Demographic-based learning in a question answering system

• US 10,275,487 B2
• Filed: 06/09/2015
• Issued: 04/30/2019
• Est. Priority Date: 06/09/2015
• Status: Active Grant

• Current Assignee: AirBNB Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Eichenlaub, Anthony W., Murch, Cynthia M., Nixa, Terrence T., Nordland, Jan M., Petri, John E., Schlicht, Michelle A.
• Primary Examiner(s): Hershley, Mark E
• Application Number:
  US14/734,481
• Publication Number:
  US 20160364391A1
• Time in Patent Office: 1,421 Days
• Field of Search: 707780
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/2468 Fuzzy queries
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

A first question may be received. A first tag may be identified. The first tag may correspond to a first demographic trait. The first tag may be for use in providing a context for generating a first answer estimate to the first question. The first answer estimate may be generated using natural language processing and based on the first tag.

Claims

1. A computer-implemented method for generating an answer estimate to a question based on demographic traits, the method comprising:

identifying that a user has selected a first tag and a second tag, wherein the user selects the first tag and the second tag from a plurality of demographic trait tags displayed within a graphical user interface (GUI), the first tag corresponding to a first demographic trait and the second tag corresponding to a second demographic trait, and wherein the first tag and the second tag provide a context for generating a first answer estimate to a first question;

receiving the first question, wherein the first question is displayed on the GUI with the plurality of selections, and wherein the first question is presented in each of a text format, a diagrams with no text format, a diagrams with text format, and an audio format; and

generating, using natural language processing and based on the first tag and the second tag, the first answer estimate to the first question, wherein the first answer estimate is generated according to the first and second demographic traits, and the first answer estimate provides associated data for each of the plurality of selections.

2. The method of claim 1, further comprising providing, in response to receiving the first question, a first correct answer, the first correct answer for use in comparing the first correct answer to the first answer estimate in order to make an observation, the observation includes a QA computing system making statistical inferences between answer estimates.

3. The method of claim 1, further comprising:

receiving a second question;

identifying that the first question and the second question has a same correct answer; and

generating, using natural language processing and based on identifying the first tag, a second answer estimate to the second question.

4. The method of claim 3, comprising:

determining, by comparing the first answer estimate with the correct answer, that the first answer estimate does not match the correct answer;

determining, by comparing the second answer estimate with the correct answer, that the second answer estimate matches the correct answer; and

providing an observation, wherein the observation identifies an association between at least two of a group consisting of;

the first question, the first answer estimate, the first tag, the correct answer, the second question, and the second answer estimate.

5. The method of claim 1, further comprising providing, in response to the generating the first answer estimate, an observation, wherein the observation identifies one or more associations between at least two of a group consisting of:

the first question, the first tag, and the first answer estimate.

6. The method of claim 1, wherein the identifying of the first tag is in response to a user selection of the first tag, the method further comprising:

providing, in response to receiving the first question, a correct answer to the first question;

receiving a second question, wherein the second question includes the first question, the second question further including additional information that the first question does not include;

determining that the second question is associated with the correct answer;

generating, using natural language processing and based on identifying the first tag, a second answer estimate to the second question;

determining, by comparing the first answer estimate with the correct answer, that the first answer estimate does not match the correct answer;

determining, by comparing the second answer estimate with the correct answer, that the second answer estimate matches the correct answer; and

providing an observation, wherein the observation identifies that after adding the additional information in the second question, the second answer estimate matches the correct answer.

7. A system for generating an answer estimate to a question based on demographic traits, the system comprising:

a memory;

at least one processor circuit in communication with the memory, wherein the at least one processor circuit is configured to perform a method, the method comprising;

identifying that a user has selected a first tag and a second tag, wherein the user selects the first tag and the second tag from a plurality of demographic trait tags within a graphical user interface (GUI), the first tag corresponding to a first demographic trait and the second tag corresponding to a second demographic trait, and wherein the first tag and the second tag provide a context for generating a first answer estimate to a first question;

receiving a first question, wherein the first question is displayed on the GUI with the plurality of selections, and wherein the first question is presented in each of a text format, a diagrams with no text format, a diagrams with text format, and an audio format; and

generating, using natural language processing and based on the first tag and the second tag, the first answer estimate to the first question, wherein the first answer estimate is generated according to the first and second demographic traits, and the first answer estimate provides associated data for each of the plurality of selections.

8. The system of claim 7, wherein the method further comprises:

providing, in response to receiving the first question, a first correct answer, the first correct answer for use in comparing the first correct answer to the first answer estimate in order to make an observation.

9. The system of claim 7, wherein the method further comprises:

receiving a second question;

associating the first question and the second question with a same correct answer;

generating, using natural language processing and based on identifying the plurality of tags, a second answer estimate to the second question;

determining, by comparing the first answer estimate with the correct answer, that the first answer estimate does not match the correct answer;

determining, by comparing the second answer estimate with the correct answer, that the second answer estimate matches the correct answer; and

providing an observation, wherein the observation identifies an association between at least two of a group consisting of;

the first question, the first answer estimate, the plurality of tags, the correct answer, the second question, and the second answer estimate.

10. The system of claim 9, further comprising:

determining that the first and second answer estimates are different based on the observation; and

generating, using the natural language processing and based on the determining that the first and second answer estimates are different, a second observation, the second observation making statistical inferences associated with the first and second answer estimates.

11. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method, the method comprising:

identifying that a user has selected a first tag and a second tag, wherein the user selects the first tag and the second tag from a plurality of demographic trait tags displayed within a graphical user interface (GUI), the first tag corresponding to a first demographic trait and the second tag corresponding to a second demographic trait, and wherein the first tag and the second tag provide a context for generating a first answer estimate to a first question;

receiving a first question, wherein the first question is displayed on the GUI with the plurality of selections, and wherein the first question is presented in each of a text format, a diagrams with no text format, a diagrams with text format, and an audio format;

identifying, using natural language processing, an answer to the first question;

determining that the answer is incorrect;

training, in response to the determining that the answer is incorrect, a training module until the first question is answered correctly;

generating, using the natural language processing and based on the first tag and the second tag, the first answer estimate to the first question, wherein the first answer estimate is generated according to the first and second demographic traits, and the first answer estimate provides associated data for each of the plurality of selections.

12. The computer program product of claim 11, wherein the method further comprises:

providing, in response to the generating the first answer estimate, an observation, wherein the observation identifies one or more associations between at least two of a group consisting of;

the first question, the first tag, and the first answer estimate.

13. The computer program product of claim 11, wherein the first answer estimate is presented in at least one of the following formats:

text, diagrams, diagrams with text, or audio.

14. The program product of claim 11, wherein the identifying of the first tag is in response to a user selection of the first tag, the method further comprising:

providing, in response to receiving the first question, a correct answer to the first question;

receiving a second question, wherein the second question includes the first question, the second question further including additional information that the first question does not include;

determining that the second question is associated with the correct answer;

generating, using natural language processing and based on identifying the first tag, a second answer estimate to the second question;

determining, by comparing the first answer estimate with the correct answer, that the first answer estimate does not match the correct answer;

determining, by comparing the second answer estimate with the correct answer, that the second answer estimate matches the correct answer; and

providing an observation, wherein the observation identifies that after adding the additional information in the second question, the second answer estimate matches the correct answer.