Natural Language Processing Utilizing Logical Tree Structures

• US 20160098394A1
• Filed: 10/06/2014
• Published: 04/07/2016
• Est. Priority Date: 10/06/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bruno, Nicholas V., Smythe, Jared M.D., Woods, Eric
• Granted Patent: US 9,665,564 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Mechanisms are provided for processing logical relationships in natural language content. Natural language content is received, upon which a reasoning operation is to be performed. A first parse representation of the natural language content is generated, by a parser, by performing natural language processing on the natural language content. A logical parse of the first parse is generated by identifying latent logical operators within the first parse indicative of logical relationships between elements of the natural language content. A reasoning operation on the logical parse is executed to generate a knowledge output indicative of knowledge associated with one or more of the logical relationships between elements of the natural language content.

Claims

1. A method, in a data processing system comprising a processor and a memory, for processing logical relationships in natural language content, the method comprising:

receiving, by the data processing system, natural language content upon which a reasoning operation is to be performed;

generating, by a parser implemented in the data processing system, a first parse representation of the natural language content by performing natural language processing on the natural language content;

generating, by the data processing system, a logical parse of the first parse by identifying latent logical operators within the first parse indicative of logical relationships between elements of the natural language content; and

performing, by the data processing system, a reasoning operation on the logical parse to generate a knowledge output indicative of knowledge associated with one or more of the logical relationships between elements of the natural language content.

2. The method of claim 1, wherein generating, by the parser, the first parse representation of the natural language content comprises:

parsing the natural language content into one or more atomic logical terms that lack explicit or implicit logic; and

connecting the one or more atomic logical terms by logical operators in the first parse representation to specify a logical relationship between the one or more atomic logical terms.

3. The method of claim 1, wherein generating a logical parse of the first parse representation of the natural language content comprises:

inputting the first parse into a logical parser implemented in the data processing system; and

normalizing, by the logical parser, logical triggers in the first parse indicative of a logical relationship to a predefined set of logical operators to thereby generate the logical parse.

4. The method of claim 3, wherein generating a logical parse of the first parse representation of the natural language content further comprises at least one of expanding negation contractions or conjoining comma-delimited lists.

5. The method of claim 3, wherein generating a logical parse of the first parse representation of the natural language content further comprises:

simplifying the logical parse by removing redundancy in the logical parse to generate a simplified logical parse that replaces the logical parse, wherein simplifying the logical parse comprises implementing DeMorgan'"'"'s Law, Factorization, and Disjunctive Syllogism.

6. The method of claim 3, wherein the set of predefined logical operators comprises a logic AND operator, a logic OR operator, and a logic NOT operator.

7. The method of claim 3, wherein normalizing logical triggers in the first part comprises identifying terms or phrases within the first parse that match one or more elements of a set of logical triggers, and wherein the set of logical triggers comprises at least one of negation identifiers, noun relations, or modifiers that are indicative of a logical relationship.

8. The method of claim 1, wherein generating a logical parse of the first parse representation of the natural language content further comprises:

processing each logical operator in the logical parse within a scope of the logical operator, where upper scopes are processed before lower scopes to thereby elevate a logical operator above a top unprocessed node in the logical parse and set a sub-tree under the logical operator as a first argument of the logical operator and a copy of the sub-tree as a second argument of the logical operator.

9. The method of claim 1, wherein processing each of the logical operators in the logical parse further comprises processing logical OR operators first, processing logical AND operators second, and processing logical NOT operators last.

10. The method of claim 1, wherein generating a logical parse of the first parse representation of the natural language content further comprises:

performing a primitivization process on the logical parse to break down sub-tree structures into primitive leaf nodes connected to the logical operators, wherein the primitivization process implements a one modifier rule and verb exception to the one modifier rule.

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 data processing system, causes the data processing system to:

receive a natural language content upon which a reasoning operation is to be performed;

generate a first parse representation of the natural language content by performing natural language processing on the natural language content;

generate a logical parse of the first parse by identifying latent logical operators within the first parse indicative of logical relationships between elements of the natural language content; and

perform a reasoning operation on the logical parse to generate a knowledge output indicative of knowledge associated with one or more of the logical relationships between elements of the natural language content.

12. The computer program product of claim 11, wherein the computer readable program further causes the data processing system to generate the first parse representation of the natural language content at least by:

parsing the natural language content into one or more atomic logical terms that lack explicit or implicit logic; and

connecting the one or more atomic logical terms by logical operators in the first parse representation to specify a logical relationship between the one or more atomic logical terms.

13. The computer program product of claim 11, wherein the computer readable program further causes the data processing system to generate a logical parse of the first parse representation of the natural language content at least by:

inputting the first parse into a logical parser implemented in the data processing system; and

normalizing, by the logical parser, logical triggers in the first parse indicative of a logical relationship to a predefined set of logical operators to thereby generate the logical parse.

14. The computer program product of claim 13, wherein the computer readable program further causes the data processing system to generate a logical parse of the first parse representation of the natural language content at least by at least one of expanding negation contractions or conjoining comma-delimited lists.

15. The computer program product of claim 13, wherein the computer readable program further causes the data processing system to generate a logical parse of the first parse representation of the natural language content at least by:

simplifying the logical parse by removing redundancy in the logical parse to generate a simplified logical parse that replaces the logical parse, wherein simplifying the logical parse comprises implementing DeMorgan'"'"'s Law, Factorization, and Disjunctive Syllogism.

16. The computer program product of claim 13, wherein the set of predefined logical operators comprises a logic AND operator, a logic OR operator, and a logic NOT operator.

17. The computer program product of claim 13, wherein the computer readable program further causes the data processing system to normalize logical triggers in the first part at least by identifying terms or phrases within the first parse that match one or more elements of a set of logical triggers, and wherein the set of logical triggers comprises at least one of negation identifiers, noun relations, or modifiers that are indicative of a logical relationship.

18. The computer program product of claim 11, wherein the computer readable program further causes the data processing system to generate a logical parse of the first parse representation of the natural language content at least by:

processing each logical operator in the logical parse within a scope of the logical operator, where upper scopes are processed before lower scopes to thereby elevate a logical operator above a top unprocessed node in the logical parse and set a sub-tree under the logical operator as a first argument of the logical operator and a copy of the sub-tree as a second argument of the logical operator.

19. The computer program product of claim 11, wherein the computer readable program further causes the data processing system to process each of the logical operators in the logical parse at least by processing logical OR operators first, processing logical AND operators second, and processing logical NOT operators last.

20. The computer program product of claim 11, wherein the computer readable program further causes the data processing system to generate a logical parse of the first parse representation of the natural language content at least by:

performing a primitivization process on the logical parse to break down sub-tree structures into primitive leaf nodes connected to the logical operators, wherein the primitivization process implements a one modifier rule and verb exception to the one modifier rule.

21. An apparatus comprising:

a processor; and

a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to;

receive a natural language content upon which a reasoning operation is to be performed;

generate a first parse representation of the natural language content by performing natural language processing on the natural language content;

generate a logical parse of the first parse by identifying latent logical operators within the first parse indicative of logical relationships between elements of the natural language content; and

perform a reasoning operation on the logical parse to generate a knowledge output indicative of knowledge associated with one or more of the logical relationships between elements of the natural language content.