Method and system for self-learning natural language predictive searching

Patent metadata

Field	Value
Publication number	US-11514036-B2
Application number	US-201916535500-A
Country	US
Kind code	B2
Filing date	Aug 8, 2019
Priority date	Aug 14, 2018
Publication date	Nov 29, 2022
Grant date	Nov 29, 2022

Abstract

Systems and methods are provided for self-learning natural language predictive searching including receiving a first input, the first input being related to the desired outcome; retrieving a first information related to the first input; determining a first output based on at least the first input and the first information; outputting the first output; receiving a second input based on the outputted first output in response to the first output being different from the desired outcome, the second input being related to the desired outcome; retrieving, by the processor, a second information related to the second input; determining a second output based on at least the second input, the second information, the first input and the first information; and outputting the second output.

First claim

What is claimed is:

1. A method for performing self-learning natural language predictive searching to reach a desired outcome, the method being implemented by a processor on a computing device, the method comprising:
receiving, by the processor, a first input related to the desired outcome, the first input corresponding to a natural language input;
segmenting, by the processor, a natural language sentence from the first input into a plurality of tokens;
generating, by the processor, at least one feature that is usable by a machine learning algorithm based on domain knowledge of the first input;
locating, by the processor, at least one entity mentioned in the natural language input based on extracted data from the first input;
retrieving, by the processor, a first information related to the first input;
determining, by the processor, a first output based on the first input, the plurality of tokens, the at least one feature, the at least one entity, and the first information;
outputting, by the processor, the first output;
receiving, by the processor, a second input based on the outputted first output in response to the first output being different from the desired outcome, the second input being related to the desired outcome;
retrieving, by the processor, a second information related to the second input;
determining, by the processor, a second output based on at least the second input, the second information, the first input, the first information and the first output; and
outputting, by the processor, the second output.

2. The method of claim 1 , wherein at least one of the retrieving the first information and the retrieving the second information comprises:
accessing a past history of a user entering the first input and the second input, the past history including past requests from the user and outcomes of the past requests, the first input and the second input being related to a specific domain;
accessing a past history of other users related to the specific domain, the past history including other past requests from the other users and other outcomes of the other past requests.

3. The method of claim 2 , wherein at least one of the determining the first output and the determining the second output comprises utilizing a combination of natural processing language and machine learning.

4. The method of claim 2 , wherein at least one of the determining the first output and the determining the second output comprises:
processing the first input and the first information via machine learning;
determining at least one discovery path to the desired outcome based on the processed first input and first information; and
modifying the determined at least one discovery path based on the second input, the second information and the first output.

5. The method of claim 4 , wherein the utilizing the combination of natural processing language and machine learning comprises utilizing machine learning based on reinforced learning.

6. The method of claim 1 , wherein the first information and the second information comprise data defining a context of the first and second input, respectively.

7. The method of claim 1 , wherein at least one of the first input and the second input comprises at least one of a question asked by a user, a menu selection, an identifier of the user, search parameters of the user, and search parameters of other users.

8. The method of claim 1 , wherein at least one of the outputting the first output and the outputting the second output comprises displaying a first dashboard illustrating the first output and displaying a second dashboard illustrating the second output, the second dashboard being a modified version of the first dashboard.

9. The method of claim 8 , wherein second information displayed on the second dashboard is closer to the desired outcome than first information displayed on the first dashboard.

10. The method of claim 9 , wherein:
an iteration comprises receiving an input, retrieving information related to the input and outputting an output; and
the second information is closer to the desired outcome than the first information when a number of iterations to reach the desired outcome starting from the second output is lower than the number of iterations to reach the desired outcome starting from the first output.

11. A computing device configured to implement an execution of a method for self-learning natural language predictive searching to reach a desired outcome, the computing device comprising:
a display screen;
a processor;
a memory; and
a communication interface coupled to each of the processor, the memory, and the display screen,
wherein, when the method is being executed, the processor is configured to:
receive, via the communication interface, a first input related to the desired outcome, the first input corresponding to a natural language input;
segment a natural language sentence from the first input into a plurality of tokens;
generate at least one feature that is usable by a machine learning algorithm based on domain knowledge of the first input;
locate at least one entity mentioned in the natural language input based on extracted data from the first input;
retrieve a first information related to the first input;
determine a first output based on at least the first input, the plurality of tokens, the at least one feature, the at least one entity, and the first information;
output the first output;
receive a second input based on the outputted first output in response to the first output being different from the desired outcome, the second input being related to the desired outcome;
retrieve a second information related to the second input;
determine a second output based on at least the second input, the second information, the first input, the first information and the first output; and
output the second output.

12. The computing device of claim 11 , wherein the processor, in order to at least one of retrieve the first information and retrieve the second information, is further configured to:
access a past history of a user entering the first input and the second input, the past history including past requests from the user and outcomes of the past requests, the first input and the second input being related to a specific domain;
access a past history of other users related to the specific domain, the past history including other past requests from the other users and other outcomes of the other past requests.

13. The computing device of claim 12 , wherein, in order to determine at least one of the first output and the second output, the processor is further configured to utilize a combination of natural processing language and machine learning.

14. The computing device of claim 12 , wherein, in order to determine the at least one of the first output and the second output, the processor is further configured to:
process the first input and the first information via machine learning;
determine at least one discovery path to the desired outcome based on the processed first input and first information; and
modify the determined at least one discovery path based on the second input, the second information and the first output.

15. The computing device of claim 14 , wherein the processor is further configured to utilize machine learning based on reinforced learning.

16. The computing device of claim 11 , wherein the first and second information comprise data defining a context of the first and second input, respectively.

17. The computing device of claim 11 , wherein at least one of the f

Assignees

• Jpmorgan Chase Bank Na

Inventors

• Chin Ke-Vin
• Balasubramanian Chandrasekaran
• Sankaran Krishnan P

Classifications

• G06N5/042

  Backward inferencing · CPC title

• G06F3/0484

  for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range · CPC title

• G06F16/24522 Primary

  Translation of natural language queries to structured queries · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/248

  Presentation of query results · CPC title