Artificial intelligence (AI) trained data model selection

Patent metadata

Field	Value
Publication number	US-12444190-B2
Application number	US-202017107866-A
Country	US
Kind code	B2
Filing date	Nov 30, 2020
Priority date	Nov 30, 2020
Publication date	Oct 14, 2025
Grant date	Oct 14, 2025

Abstract

This disclosure describes techniques for continuous improvement of machine learning models (also called data models) in a Content Management System (CMS). In one example, a CMS may store a set of data models for each application such as plate number recognition, facial recognition, a determination of likelihood of assault to a law enforcement officer in a traffic violation or robbery scenario, and car identification. In an example embodiment, a predictive model may be used to select a data model from the plurality of data models. The selected data model may be further improved or trained to a new sample of data features to generate an output pattern (e.g., likelihood of assault to a law enforcement officer).

First claim

What is claimed is:

1. One or more non-transitory computer-readable storage media storing computer-executable instructions that upon execution cause one or more computers to perform acts comprising:
operating, by a content management system (CMS), a prediction model stored in the CMS to select a data model from a plurality of stored data models in the CMS for a law enforcement application;
retrieving a first set of data features used to train the selected data model and a second set of data features used to train at least one historical version of the selected data model;
incorporating the first set of data features and the second set of data features to generate an incorporated data set;
generating a new data model based upon the incorporated data set, the new data model being configured to generate an output pattern;
comparing an expected accuracy of the new data model with an associated expected accuracy of the selected data model; and
operating the new data model for the law enforcement application by the CMS when the expected accuracy of the new data model is greater than the associated expected accuracy of the selected data model by at least a threshold value, wherein the operating the new data model comprises:
processing, by the new data model, real-time data received from the law enforcement application; and
sending the output pattern generated by the new data model as a real-time notification.

2. The one or more non-transitory computer-readable storage media of claim 1 , wherein the selected data model includes an output pattern that is based from an output pattern of the law enforcement application.

3. The one or more non-transitory computer-readable storage media of claim 1 , wherein a version of the retrieved first set of data features for the selected data model is different from a version of the retrieved second set of data features.

4. The one or more non-transitory computer-readable storage media of claim 1 , wherein the new data model is stored as another version of the selected data model.

5. The one or more non-transitory computer-readable storage media of claim 1 further comprising: tracking, by the CMS, of defective data set sources; and marking data models that are associated with the defective data set sources.

6. The one or more non-transitory computer-readable storage media of claim 5 , wherein the defective data set sources include telemetry data that are collected from defective devices.

7. The one or more non-transitory computer-readable storage media of claim 1 , wherein the CMS is configured to access different prediction models that are used for different applications.

8. The one or more non-transitory computer-readable storage media of claim 1 , wherein the historical version of the selected data model is associated with data features that contributed to an improvement of another data model in the plurality of stored data models.

9. A computer system, comprising:
at least one processor;
a memory coupled to the at least one processor, the memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
operating a prediction model stored in a content management system (CMS) to select a data model from a plurality of stored data models in the CMS for a law enforcement application;
retrieving a first set of data features used to train the selected data model and a second set of data features used to train at least one historical version of the selected data model;
incorporating the first set of data features and the second set of data features to generate an incorporated data set;
generating a new data model based upon the incorporated data set, the new data model being configured to generate an output pattern;
comparing an expected accuracy of the new data model with an associated expected accuracy of the selected data model; and
operating the new data model for the law enforcement application by the CMS when the expected accuracy of the new data model is greater than the associated expected accuracy of the selected data model by at least a threshold value, wherein the operating the new data model comprises:
processing, by the new data model, real-time data received from the law enforcement application; and
sending the output pattern generated by the new data model as a real-time notification.

10. The computer system of claim 9 , wherein the selected data model includes an output pattern that is based from an output pattern of the law enforcement application.

11. The computer system of claim 9 , wherein a version of the retrieved first set of data features for the selected data model is different from a version of the retrieved second set of data features.

12. The computer system of claim 9 , wherein the new data model is stored as another version of the selected data model.

13. The computer system of claim 9 , wherein the computer system is configured to track defective data set sources and mark data models that are associated with the defective data set sources.

14. The computer system of claim 13 , wherein the defective data set sources include telemetry data that are gathered from defective devices.

15. The computer system of claim 9 , wherein the computer system is configured to access different prediction models that are used for different applications.

16. A computer-implemented method, comprising:
operating, by a content management system (CMS), a prediction model stored in the CMS to select a data model from a plurality of stored data models in the CMS for a law enforcement application;
retrieving a first set of data features used to train the selected data model and a second set of data features used to train an original version of the selected data model;
incorporating the first set of data features and the second set of data features to generate an incorporated data set;
generating a new data model based upon the incorporated data set, the new data model being configured to generate an output pattern;
comparing an expected accuracy of the new data model with an associated expected accuracy of the selected data model; and
operating the new data model for the law enforcement application by the CMS when the expected accuracy of the new data model is greater than the associated expected accuracy of the selected data model by at least a threshold value, wherein the operating the new data model comprises:
processing, by the new data model, real-time data received from the law enforcement application; and
sending the output pattern generated by the new data model as a real-time notification.

17. The computer-implemented method of claim 16 , wherein the CMS is configured to access different prediction models that are used for different applications.

Assignees

• Getac Technology Corp
• Whp Workflow Solutions Inc

Inventors

• Guzik Thomas
• Adeel Muhammad

Classifications

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06V20/52

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/435 Primary

  Filtering based on additional data, e.g. user or group profiles · CPC title

• G06V20/41 Primary

  Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items (segmenting video sequences G06V20/49) · CPC title