Machine learning to integrate knowledge and augment natural language processing

Patent metadata

Field	Value
Publication number	US-10776586-B2
Application number	US-201815866702-A
Country	US
Kind code	B2
Filing date	Jan 10, 2018
Priority date	Jan 10, 2018
Publication date	Sep 15, 2020
Grant date	Sep 15, 2020

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.

First claim

What is claimed is:

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 leverage a machine learning model (MLM) manager, including:
receive natural language (NL) input and query the input against context, wherein the context includes a specified knowledge graph (KG) and corresponding blockchain (BC) ledger; and
extract one or more triplets from the specified KG, wherein each triplet includes a subject, object, and a relationship;
obtain a BC identifier; and
identify a corresponding veracity value in the BC ledger;
generate a list of triplets together with the identified veracity values, and sort the generated list of triplets based the veracity values; and

the MLM manager to augment one or more MLMs with the received natural language input.

2. The system of claim 1 , further comprising the knowledge engine to identify a conflict between the NL input and one or more of the triplets in the generated list, and further comprising the knowledge engine to correct the received NL input by replacement with an identified triplet in the generated list.

3. The system of claim 1 , further comprising the knowledge engine to identify a match between the NL input and at least one of the triplets in the generated list, and further comprising the knowledge engine to create an entry of the NL input in the KG and corresponding BC ledger.

4. The system of claim 1 , further comprising the knowledge engine to identify a conflict between the NL input and at least one of the triplets in the generated list, and further comprising the knowledge engine to sort the triplets in the generated list with a select component of the identified veracity value, and return the triplet in the sorted list corresponding to the selected veracity value component.

5. The system of claim 1 , wherein the knowledge engine identifies an immutable factor, and a conflict between the NL input and at least one entry in the generated list associated with the immutable factor, and further comprising the knowledge engine to return the triplet associated from the list entry having the immutable factor and a corresponding BC identifier.

6. The system of claim 1 , further comprising the knowledge engine to identify a partial match between the NL input and at least one of the triplets in the generated list, and further comprising the knowledge engine to create a new entry in the KG and a corresponding BC ledger, and to connect the created new entry with an entry corresponding to the partial match.

7. The system of claim 1 , wherein the generated list of triplets is empty, and further comprising:
the knowledge engine to: create a new triplet corresponding to the received NL input, assign a veracity score to the created triplet, create an entry for the new triplet in the KG and create a corresponding entry for the new triplet in the BC ledger.

8. The system of claim 7 , further comprising the knowledge engine to store a BC identifier associated with the BC ledger entry with the new triplet in the KG, and to store the assigned veracity score with the BC ledger entry.

9. A computer program product to process natural language (NL), the computer program product comprising a computer readable storage device having program code embodied therewith, the program code executable by a processing unit to:
leverage a machine learning model (MLM), including:
receive a NL input and query the input against context, wherein the context includes a specified knowledge graph (KG) and corresponding blockchain (BC) ledger; and
extract one or more triplets from the specified KG, wherein each triplet includes a subject, object, and a relationship;

for each extracted triplet, program code is provided to:
obtain a BC identifier; and
identify a corresponding veracity value in the BC ledger;

generate a list of triplets together with the identified veracity values, and sort the generated list of triplets based on the identified veracity values; and
augment one or more MLMs with the received NL input.

10. The computer program product of claim 9 , further comprising program code to identify a conflict between the NL input and one or more of the triplets in generated list, and further comprising program code to:
sort the triplets in the generated list with a select component of the identified veracity value, and return the triplet in the sorted list corresponding to the selected veracity value; and
replace the received NL input with an identified triplet in the sorted list.

11. The computer program product of claim 9 , further comprising program code to identify a match between the NL input and at least one of the triplets in the generated list, and further comprising program code to:
create an entry of a triplet created from the NL input in the KG and corresponding BC ledger.

12. The computer program product of claim 9 , further comprising program code to identify a conflict between the NL input and at least one entry in the generated list associated with an immutable factor, and further comprising program code to:
return the triplet associated from the list entry having the immutable factor and a corresponding BC identifier for the returned triplet.

13. The computer program product of claim 9 , further comprising program code to identify a partial match between the NL input and at least one of the triplets in the generated list, and further comprising program code to:
create a new entry in the KG and a corresponding BC ledger, and to connect the created new entry with an entry corresponding to the partial match.

14. The computer program product of claim 9 , wherein the generated list of triplets is empty, and further comprising program code to:
create a new triplet corresponding to the received NL input;
assign a veracity score to the created triplet; and
create an entry for the new triplet in the KG and a corresponding entry for the new triplet in the BC ledger.

15. A method for processing natural language (NL), comprising:
receiving a natural language input and querying the input against context, wherein the context includes a specified knowledge graph (KG) and corresponding blockchain (BC) ledger;
extracting one or more triplets from the specified KG, wherein each triplet includes a subject, object, and a relationship;
for each extracted triplet, obtaining a BC identifier identifying a corresponding veracity value in the BC ledger;
generating a list of triplets together with the identified veracity values, and sorting the generated list of triplets based on the identified veracity values; and
augmenting one or more machine learning models MLMs) with the received natural language (NL) input.

16. The method of claim 15 , wherein augmenting the NL input identifies a conflict between the NL input and one or more of the triplets in generated list, and further comprising:
sorting the triplets in the generated list with a select component of the identified veracity value, and returning the triplet in the sorted list corresponding to the selected veracity value; and
replacing the received NL input with an identified triplet in the sorted list.

17. The method of claim 15 , wherein augmenting the NL input identifies a match between the NL input and at least one of the triplets in the generated list, and further comprising:
creating an entry of the NL input in the KG and corresponding BC ledger.

18. The method of claim 15

Assignees

• IBM

Inventors

• Bacarella David
• Barnebee Iv James H
• Lawrence Nicholas
• Patel Sumit

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• H04L9/50

  using hash chains, e.g. blockchains or hash trees · CPC title

• G06F40/20 Primary

  Natural language analysis (semantic analysis of natural language G06F40/30) · CPC title

• G06F40/40 Primary

  Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30) · CPC title