Natural language processing utilizing transaction based knowledge representation

• US 9,904,668 B2
• Filed: 07/05/2017
• Issued: 02/27/2018
• 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
• Primary Examiner(s): Riley, Marcus T
• Application Number:
  US15/641,904
• Publication Number:
  US 20170308521A1
• Time in Patent Office: 237 Days
• Field of Search: None
• CPC Class Codes: G06F 40/205 Parsing

Abstract

Mechanisms are provided for processing logical relationships in natural language content. A logical parse of a first parse of the natural language content is generated by identifying latent logical terms within the first parse indicative of logical relationships between elements of the natural language content. The logical parse comprises nodes and edges linking nodes. At least one knowledge value is associated with each node in the logical parse. The at least one knowledge value associated with at least a subset of the nodes in the logical parse is propagated to one or more other nodes in the logical parse based on propagation rules. The propagating of the at least one knowledge value generates transaction records in a transaction knowledgebase data structure. A reasoning operation is executed based on the transaction knowledgebase data structure.

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:

generating, by the data processing system, a logical parse hierarchical representation of a first parse of the natural language content by identifying latent logical terms within the first parse indicative of logical relationships between elements of the natural language content, wherein the logical parse hierarchical representation comprises nodes and edges linking nodes;

associating, by the data processing system, at least one knowledge value with each node in the logical parse hierarchical representation;

propagating, by the data processing system, the at least one knowledge value associated with at least a subset of the nodes in the logical parse hierarchical representation to one or more other nodes in the logical parse hierarchical representation based on propagation rules, wherein the propagating of the at least one knowledge value generates transaction records in a transaction knowledgebase data structure; and

performing, by the data processing system, a reasoning operation based on the transaction knowledgebase data structure, wherein the at least one knowledge value associated with each node in the logical parse hierarchical representation comprises a knowledge tuple having one or more supporting evidence measure knowledge values and a relevance knowledge value.

2. The method of claim 1, wherein propagating the at least one knowledge value comprises iteratively applying, by the data processing system, one or more knowledge reasoners to the logical parse hierarchical representation to propagate the at least one knowledge value of at least a subset of the nodes in the logical parse hierarchical representation to one or more other nodes in the logical parse hierarchical representation based on propagation rules, wherein each application of a knowledge reasoner generates a transaction record in the transaction knowledgebase data structure identifying nodes affected by the application of the knowledge reasoner.

3. The method of claim 2, wherein the one or more knowledge reasoners comprises at least one of an evidential support reasoner, a relevance reasoner, or a co-reference knowledge reasoner.

4. The method of claim 2, wherein the one or more knowledge reasoners are iteratively applied until no further logical deductions are able to be made by further application of the one or more knowledge reasoners.

5. The method of claim 4, wherein no further logical deductions being able to be made is detected by the detection of an empty transaction being generated in response to the one or more knowledge reasoners being applied.

6. The method of claim 2, wherein each transaction record is generated by exactly one invocation of a knowledge reasoner and stores data representing knowledge contributions from the one invocation of the knowledge reasoner.

7. The method of claim 2, wherein each transaction record stores identifiers of nodes in the logical parse hierarchical representation affected by application of a corresponding knowledge reasoner and knowledge value states of the nodes corresponding to the identifiers of nodes in the transaction record.

8. The method of claim 2, wherein performing a reasoning operation based on the transaction knowledgebase data structure comprises determining a current state of a node in the logical parse hierarchical representation by retrieving a most recent transaction that affected the node from the transaction knowledgebase data structure.

9. The method of claim 2, wherein performing a reasoning operation based on the transaction knowledgebase data structure comprises analyzing the transaction knowledgebase data structure to identify individual contributions to the transaction knowledgebase data structure by a specific knowledge reasoner based on the transaction records stored in the transaction knowledgebase data structure.

10. 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:

generate a logical parse hierarchical representation of a first parse of natural language content by identifying latent logical terms within the first parse indicative of logical relationships between elements of the natural language content, wherein the logical parse hierarchical representation comprises nodes and edges linking nodes;

associate at least one knowledge value with each node in the logical parse hierarchical representation;

propagate the at least one knowledge value associated with at least a subset of the nodes in the logical parse hierarchical representation to one or more other nodes in the logical parse hierarchical representation based on propagation rules, wherein the propagating of the at least one knowledge value generates transaction records in a transaction knowledgebase data structure; and

perform a reasoning operation based on the transaction knowledgebase data structure, wherein the at least one knowledge value associated with each node in the logical parse hierarchical representation comprises a knowledge tuple having one or more supporting evidence measure knowledge values and a relevance knowledge value.

11. The computer program product of claim 10, wherein the computer readable program further causes the computing device to propagate the at least one knowledge value at least by iteratively applying, by the computing device, one or more knowledge reasoners to the logical parse hierarchical representation to propagate the at least one knowledge value of at least a subset of the nodes in the logical parse hierarchical representation to one or more other nodes in the logical parse hierarchical representation based on propagation rules, wherein each application of a knowledge reasoner generates a transaction record in the transaction knowledgebase data structure identifying nodes affected by the application of the knowledge reasoner.

12. The computer program product of claim 11, wherein the one or more knowledge reasoners comprises at least one of an evidential support reasoner, a relevance reasoner, or a co-reference knowledge reasoner.

13. The computer program product of claim 11, wherein the one or more knowledge reasoners are iteratively applied until no further logical deductions are able to be made by further application of the one or more knowledge reasoners.

14. The computer program product of claim 13, wherein no further logical deductions being able to be made is detected by the detection of an empty transaction being generated in response to the one or more knowledge reasoners being applied.

15. The computer program product of claim 11, wherein each transaction record is generated by exactly one invocation of a knowledge reasoner and stores data representing knowledge contributions from the one invocation of the knowledge reasoner.

16. The computer program product of claim 11, wherein each transaction record stores identifiers of nodes in the logical parse hierarchical representation affected by application of a corresponding knowledge reasoner and knowledge value states of the nodes corresponding to the identifiers of nodes in the transaction record.

17. The computer program product of claim 11, wherein the computer readable program further causes the computing device to perform a reasoning operation based on the transaction knowledgebase data structure at least by determining a current state of a node in the logical parse hierarchical representation by retrieving a most recent transaction that affected the node from the transaction knowledgebase data structure.

18. The computer program product of claim 11, wherein the computer readable program further causes the computing device to perform a reasoning operation based on the transaction knowledgebase data structure at least by analyzing the transaction knowledgebase data structure to identify individual contributions to the transaction knowledgebase data structure by a specific knowledge reasoner based on the transaction records stored in the transaction knowledgebase data structure.