Correction of misspellings in QA system

• US 10,803,242 B2
• Filed: 10/26/2018
• Issued: 10/13/2020
• Est. Priority Date: 10/26/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Boxwell, Stephen A., Frost, Keith G., Vernier, Stanley J., Brake, Kyle M.
• Primary Examiner(s): Baderman, Scott T
• Assistant Examiner(s): Smith, Benjamin
• Application Number:
  US16/171,782
• Publication Number:
  US 20200134010A1
• Time in Patent Office: 718 Days
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 40/20 Natural language analysis semantic analysis of natural language G06F40/30
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/153 using recognition of characters or words
• CPC Class Codes: G06V 30/268 Lexical context
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

Embodiments provide a computer implemented method for identifying and correcting a misspelling in a question answering (QA) system, wherein the QA system is coupled to a document corpus, and the document corpus includes a plurality of documents related to a particular domain. The method includes the following steps: receiving an input question and a plurality of passages, wherein the plurality of passages are extracted from the document corpus by the QA system; providing at least one alternate form for each token extracted from the input question and the plurality of passages; identifying at least one misspelled token; and scoring at least one alternate form of each identified misspelled token.

Claims

1. A computer implemented method for identifying and correcting a misspelling in a question answering (QA) system, wherein the QA system is coupled to a document corpus, and the document corpus includes a plurality of documents related to a particular domain, the method comprising:

receiving, by a processor coupled to one or more user devices, an input question and a plurality of passages, wherein the plurality of passages are extracted from the document corpus by the QA system;

providing, by the processor, at least one alternate form for each token extracted from the input question and the plurality of passages;

the step of providing at least one alternate form for each token further comprising;

providing, by the processor, a first substitution confusion matrix for the input question through a speech to text process, wherein a first group of characters are substituted with a first single character in the first substitution confusion matrix, wherein in the first substitution confusion matrix, a first modified Levenshtein distance value between the character “

e” and

the character “

o”

is 0, a first modified Levenshtein distance value between the character “

i” and

the character “

t”

is 0, a first modified Levenshtein distance value between the character “

a” and

the character “

o”

is 0, and a first modified Levenshtein distance value between the character “

rn” and

the character “

m”

is 0;

providing, by the processor, the first modified Levenshtein distance value for each alternate form of each token extracted from the input question;

providing, by the processor, a second substitution confusion matrix for the plurality of passages through an optical character recognition process, wherein a second group of characters are substituted with a second single character in the second substitution confusion matrix, wherein in the second substitution confusion matrix, a second modified Levenshtein distance value between the character “

sh” and

the character “

s”

is 0, a second modified Levenshtein distance value between the character “

ai” and

the character “

e”

is 0, a second modified Levenshtein distance value between the character “

ed” and

the character “

te”

is 0, a second modified Levenshtein distance value between the character “

s” and

the character “

th”

is 0, and a second modified Levenshtein distance value between the character “

t” and

the character “

k”

is 0;

providing, by the processor, the second modified Levenshtein distance value for each alternate form of each token extracted from the plurality of passages;

identifying, by the processor, at least one misspelled token from the input question and the plurality of passages; and

scoring, by the processor, at least one alternate form of each identified misspelled token.

2. The method as recited in claim 1, wherein the at least one alternate form includes a synonym and a corrected spelling of the at least one misspelled token.

3. A computer program product for identifying and correcting a misspelling in a question answering (QA) system, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

receive, by a processor coupled to one or more user devices, an input question and a plurality of passages, wherein the plurality of passages are extracted from the document corpus by the QA system;

provide, by the processor, at least one alternate form for each token extracted from the input question and the plurality of passages;

the step of providing at least one alternate form for each token further causes the processor to;

provide, by the processor, a first substitution confusion matrix for the input question through a speech to text process, wherein a first group of characters are substituted with a first single character in the first substitution confusion matrix, wherein in the first substitution confusion matrix, a first modified Levenshtein distance value between the character “

e” and

the character “

o”

is 0, a first modified Levenshtein distance value between the character “

i” and

the character “

t”

is 0, a first modified Levenshtein distance value between the character “

a” and

the character “

o”

is 0, and a first modified Levenshtein distance value between the character “

rn” and

the character “

m”

is 0;

provide, by the processor, the first modified Levenshtein distance value for each alternate form of each token extracted from the input question;

provide, by the processor, a second substitution confusion matrix for the plurality of passages through an optical character recognition process, wherein a second group of characters are substituted with a second single character in the second substitution confusion matrix, wherein in the second substitution confusion matrix, a second modified Levenshtein distance value between the character “

sh” and

the character “

s”

is 0, a second modified Levenshtein distance value between the character “

ai” and

the character “

e”

is 0, a second modified Levenshtein distance value between the character “

ed” and

the character “

te”

is 0, a second modified Levenshtein distance value between the character “

s” and

the character “

th”

is 0, and a second modified Levenshtein distance value between the character “

t” and

the character “

k”

is 0;

provide, by the processor, the second modified Levenshtein distance value for each alternate form of each token extracted from the plurality of passages;

identify, by the processor, at least one misspelled token and the plurality of passages; and

score, by the processor, at least one alternate form of each identified misspelled token.

4. The computer program product as recited in claim 3, wherein the at least one alternate form includes a synonym and a corrected spelling of the at least one misspelled token.

5. A system for identifying and correcting a misspelling in a question answering (QA) system, comprising:

a processor configured to;

receive, by a processor coupled to one or more user devices, an input question and a plurality of passages, wherein the plurality of passages are extracted from the document corpus by the QA system;

provide, by the processor, at least one alternate form for each token extracted from the input question and the plurality of passages;

the step of providing at least one alternate form for each token further configures the processor to;

provide, by the processor, a first substitution confusion matrix for the input question through a speech to text process, wherein a first group of characters are substituted with a first single character in the first substitution confusion matrix, wherein in the first substitution confusion matrix, a first modified Levenshtein distance value between the character “

e” and

the character “

o”

is 0, a first modified Levenshtein distance value between the character “

i” and

the character “

t”

is 0, a first modified Levenshtein distance value between the character “

a” and

the character “

o”

is 0, and a first modified Levenshtein distance value between the character “

rn” and

the character “

m”

is 0;

provide, by the processor, the first modified Levenshtein distance value for each alternate form of each token extracted from the input question;

provide, by the processor, a second substitution confusion matrix for the plurality of passages through an optical character recognition process, wherein a second group of characters are substituted with a second single character in the second substitution confusion matrix, wherein in the second substitution confusion matrix, a second modified Levenshtein distance value between the character “

sh” and

the character “

s”

is 0, a second modified Levenshtein distance value between the character “

ai” and

the character “

e”

is 0, a second modified Levenshtein distance value between the character “

ed” and

the character “

te”

is 0, a second modified Levenshtein distance value between the character “

s” and

the character “

th”

is 0, and a second modified Levenshtein distance value between the character “

t” and

the character “

k”

is 0;

provide, by the processor, the second modified Levenshtein distance value for each alternate form of each token extracted from the plurality of passages;

identify, by the processor, at least one misspelled token; and

score, by the processor, at least one alternate form of each identified misspelled token.

6. The system as recited in claim 5, wherein the at least one alternate form includes a synonym and a corrected spelling of the at least one misspelled token.