein the at least one processor is further operative to:

track chat bot dispositions about different entities as provided in replies by the chat bot;

form an entity-based disposition memory graph for the chat bot based on the tracking of the chat bot dispositions for the different entities;

identify that the reply relates to an entity;

search the entity-based disposition memory graph for the chat bot for the entity;

identify a chat bot disposition associated with the entity based on the search of the entity-based disposition memory graph for the chat bot;

wherein generating the reply is further based on the chat bot disposition for the entity.

19. The system of claim 18, wherein the chat bot dispositions for the different entities include a chat bot emotion and an associated strength for the chat bot emotion.

20. The system of claim 17, wherein the user emotional state is at least one of:

happy, sad, angry, neutral, healthy, sick, disgust, contempt, fear, excitement, mature, and immature.