System and Method for Evaluating Intent of a Human Partner to a Dialogue Between Human User and Computerized System

• US 20140222433A1
• Filed: 09/19/2012
• Published: 08/07/2014
• Est. Priority Date: 09/19/2011
• Status: Active Grant

• Current Assignee: Personetics Technologies Limited (Personetics Technologies, Inc.)
• Original Assignee: Personetics Technologies Limited (Personetics Technologies, Inc.)
• Inventors: Govrin, David D., Sosna, David, Ophir, Ido, Vatnik, Dan
• Granted Patent: US 9,495,962 B2
• Time in Patent Office: Days
• US Class Current: 704/260
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2013/083 Special characters, e.g. punctuation marks
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages

Abstract

A system and method for assigning relative scores to various possible intents on the part of a user approaching a virtual agent, the method comprising predicting priority topics, including gathering first data and employing the first data to discern and seek user confirmation of at least one possible intent on the part of the user; and subsequent to receipt of the confirmation, gathering second data and employing the second data to provide service to the user to suit the user'"'"'s confirmed intent.

Claims

1. A computerized method for assigning relative scores to various possible intents on the part of a user approaching a virtual agent, the method comprising:

employing a processor for predicting priority topics, including gathering first data and employing the first data to discern and seek user confirmation of at least one possible intent on the part of the user; and

subsequent to receipt of said confirmation, employing a processor for gathering second data and employing the second data to provide service to the user to suit the user'"'"'s confirmed intent.

2. A method according to claim 1 wherein said predicting employs a prediction model to generate, from said first data, a list of topics which the customer is likely to raise.

3. A method according to claim 1 wherein said first data comprises historical information regarding transactional activities involving at least the user.

4. A method according to claim 1 wherein the system is operative to prompt at least one user from among a population of users by presenting, and seeking user confirmation, of at least one possible intent before the user has presented any request.

5. A method according to claim 4 wherein the system generates a combined score which includes user input for improving the accuracy of a text retrieval system.

6. A method according to claim 3 wherein said activities comprise at least one credit card transaction.

7. A method according to claim 3 wherein said historical information comprises billing information.

8. A method according to claim 1 wherein the user is also a user of a computerized enterprise and wherein said first data includes enterprise data obtained from said enterprise which establishes a business, data and dialogue context for dialog with the user.

9. A method according to claim 1 also comprising using a result of said predicting to automatically construct a guide for the user on how to phrase requests, in contrast to conventionally pure NLP-based dialogue systems in which users are provided with empty text boxes and are expected to phrase their requests without guidance.

10. A method according to claim 1 wherein said predicting uses executable code having forward-chaining rule-based deduction logic.

11. A method according to claim 9 wherein said using comprises presenting a multiple-choice selection to the user including a plurality of options for selection wherein each option suggests a specific topic closure as a next step in a dialogue with the user.

12. A method according to claim 11 wherein relevance scores generated by said predicting are used to sort selection options such that highest relevance are on top.

13. A method according to claim 11 wherein variable binding is used to automatically generate appropriate natural language names for topic closures.

15. A method according to claim 5 wherein the user input includes text retrieved from the user as a reply to a “

how can I help you”

-type question.

16. A method according to claim 15 wherein the user input also includes text entered by the user later in the dialog to direct a next step in the dialog.

14. A system for assigning relative scores to various possible intents on the part of a user approaching a virtual agent, the system comprising:

a dialog topic predictor operative for predicting priority topics, including gathering first data and employing the first data to discern and seek user confirmation of at least one possible intent on the part of the user; and

a dialog generator operative, subsequent to receipt of said confirmation, for gathering second data and employing the second data to provide dialog with a user to suit the user'"'"'s confirmed intent.

17. A computer program product, comprising a non-transitory tangible computer-readable medium having computer-readable program code embodied therein, said computer-readable program code adapted to be executed to implement a method for assigning relative scores to various possible intents on the part of a user approaching a virtual agent, the method comprising:

predicting priority topics, including gathering first data and employing the first data to discern and seek user confirmation of at least one possible intent on the part of the user; and

subsequent to receipt of said confirmation, gathering second data and employing the second data to provide service to the user to suit the user'"'"'s confirmed intent.