METHOD AND SYSTEM FOR FACILITATING A USER-MACHINE CONVERSATION

• US 20180181558A1
• Filed: 12/22/2016
• Published: 06/28/2018
• Est. Priority Date: 12/22/2016
• Status: Active Grant

• Current Assignee: Verizon Patent and Licensing Incorporated (Verizon Communications Inc.)
• Original Assignee: Yahoo Holdings, Inc. (Verizon Communications Inc.)
• Inventors: Emery, Michael, Viswanathan, Lavanya Colinjivadi
• Granted Patent: US 11,138,388 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/3346 using probabilistic model
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00

Abstract

The present teaching relates to facilitating a user-machine conversation. In one example, a query is obtained from a user. The query is directed to a first conversational bot. A reply in response to the query is obtained from the first conversational bot. A degree of validity of the reply is determined based on the reply and the query. A second conversational bot is determined based on the query and the degree of validity. The conversation is directed to the second conversational bot with the query.

Claims

1. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network for facilitating a user-machine conversation, the method comprising:

obtaining a query from a user, wherein the query is directed to a first conversational bot;

obtaining, from the first conversational bot, a reply in response to the query;

determining a degree of validity of the reply based on the reply and the query;

determining a second conversational bot based on the query and the degree of validity; and

directing the conversation to the second conversational bot with the query.

2. The method of claim 1, wherein determining the degree of validity comprises:

determining a degree of matching between the reply and the query based on their semantic meanings;

obtaining a confidence score related to the reply from the first conversational bot; and

determining the degree of validity of the reply based on the degree of matching and the confidence score.

3. The method of claim 1, further comprising:

comparing the degree of validity with a threshold;

providing the reply as a response to the query and providing the query and the reply for training a model for bot recommendation, when the degree of validity exceeds the threshold; and

determining the second conversational bot when the degree of validity does not exceed the threshold.

4. The method of claim 1, wherein determining the second conversational bot comprises:

parsing the query to determine one or more features;

obtaining a model for bot recommendation; and

determining the second conversational bot based on the one or more features and the model.

5. The method of claim 4, further comprising:

determining a plurality of conversational bots based on the one or more features and the model;

determining, for each of the plurality of conversational bots, a confidence score;

ranking the plurality of conversational bots based on their respective confidence scores;

determining one or more thresholds based on the reply; and

selecting the second conversational bot from the plurality of conversational bots based on the ranking and the one or more thresholds.

6. The method of claim 5, further comprising:

generating a new conversational bot based on the query when none of the confidence scores of the plurality of conversational bots passes the one or more thresholds.

7. The method of claim 4, further comprising:

obtaining a plurality of query-reply pairs, wherein each of the plurality of query-reply pairs includes a historical query and a historical reply corresponding to the historical query, wherein the historical reply was provided by a conversational bot with an associated confidence score;

selecting one or more query-reply pairs from the plurality of query-reply pairs based on their respective associated confidence scores and a threshold;

assigning one or more conversational bots to each historical query in the one or more query-reply pairs based on the conversational bots providing historical replies corresponding to the historical query;

generating training data based on the selecting and assigning; and

training the model for bot recommendation based on the training data.

8. The method of claim 1, further comprising:

recommending the second conversational bot to the user as a response to the query.

9. The method of claim 1, wherein the first and second conversational bots are independent and related to different topics, wherein the topics include at least one of:

weather, restaurant, hotel, travel, sports, finance, and news.

10. A system, having at least one processor, storage, and a communication platform connected to a network for facilitating a user-machine conversation, comprising:

a user request analyzer configured for obtaining a query from a user, wherein the query is directed to a first conversational bot;

a bot reply analyzer configured for obtaining, from the first conversational bot, a reply in response to the query and determining a degree of validity of the reply based on the reply and the query;

a conversational bot recommender configured for determining a second conversational bot based on the query and the degree of validity; and

a conversational bot determiner configured for directing the conversation to the second conversational bot with the query.

11. The system of claim 10, wherein the bot reply analyzer comprises:

a query/reply semantic matcher configured for determining a degree of matching between the reply and the query based on their semantic meanings;

a reply confidence score determiner configured for obtaining a confidence score related to the reply from the first conversational bot; and

a bot switch request generator configured for determining the degree of validity of the reply based on the degree of matching and the confidence score.

12. The system of claim 10, wherein:

the bot reply analyzer is further configured for;

comparing the degree of validity with a threshold, and

providing the reply as a response to the query and providing the query and the reply for training a model for bot recommendation, when the degree of validity exceeds the threshold; and

the second conversational bot is determined when the degree of validity does not exceed the threshold.

13. The system of claim 10, wherein the conversational bot recommender comprises:

a user query parser configured for parsing the query to determine one or more features;

a matching bot determiner configured for obtaining a model for bot recommendation; and

a bot switch instruction generator configured for determining the second conversational bot based on the one or more features and the model.

14. The system of claim 13, wherein the conversational bot recommender comprises:

the matching bot determiner configured for determining a plurality of conversational bots based on the one or more features and the model;

a matching bot confidence score determiner configured for determining, for each of the plurality of conversational bots, a confidence score;

a matching bot ranker configured for ranking the plurality of conversational bots based on their respective confidence scores;

an original reply based threshold generator configured for determining one or more thresholds based on the reply; and

the bot switch instruction generator configured for selecting the second conversational bot from the plurality of conversational bots based on the ranking and the one or more thresholds.

15. The system of claim 14, further comprising:

a new bot generator configured for generating a new conversational bot based on the query when none of the confidence scores of the plurality of conversational bots passes the one or more thresholds.

16. A non-transitory machine-readable medium having information recorded thereon for facilitating a user-machine conversation, wherein the information, when read by the machine, causes the machine to perform the following:

obtaining a query from a user, wherein the query is directed to a first conversational bot;

obtaining, from the first conversational bot, a reply in response to the query;

determining a degree of validity of the reply based on the reply and the query;

determining a second conversational bot based on the query and the degree of validity; and

directing the conversation to the second conversational bot with the query.

17. The medium of claim 16, wherein determining the degree of validity comprises:

determining a degree of matching between the reply and the query;

obtaining a confidence score related to the reply from the first conversational bot; and

determining the degree of validity of the reply based on the degree of matching and the confidence score.

18. The medium of claim 16, wherein the information, when read by the machine, further causes the machine to perform the following:

comparing the degree of validity with a threshold;

providing the reply as a response to the query and providing the query and the reply for training a model for bot recommendation, when the degree of validity exceeds the threshold; and

determining the second conversational bot when the degree of validity does not exceed the threshold.

19. The medium of claim 16, wherein determining the second conversational bot comprises:

parsing the query to determine one or more features;

obtaining a model for bot recommendation; and

determining the second conversational bot based on the one or more features and the model.

20. The medium of claim 19, wherein the information, when read by the machine, further causes the machine to perform the following:

determining a plurality of conversational bots based on the one or more features and the model;

determining, for each of the plurality of conversational bots, a confidence score;

ranking the plurality of conversational bots based on their respective confidence scores;

determining one or more thresholds based on the reply; and

selecting the second conversational bot from the plurality of conversational bots based on the ranking and the one or more thresholds.