Disambiguation of the meaning of terms based on context pattern detection

• US 10,255,271 B2
• Filed: 02/06/2017
• Issued: 04/09/2019
• Est. Priority Date: 02/06/2017
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Alba, Alfredo, Drews, Clemens, Gruhl, Daniel F., Kato, Linda H., Kau, Christian B., Lewis, Neal R., Mendes, Pablo N., Nagarajan, Meenakshi, Ramakrishnan, Cartic
• Primary Examiner(s): Leland, III, Edwin S
• Application Number:
  US15/425,751
• Publication Number:
  US 20180225278A1
• Time in Patent Office: 792 Days
• Field of Search: 704 9
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A method, system and computer program product for disambiguating meaning of terms. A natural language processing system obtains a first corpora of words used in a first sense and a second corpora of the same words that are used in a second sense. Each of these corpora of words may be associated with different linguistic domains. The natural language processing system generates a first and a second set of patterns using both the first and second corpora of words, respectively. A question passage is then received by the natural language processing system. The natural language processing system examines a word of interest in the question passage. The user is then notified that the word of interest is being used in the first sense or the second sense in response to identifying the word of interest in only the first set of patterns or the second set of patterns, respectively.

Claims

1. A method for disambiguating meaning of terms, the method comprising:

obtaining a first corpora of words used in a first sense;

obtaining a second corpora of words used in a second sense;

generating, by a processor, a first set of patterns and a second set of patterns using said first corpora of words and said second corpora of words, respectively;

receiving a question passage;

examining, by said processor, a word of interest in said question passage; and

notifying, by said processor, a user that said word of interest is being used in said first sense or said second sense in response to identifying said word of interest in only said first set of patterns or said second set of patterns, respectively.

2. The method as recited in claim 1, wherein said first and second sets of patterns are generated using an iteratively supervised lexicon induction algorithm.

3. The method as recited in claim 1 further comprising:

searching and identifying in said first and second sets of patterns for matching patterns; and

removing matching patterns from said first and second sets of patterns.

4. The method as recited in claim 1 further comprising:

notifying said user cannot conclude sense of usage of said word of interest in response to not identifying said word of interest in either said first set of patterns or said second set of patterns.

5. The method as recited in claim 1, wherein said first corpora of words is associated with a first linguistic domain and said second corpora of words is associated with a second linguistic domain.

6. The method as recited in claim 1 further comprising:

identifying said word of interest in both said first and second sets of patterns;

comparing said question passage to patterns in said first and second sets of patterns containing said word of interest;

generating scores for each pattern in said first and second sets of patterns containing said word of interest based on similarity of said question passage to each pattern in said first and second sets of patterns containing said word of interest;

identifying a pattern in said first and second sets of patterns containing said word of interest attributed with a highest score; and

notifying said user that said word of interest is being used in said first sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said first corpora of words and said highest value exceeding a threshold value or notifying said user that said word of interest is being used in said second sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said second corpora of words and said highest value exceeding said threshold value.

7. A computer program product for disambiguating meaning of terms, the computer program product comprising a computer readable storage medium having program code embodied therewith, the program code comprising the programming instructions for:

obtaining a first corpora of words used in a first sense;

obtaining a second corpora of words used in a second sense;

generating a first set of patterns and a second set of patterns using said first corpora of words and said second corpora of words, respectively;

receiving a question passage;

examining a word of interest in said question passage; and

notifying a user that said word of interest is being used in said first sense or said second sense in response to identifying said word of interest in only said first set of patterns or said second set of patterns, respectively.

8. The computer program product as recited in claim 7, wherein said first and second sets of patterns are generated using an iteratively supervised lexicon induction algorithm.

9. The computer program product as recited in claim 7, wherein the program code further comprises the programming instructions for:

searching and identifying in said first and second sets of patterns for matching patterns; and

removing matching patterns from said first and second sets of patterns.

10. The computer program product as recited in claim 7, wherein the program code further comprises the programming instructions for:

notifying said user cannot conclude sense of usage of said word of interest in response to not identifying said word of interest in either said first set of patterns or said second set of patterns.

11. The computer program product as recited in claim 7, wherein said first corpora of words is associated with a first linguistic domain and said second corpora of words is associated with a second linguistic domain.

12. The computer program product as recited in claim 7, wherein the program code further comprises the programming instructions for:

identifying said word of interest in both said first and second sets of patterns;

comparing said question passage to patterns in said first and second sets of patterns containing said word of interest;

generating scores for each pattern in said first and second sets of patterns containing said word of interest based on similarity of said question passage to each pattern in said first and second sets of patterns containing said word of interest;

identifying a pattern in said first and second sets of patterns containing said word of interest attributed with a highest score; and

notifying said user that said word of interest is being used in said first sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said first corpora of words and said highest value exceeding a threshold value or notifying said user that said word of interest is being used in said second sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said second corpora of words and said highest value exceeding said threshold value.

13. A system, comprising:

a memory unit for storing a computer program for disambiguating meaning of terms; and

a processor coupled to the memory unit, wherein the processor is configured to execute the program instructions of the computer program comprising;

obtaining a first corpora of words used in a first sense;

obtaining a second corpora of words used in a second sense;

generating a first set of patterns and a second set of patterns using said first corpora of words and said second corpora of words, respectively;

receiving a question passage;

examining a word of interest in said question passage; and

notifying a user that said word of interest is being used in said first sense or said second sense in response to identifying said word of interest in only said first set of patterns or said second set of patterns, respectively.

14. The system as recited in claim 13, wherein said first and second sets of patterns are generated using an iteratively supervised lexicon induction algorithm.

15. The system as recited in claim 13, wherein the program instructions of the computer program further comprise:

searching and identifying in said first and second sets of patterns for matching patterns; and

removing matching patterns from said first and second sets of patterns.

16. The system as recited in claim 13, wherein the program instructions of the computer program further comprise:

notifying said user cannot conclude sense of usage of said word of interest in response to not identifying said word of interest in either said first set of patterns or said second set of patterns.

17. The system as recited in claim 13, wherein said first corpora of words is associated with a first linguistic domain and said second corpora of words is associated with a second linguistic domain.

18. The system as recited in claim 13, wherein the program instructions of the computer program further comprise:

identifying said word of interest in both said first and second sets of patterns;

comparing said question passage to patterns in said first and second sets of patterns containing said word of interest;

generating scores for each pattern in said first and second sets of patterns containing said word of interest based on similarity of said question passage to each pattern in said first and second sets of patterns containing said word of interest;

identifying a pattern in said first and second sets of patterns containing said word of interest attributed with a highest score; and

notifying said user that said word of interest is being used in said first sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said first corpora of words and said highest value exceeding a threshold value or notifying said user that said word of interest is being used in said second sense in response to said pattern attributed with said highest value in said first and second sets of patterns being associated with said second corpora of words and said highest value exceeding said threshold value.