Generating and modifying ontologies for machine learning models

Patent metadata

Field	Value
Publication number	US-12579478-B2
Application number	US-202117532774-A
Country	US
Kind code	B2
Filing date	Nov 22, 2021
Priority date	Nov 20, 2020
Publication date	Mar 17, 2026
Grant date	Mar 17, 2026

Abstract

A method performed by a machine learning system that involves obtaining a first ontology that includes one or more labels. Each label is associated with a sample that includes text. The ML system is configured to use a particular label to retrieve one or more samples associated with the particular label. The method further involves receiving an identification of a label of a first ontology associated with a first machine learning model to share with a second ontology associated with a second machine learning model and sharing the label and the information with the second ontology. The method further involves training the second machine learning model using the shared information associated with the label.

First claim

What is claimed is:

1 . A method performed by a machine learning system comprising one or more processors, the method comprising:
obtaining, from a computer readable medium, a first ontology to train a first machine learning model received from a first computing device, the first ontology including one or more labels, each label of the one or more labels being associated with a sample that includes text, wherein the machine learning system is configured to use a particular label of the one or more labels to retrieve one or more samples associated with the particular label; receiving an identification of a label of the first ontology associated with the first machine learning model to share with a second ontology associated with a second machine learning model received from a second computing device, the label including information for training a model, the second ontology stored on the computer readable medium with the first ontology; sharing with the second ontology, the label and the information associated with the label while restricting the second ontology from accessing additional information of the first ontology to maintain the security of the additional information of the first ontology; and training the second machine learning model using the shared information associated with the label.

2 . The method of claim 1 wherein the information associated with the label includes a sample of text representative of a category of information for the label.

3 . The method of claim 2 wherein training the second machine learning model comprises using the sample of the text representative of the category of information for the label to identify the category of information in a document analyzed by the second model.

4 . The method of claim 1 wherein the information associated with the label includes at least one of text representative of a category of information for the label, a vector representation of the text representative of a category of information for the label, a feature representation of the label, and machine learning model weights.

5 . The method of claim 1 further storing the information associated with the label to restrict access to the information by a system utilizing processing data with the second model.

6 . The method of claim 5 wherein the system includes a device of a user processing data with the second model.

7 . The method of claim 1 wherein the second ontology is an existing ontology that includes one or more labels.

8 . The method of claim 1 , further comprising automatically including a new information for the label updated with respect to the first ontology with the shared label of the second ontology.

9 . The method of claim 1 further comprising automatically including a new information for the shared label of the second ontology with the label of the first ontology.

10 . The method of claim 1 , wherein the information related to the label of the first ontology can include one or more machine learning parameters and a weight associated with each of the one or more machine learning parameters.

11 . The method of claim 1 , wherein the information related to the label of the first ontology can include features associated with the label of the first ontology.

12 . A system comprising:
a first ontology stored in a first computer readable medium, the first ontology including a first label and information associated with the first label, the information utilized to train a first machine learning model received from a first computing device to identify a category of information represented by the first label; a second ontology stored in a second computer readable medium, wherein the first computer readable medium and the second computer readable medium are the same computer readable medium, and a processor to run computer executable instructions to cause the processor to:
generate a user interface, the user interface depicting a representation of the first ontology, and operable to share the first label and information associated with the first label with the second ontology while restricting the second ontology from accessing additional information of the first ontology to maintain the security of the additional information of the first ontology, whereby the second ontology includes the first label and information associated with the first label, the information utilized to train a second machine learning model received from a second computing device to identify the category of information represented by the first label.

13 . The system of claim 12 wherein the second ontology including a second label and information associated with the second label, the information associated with shared first label and the information associated with the second label utilized to train the second machine learning model to identify a category of information represented by the shared first label and a category of information associated with the second label.

14 . The system of claim 12 wherein access to the shared information of the second label is restricted based on a user ID.

15 . The system of claim 12 wherein the information associated with the first label includes a sample of text representative of a category of information for the label.

16 . The system of claim 15 the computer executable instructions to train the second machine learning model using the sample of the text representative of the category of information for the first label to identify the category of information in a document analyzed by the second model.

17 . The system of claim 12 wherein the information associated with the first label includes at least one of text representative of a category of information for the first label, a vector representation of the text representative of a category of information for the first label, a feature representation of the first label, and machine learning model weights.

18 . A method performed by a machine learning system comprising one or more processors, the method comprising:
obtaining, from a computer readable medium, a first ontology to train a first machine learning model received from a first computing device, the first ontology including one or more labels, each label of the one or more labels being associated with a sample that includes content, wherein the machine learning system is configured to use a particular label of the one or more labels to retrieve one or more samples associated with the particular label; receiving an identification of a label of the first ontology associated with the first machine learning model to share with a second ontology associated with a second machine learning model received from a second computing device, the label including information for training a model, the second ontology stored on the computer readable medium with the first ontology; sharing with the second ontology, the label and the information associated with the label while restricting the second ontology from accessing additional information of the first ontology to maintain the security of the additional information of the first ontology; and training the second machine learning model using the shared information associated with the label.

19 . The method of claim 18 wherein the content of the sample is textual content from one of the first computing device or the second computing device, and the system not allowing access by the second computing device to the sample if the sample was obtained from the first computing device, or the system not allowing access by the first c

Assignees

• Thomson Reuters Entpr Centre Gmbh

Inventors

• Hron Ii Joel M

Classifications

• G06F40/279

  Recognition of textual entities · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N20/20 Primary

  Ensemble learning · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title