Machine learning to integrate knowledge and natural language processing

• US 10,423,726 B2
• Filed: 01/10/2018
• Issued: 09/24/2019
• Est. Priority Date: 01/10/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bacarella, David, Barnebee, IV, James H., Lawrence, Nicholas, Patel, Sumit
• Primary Examiner(s): Hoque, Nafiz E
• Application Number:
  US15/866,698
• Publication Number:
  US 20190213258A1
• Time in Patent Office: 622 Days
• CPC Class Codes: G06F 16/3331 Query processing
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/226 Validation
• CPC Class Codes: G06F 40/237 Lexical tools
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06N 5/043 Distributed expert systems; Blackboards
• CPC Class Codes: H04L 67/104 Peer-to-peer [P2P] networks
• CPC Class Codes: H04L 9/0637 Modes of operation, e.g. cipher block chaining [CBC], electronic codebook [ECB] or Galois/counter mode [GCM]
• CPC Class Codes: H04L 9/3239 involving non-keyed hash functions, e.g. modification detection codes [MDCs], MD5, SHA or RIPEMD
• CPC Class Codes: H04L 9/50 using hash chains, e.g. blockchains or hash trees

Abstract

A system, computer program product, and method are provided to automate a framework for knowledge graph based persistence of data, and to resolve temporal changes and uncertainties in the knowledge graph. Natural language understanding, together with one or more machine learning models (MLMs), is used to extract data from unstructured information, including entities and entity relationships. The extracted data is populated into a knowledge graph. As the KG is subject to change, the KG is used to create new and retrain existing machine learning models (MLMs). Weighting is applied to the populated data in the form of veracity value. Blockchain technology is applied to the populated data to ensure reliability of the data and to provide auditability to assess changes to the data.

Claims

1. A computer system comprising:

a processing unit operatively coupled to memory;

an artificial intelligence platform, in communication with the processing unit and memory;

a knowledge engine in communication with the processing unit to manage data, including;

extract data and a data relationship from data selected from the group consisting of;

structured data, unstructured data, and combinations thereof;

create an entry for the extracted data and data relationship in a knowledge graph (KG) and selectively store the extracted data and data relationship in the KG, including assign a veracity value to the stored data;

create an asset value entry in a blockchain (BC) ledger corresponding to the KG, the entry including the assigned veracity value;

create a BC identifier corresponding to the BC ledger entry; and

store the created BC identifier with the KG entry;

evaluate select data stored in the KG, including employ the BC identifier to determine provenance of the select data and to quantify the data; and

generate a list of the evaluated data, and sort the data in the generated list based on the assigned veracity value; and

a data element returned from the sorted list with a strongest veracity score.

2. The system of claim 1, further comprising the knowledge engine to:

create a first partition within the KG and populate and assign a first reliability value to first data in the first partition;

create a second partition within the KG and populate and assign a second reliability value to second data in the second partition, wherein the first and second reliability values are different.

3. The system of claim 2, further comprising the knowledge engine to automatically perform a veracity evaluation within the KG, including comparison of the first and second data.

4. The system of claim 1, further comprising the knowledge engine to:

establish a link between two knowledge graphs, including compare and evaluate data elements in a second KG with data elements in a first KG, and selectively replace data elements based on a value selected from the group consisting of;

reliability, feedback, and combinations thereof.

5. The system of claim 4, further comprising the knowledge engine to maintain a structure of the KG constant following establishment of the link between the first KG and the second KG.

6. The system of claim 1, wherein the data is stored in a node in the KG and the relationship is represented as an edge connecting two nodes, each node having a node level veracity value and each relationship having a relationship veracity value, wherein the relationship value is calculated based on the veracity values of the nodes in the relationship.

7. A computer program product to process natural language, the computer program product comprising a computer readable storage device having program code embodied therewith, the program code executable by a processing unit to:

store data in a knowledge graph (KG), comprising;

extract data and a data relationship from data selected from the group consisting of;

structured data, unstructured data, and combinations thereof;

create an entry in the KG and selectively store the extracted data and data relationship in the KG, including assign a veracity value to the stored data;

create an asset value entry in a blockchain (BC) ledger corresponding to the KG, the entry including the assigned veracity value;

create a BC identifier corresponding to the BC ledger entry; and

store the created BC identifier with the KG entry;

evaluate select data stored in the KG, including employ the BC identifier to determine provenance of the select data and to quantify the data;

generate a list of the evaluated data, and sort the data in the generated list based on the assigned veracity value; and

generate an outcome, wherein the outcome is a data element returned from the sorted list with a strongest veracity score.

8. The computer program product of claim 7, further comprising program code to:

create a first partition within the KG and populate and assign a first reliability value to first data in the first partition;

create a second partition within the KG and populate and assign a second reliability value to second data in the second partition, wherein the first and second reliability values are different.

9. The computer program product of claim 8, further comprising program code to automatically perform a veracity evaluation within the KG, including comparison of the first and second data.

10. The computer program product of claim 7, further comprising program code to:

establish a link between two knowledge graphs, including compare and evaluate data elements in a second KG with data elements in a first KG, and selectively replace data elements based on a value selected from the group consisting of;

reliability, feedback, and combinations thereof.

11. The computer program product of claim 10, further comprising program code to maintain a structure of the KG constant following establishment of the link between the first KG and the second KG.

12. The computer program product of claim 7, wherein the data is stored in a node in the KG and the relationship is represented as an edge connecting two nodes, each node having a node level veracity value and each relationship having a relationship veracity value, wherein the relationship value is calculated based on the veracity values of the nodes in the relationship.

13. A method for processing natural language, comprising:

storing data in a knowledge graph (KG), comprising;

extracting data and a data relationship from data selected from the group consisting of;

structured data, unstructured data, and combinations thereof;

creating an entry in the KG and selectively storing the extracted data and data relationship in the KG, including assigning a veracity value to the stored data;

creating an asset value entry in a blockchain (BC) ledger corresponding to the KG, the entry including the assigned veracity value;

creating a BC identifier corresponding to the BC ledger entry; and

storing the created BC identifier with the KG entry;

evaluating select data stored in the KG, including employing the BC identifier to determine provenance of the select data and to quantify the data;

generating a list of the evaluated data, and sorting the data in the generated list based on the assigned veracity value; and

a data element returned from the sorted list with a strongest veracity score.

14. The method of claim 13, further comprising

creating a first partition within the KG and populating and assigning a first reliability value to first data in the first partition;

creating a second partition within the KG and populating and assigning a second reliability value to second data in the second partition, wherein the first and second reliability values are different.

15. The method of claim 14, further comprising automatically performing a veracity evaluation within the KG, including comparison of the first and second data.

16. The method of claim 13, further comprising:

establishing a link between two knowledge graphs, including comparing and evaluating data elements in a second KG with data elements in a first KG, and selectively replacing data elements based on a value selected from the group consisting of;

reliability, feedback, and combinations thereof.

17. The method of claim 16, further comprising maintaining a structure of the KG constant following establishment of the link between the first KG and the second KG.

18. The method of claim 13, wherein the data is stored in a node in the KG and the relationship is represented as an edge connecting two nodes, each node having a node level veracity value and each relationship having a relationship veracity value, wherein the relationship value is calculated based on the veracity values of the nodes in the relationship.