Question Correction and Evaluation Mechanism for a Question Answering System

• US 20160062980A1
• Filed: 08/29/2014
• Published: 03/03/2016
• Est. Priority Date: 08/29/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Boguraev, Branimir K., Bufe, John P. III, Hatem, Matthew T., Smythe, Jared M.D.
• Granted Patent: US 10,671,929 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

Mechanisms are provided in a question answering (QA) system comprising a QA system pipeline that analyzes an input question and generates an answer to the input question, for pre-processing the input question. The mechanisms receive an input question and input the input question to a pre-processor flow path having one or more pre-processors. The one or more pre-processors transform the input question into a transformed question by correcting errors in a formulation of the input question that are determined to be detrimental to efficient and accurate processing of the input question by a QA system pipeline of the QA system. The transformed question is then input to the QA system pipeline of the QA system which processes the transformed question to generate and output an answer to the input question.

Claims

1. A method, in a question answering (QA) system comprising a QA system pipeline that analyzes an input question and generates an answer to the input question, for pre-processing the input question, the method comprising:

receiving, by the QA system, an input question for which an answer is sought by a submitter of the input question;

inputting, by the QA system, the input question to a pre-processor flow path comprising at least one pre-processor;

transforming, by the at least one pre-processor, the input question into a transformed question by correcting errors in a formulation of the input question that are determined to be detrimental to efficient and accurate processing of the input question by a QA system pipeline of the QA system;

inputting the transformed question to the QA system pipeline of the QA system; and

processing the transformed question via the QA system pipeline to generate and output at least one answer to the input question.

2. The method of claim 1, wherein the at least one pre-processor comprises a superfluous information pre-processor that identifies superfluous information content in the input question and transforms the superfluous information content.

3. The method of claim 2, wherein transforming the input question comprises:

identifying at least one of a term, phrase, or language pattern indicative of superfluous information content in the input question that is not relevant to the underlying question being asked by the input question;

removing the superfluous information content from the input question to generate a modified input question;

in response to determining that the modified input question requires reformatting, reformatting the input question into a logically correct input question; and

outputting the logically correct input question as the transformed question.

4. The method of claim 3, wherein identifying at least one of a term, phrase, or language pattern indicative of superfluous information content in the input question comprises identifying at least one of a vocative clause or a common phrase in the input question.

5. The method of claim 2, wherein the superfluous information content in the input question comprises at least one of a vocative clause or a common phrase in the input question.

6. The method of claim 1, wherein the at least one pre-processor comprises an elided information pre-processor that identifies elided information content associated with the input question and transforms the input question based on the identified elided information.

7. The method of claim 6, wherein transforming the input question comprises:

identifying at least one of a term, phrase, or language pattern indicative of elided information content in the input question that is relevant to the underlying question being asked by the input question and is missing in the input question;

inserting the elided information content into the input question to generate a modified input question;

in response to determining that the modified input question requires reformatting, reformatting the input question into a logically correct input question; and

outputting the logically correct input question as the transformed question.

8. The method of claim 7, wherein identifying at least one of a term, phrase, or language pattern indicative of elided information content in the input question comprises identifying at least one of an elided focus, an elided verb, or an elided noun phrase.

9. The method of claim 7, wherein inserting the elided information content into the input question to generate a modified input question comprises analyzing at least one of a domain or a context associated with the input question to determine the elided information content to be inserted into the input question.

10. The method of claim 6, wherein the elided information content in the input question comprises at least one of an elided focus, an elided verb, or an elided noun phrase.

11. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device implementing a Question Answering (QA) system, causes the computing device to:

receive, by the QA system, an input question for which an answer is sought by a submitter of the input question;

input, by the QA system, the input question to a pre-processor flow path comprising at least one pre-processor;

transform, by the at least one pre-processor, the input question into a transformed question by correcting errors in a formulation of the input question that are determined to be detrimental to efficient and accurate processing of the input question by a QA system pipeline of the QA system;

input the transformed question to the QA system pipeline of the QA system; and

process the transformed question via the QA system pipeline to generate and output at least one answer to the input question.

12. The computer program product of claim 11, wherein the at least one pre-processor comprises a superfluous information pre-processor that identifies superfluous information content in the input question and transforms the superfluous information content.

13. The computer program product of claim 12, wherein the computer readable program further causes the computing device to transform the input question at last by:

identifying at least one of a term, phrase, or language pattern indicative of superfluous information content in the input question that is not relevant to the underlying question being asked by the input question;

removing the superfluous information content from the input question to generate a modified input question;

in response to determining that the modified input question requires reformatting, reformatting the input question into a logically correct input question; and

outputting the logically correct input question as the transformed question.

14. The computer program product of claim 13, wherein the computer readable program further causes the computing device to identify at least one of a term, phrase, or language pattern indicative of superfluous information content in the input question at least by identifying at least one of a vocative clause or a common phrase in the input question.

15. The computer program product of claim 12, wherein the superfluous information content in the input question comprises at least one of a vocative clause or a common phrase in the input question.

16. The computer program product of claim 11, wherein the at least one pre-processor comprises an elided information pre-processor that identifies elided information content associated with the input question and transforms the input question based on the identified elided information.

17. The computer program product of claim 16, wherein the computer readable program further causes the computing device to transform the input question at least by:

identifying at least one of a term, phrase, or language pattern indicative of elided information content in the input question that is relevant to the underlying question being asked by the input question and is missing in the input question;

inserting the elided information content into the input question to generate a modified input question;

in response to determining that the modified input question requires reformatting, reformatting the input question into a logically correct input question; and

outputting the logically correct input question as the transformed question.

18. The computer program product of claim 17, wherein the computer readable program further causes the computing device to insert the elided information content into the input question to generate a modified input question at least by analyzing at least one of a domain or a context associated with the input question to determine the elided information content to be inserted into the input question.

19. The computer program product of claim 16, wherein the elided information content in the input question comprises at least one of an elided focus, an elided verb, or an elided noun phrase.

20. An apparatus comprising:

a processor; and

a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor implementing a Question Answering (QA) system, cause the processor to;

receive, by the QA system, an input question for which an answer is sought by a submitter of the input question;

input, by the QA system, the input question to a pre-processor flow path comprising at least one pre-processor;

transform, by the at least one pre-processor, the input question into a transformed question by correcting errors in a formulation of the input question that are determined to be detrimental to efficient and accurate processing of the input question by a QA system pipeline of the QA system;

input the transformed question to the QA system pipeline of the QA system; and

process the transformed question via the QA system pipeline to generate and output at least one answer to the input question.