Predicate parses using semantic knowledge

• US 10,102,200 B2
• Filed: 08/25/2016
• Issued: 10/16/2018
• Est. Priority Date: 08/25/2016
• Status: Active Grant

• Current Assignee: Merative US LP
• Original Assignee: International Business Machines Corporation
• Inventors: Bull, Brendan C., Contreras, David, Felt, Paul L.
• Primary Examiner(s): Ky, Kevin
• Application Number:
  US15/246,903
• Publication Number:
  US 20180060304A1
• Time in Patent Office: 782 Days
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A mechanism is provided for improving predicate parses (or logical representations of a passage) using semantic knowledge. In response to encountering an ambiguous decision point during a syntactic analysis of a portion of natural language content, a candidate meaning of the ambiguous decision point is generated. Characteristics of the ambiguous decision point are evaluated based on a semantic knowledge base to determine a semantic meaning associated with the ambiguous decision point. A determination is made as to whether the semantic meaning supports or refutes the candidate meaning. In response to determining that the semantic meaning refutes the candidate meaning, the candidate meaning of the ambiguous decision point is overridden based on the semantic meaning to include the semantic meaning as a final meaning for the ambiguous decision point. The portion of natural language content is then processed based on the final meaning for the ambiguous decision point.

Claims

1. A method, in a data processing system comprising at least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to implement a content analysis system to perform predicate parsing using semantic knowledge in order to identify how two or more concepts are logically conjoined, the method which comprises:

performing, by the content analysis system, a syntactic analysis of a portion of natural language content;

in response to encountering an ambiguous decision point during the syntactic analysis of the portion of natural language content, generating, by the content analysis system, a candidate meaning of the ambiguous decision point, wherein the ambiguous decision point is one or more of a mutually exclusive pair, a list of mutually exclusive items, an ambiguous conjunction, a context sensitive feature expansion, or sentence grouping;

evaluating, by the content analysis system, characteristics of the ambiguous decision point based on a semantic knowledge base to determine a semantic meaning associated with the ambiguous decision point, wherein the evaluation of the characteristics of the ambiguous decision point further comprises;

determining, by the content analysis system, whether a set of concepts associated with the ambiguous decision point are distillable to a core concept meaning;

determining, by the content analysis system, whether an obvious lexical relationship exists between the set of concepts and a main concept of the ambiguous decision point;

responsive to the set of concepts being two or more concepts, determining, by the content analysis system, whether the two or more concepts are distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point; and

responsive to determining that the set of concepts associated with the ambiguous decision point fail to be distillable to the core concept meaning, responsive to determining that no obvious lexical relationship exists between the set of concepts and the main concept of the ambiguous decision point, or responsive to determining that the two or more concepts fail to be distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point, generating, by the content analysis system, a notification indicating that the portion of natural language content should be investigated for a possible ambiguous decision point;

determining, by the content analysis system, whether the semantic meaning supports or refutes the candidate meaning of the ambiguous decision point;

in response to determining that the semantic meaning refutes the candidate meaning, overriding, by the content analysis system, the candidate meaning of the ambiguous decision point based on the semantic meaning to include the semantic meaning as a final meaning for the ambiguous decision point; and

processing, by the content analysis system, the portion of natural language content based on the final meaning for the ambiguous decision point.

2. The method of claim 1, further comprising:

in response to determining that the semantic meaning supports the candidate meaning, maintaining, by the content analysis system, the candidate meaning of the ambiguous decision point as the final meaning for the ambiguous decision point.

3. The method of claim 1, further comprising:

in response to a failure to encounter the ambiguous decision point during the syntactic analysis of the portion of natural language content, providing, by the content analysis system a notification that no ambiguous decision point was encountered.

4. The method of claim 1, wherein the semantic knowledge base ranks concepts and groups of concepts on various semantic relationships comprising one or more of similarity, interchangeability, semantic interpretation, or being part of a larger set.

5. The method of claim 1, wherein the semantic knowledge base ranks concepts and groups of concepts on various semantic relationships comprising one or more of parenthetical understanding, obvious lexical relationship, or parenthetical distilment.

6. 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, causes the computing device to implement a content analysis system to perform predicate parsing using semantic knowledge in order to identify how two or more concepts are logically conjoined, and further causes the computing device to:

perform, by the content analysis system, a syntactic analysis of a portion of natural language content;

in response to encountering an ambiguous decision point during the syntactic analysis of the portion of natural language content, generate, by the content analysis system, a candidate meaning of the ambiguous decision point, wherein the ambiguous decision point is one or more of a mutually exclusive pair, a list of mutually exclusive items, an ambiguous conjunction, a context sensitive feature expansion, or sentence grouping;

evaluate, by the content analysis system, characteristics of the ambiguous decision point based on a semantic knowledge base to determine a semantic meaning associated with the ambiguous decision point, wherein the computer program product to evaluate the characteristics of the ambiguous decision point further causes the computing device to;

determine, by the content analysis system, whether a set of concepts associated with the ambiguous decision point are distillable to a core concept meaning;

determine, by the content analysis system, whether an obvious lexical relationship exists between the set of concepts and a main concept of the ambiguous decision point;

responsive to the set of concepts being two or more concepts, determine, by the content analysis system, whether the two or more concepts are distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point; and

responsive to determining that the set of concepts associated with the ambiguous decision point fail to be distillable to the core concept meaning, responsive to determining that no obvious lexical relationship exists between the set of concepts and the main concept of the ambiguous decision point, or responsive to determining that the two or more concepts fail to be distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point, generate, by the content analysis system, a notification indicating that the portion of natural language content should be investigated for a possible ambiguous decision point;

determine, by the content analysis system, whether the semantic meaning supports or refutes the candidate meaning of the ambiguous decision point;

in response to determining that the semantic meaning refutes the candidate meaning, override, by the content analysis system, the candidate meaning of the ambiguous decision point based on the semantic meaning to include the semantic meaning as a final meaning for the ambiguous decision point; and

process, by the content analysis system, the portion of natural language content based on the final meaning for the ambiguous decision point.

7. The computer program product of claim 6, wherein the computer readable program further causes the computing device to:

in response to determining that the semantic meaning supports the candidate meaning, maintain, by the content analysis system, the candidate meaning of the ambiguous decision point as the final meaning for the ambiguous decision point.

8. The computer program product of claim 6, wherein the computer readable program further causes the computing device to:

in response to a failure to encounter the ambiguous decision point during the syntactic analysis of the portion of natural language content, provide, by the content analysis system, a notification that no ambiguous decision point was encountered.

9. The computer program product of claim 6, wherein the semantic knowledge base ranks concepts and groups of concepts on various semantic relationships comprising one or more of similarity, interchangeability, semantic interpretation, or being part of a larger set.

10. The computer program product of claim 6, wherein the semantic knowledge base ranks concepts and groups of concepts on various semantic relationships comprising one or more of parenthetical understanding, obvious lexical relationship, or parenthetical distilment.

11. 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 implement a content analysis system to perform predicate parsing using semantic knowledge in order to identify how two or more concepts are logically conjoined, and further cause the processor to;

perform, by the content analysis system, a syntactic analysis of a portion of natural language content;

in response to encountering an ambiguous decision point during the syntactic analysis of the portion of natural language content, generate, by the content analysis system, a candidate meaning of the ambiguous decision point, wherein the ambiguous decision point is one or more of a mutually exclusive pair, a list of mutually exclusive items, an ambiguous conjunction, a context sensitive feature expansion, or sentence grouping;

evaluate, by the content analysis system, characteristics of the ambiguous decision point based on a semantic knowledge base to determine a semantic meaning associated with the ambiguous decision point, wherein the instructions to evaluate the characteristics of the ambiguous decision point further causes the processor to;

determine, by the content analysis system, whether a set of concepts associated with the ambiguous decision point are distillable to a core concept meaning;

determine, by the content analysis system, whether an obvious lexical relationship exists between the set of concepts and a main concept of the ambiguous decision point;

responsive to the set of concepts being two or more concepts, determine, by the content analysis system, whether the two or more concepts are distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point; and

responsive to determining that the set of concepts associated with the ambiguous decision point fail to be distillable to the core concept meaning, responsive to determining that no obvious lexical relationship exists between the set of concepts and the main concept of the ambiguous decision point, or responsive to determining that the two or more concepts fail to be distillable to a single semantic entry that can be compared to the main concept of the ambiguous decision point, generate, by the content analysis system, a notification indicating that the portion of natural language content should be investigated for a possible ambiguous decision point;

determine, by the content analysis system, whether the semantic meaning supports or refutes the candidate meaning of the ambiguous decision point;

in response to determining that the semantic meaning refutes the candidate meaning, override, by the content analysis system, the candidate meaning of the ambiguous decision point based on the semantic meaning to include the semantic meaning as a final meaning for the ambiguous decision point; and

process, by the content analysis system, the portion of natural language content based on the final meaning for the ambiguous decision point.

12. The apparatus of claim 11, wherein the instructions further cause the processor to:

in response to determining that the semantic meaning supports the candidate meaning, maintain, by the content analysis system, the candidate meaning of the ambiguous decision point as the final meaning for the ambiguous decision point.

13. The apparatus of claim 11, wherein the instructions further cause the processor to:

in response to a failure to encounter the ambiguous decision point during the syntactic analysis of the portion of natural language content, provide, by the content analysis system, a notification that no ambiguous decision point was encountered.

14. The apparatus of claim 11, wherein the semantic knowledge base ranks concepts and groups of concepts on various semantic relationships comprising one or more of similarity, interchangeability, semantic interpretation, being part of a larger set, parenthetical understanding, obvious lexical relationship, or parenthetical distilment.