IP Library › Granted Patent US 12,645,866
Granted Patent B2
US 12,645,866 · App. 18/534,297 · Granted Jun 2, 2026

System and method integrating machine learning algorithms to enrich documents in a content management system

Inventors: John Newton (Warfield, GB); Brian Remmington (Wokingham, GB); Jan Vonka (Reading, GB); Tom Morris (London, GB); Chris Hudson-Silver (London, GB); Cezar Leahu (Iasi, RO)
Assignee: Hyland UK Operations Limited
G06F40/16G06N20/00
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,645,866
App. No.
18/534,297
Granted
Jun 2, 2026
Kind
B2
Abstract

A system and method for integrating machine learning algorithms to enrich content items in a content management system. The content management system includes at least one processor, a transformation engine to execute one or more transformations on content items stored in the content management system, and a machine learning services to apply a machine learning algorithm to the content items based on the one or more transformations.

Claims (42)

1 . A content management system comprising:

at least one processor in communication with a content server having one or more storage media devices for storing one or more content items;

a first machine learning transformation model; and a second machine learning transformation model;

a first content item being transformed from a first format to a second format, wherein the transforming is performed by application of the first machine learning transformation model and the second machine learning transformation model;

wherein the first machine learning transformation model generates an association between the first format and the second format;

the first machine learning transformation model invokes a transformation policy based on the association; and

the transformation policy is invoked to update the first content item or changes in the association.

2 . The system of claim 1 , wherein transformation from the first format to the second format results in updating the first content item from a first rendition to a second rendition.

3 . The system of claim 2 , wherein a mark-up interface is provided to receive the second rendition comprising optical character recognition data having mark-up and geometry information.

4 . The system of claim 2 wherein the second rendition comprises a transcription of a media file that includes speech or encoded human-readable textual file format or a text-based summarization that is shorter than the first content item.

5 . The system of claim 1 , wherein the first and the second machine learning transformation models are sequentially applied to transform the first content item from a first rendition to a second rendition.

6 . The system of claim 1 , wherein the first machine learning transformation model transforms the first content item from a first format to an intermediate format and the second machine learning transformation model transforms the first content item from the intermediate format to a second format.

7 . The system of claim 1 , wherein the content management system is configured to: create an association between the first content item and a stored rendition;

type the association based upon the type of transformation being performed; and

invoke a policy based upon the type of the association.

8 . The system of claim 1 , wherein the association is created for a transformation type between the first content item and a rendition created by the first machine learning transformation model.

9 . The system of claim 1 , wherein the transformation policy is invoked to update the first content item or changes in the association without fully transforming the first content item again; or

the transformation policy is invoked to update changes in the association.

10 . A method of managing content in a content management system comprising:

at least one processor in communication with a content server having one or more storage media devices for storing one or more content items;

a first machine learning transformation model; and a second machine learning transformation model;

the method comprising transforming a first content item from a first format to a second format, wherein the transforming is performed by application of the first machine learning transformation model and the second machine learning transformation model;

wherein the first machine learning transformation model generates an association between the first format and the second format;

the first machine learning transformation model invokes a transformation policy based on the association; and

the transformation policy is invoked to update the first content item or changes in the association.

11 . The method of claim 10 , wherein transformation from the first format to the second format results in updating the first content item from a first rendition to a second rendition.

12 . The method of claim 11 , wherein in response to a request for access to the first content item, the first content item is rendered in the second format without invoking a transformation event.

13 . The method of claim 11 , wherein the second rendition comprises a transcription of a media file that includes speech, or encoded human-readable textual file format.

14 . The method of claim 10 , wherein responsive to detecting a change in data associated with the first content item, the first and the second machine learning transformation models are sequentially applied to transform the first content item from a first rendition to a second rendition.

15 . The method of claim 10 , wherein the first machine learning transformation model transforms the first content item from a first format to an intermediate format and the second machine learning transformation model transforms the first content item from the intermediate format to a second format.

16 . The method of claim 10 , wherein the first machine learning transformation model and the second machine learning transformation model exclude performing optical text recognition.

17 . A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:

servicing a first content item having a first rendition, in response to receiving a first transformation triggering event;

transforming the first content item by at least partially using the first rendition from a first format to a second format, in response to detecting a change to the first content item or a change to metadata associated with the first content item; and

applying two or more machine learning models to the transformation such that a second rendition of the first content item is generated and stored in one or more storage media devices;

wherein a first machine learning model generates an association between the first format and the second format;

the first machine learning transformation model invokes a transformation policy based on the association; and

the transformation policy is invoked to update the first content item or changes in the association.

18 . The computer program product of claim 17 , wherein the second rendition of the first content item is stored without re-invoking the first transformation triggering event.

19 . The computer program product of claim 17 , further comprising:

invoking an action to perform based upon the first rendition in order to perform an additional task.

20 . The computer program product of claim 19 , wherein the additional task is adding metadata or moving the first content item to a new location based upon information contained therein.

Continuity (3)
Continuation 17745735 · May 16, 2022
Continuation 16372051 · Apr 1, 2019
Related Publication 20240104293A1 · Mar 28, 2024
References Cited (12)
US 11544460B1 · Margolin · 2023 [cited by examiner]
US 20080222516A1 · Petri · 2008 [cited by examiner]
US 20090116755A1 · Neogi · 2009 [cited by examiner]
US 20110271176A1 · Menday · 2011 [cited by examiner]
US 20180196873A1 · Yerebakan · 2018 [cited by examiner]
US 20190050639A1 · Ast · 2019 [cited by examiner]
US 20190102375A1 · Goulikar · 2019 [cited by examiner]
US 20190108277A1 · Taneja · 2019 [cited by examiner]
US 20190294661A1 · Sarkar · 2019 [cited by examiner]
US 20190311227A1 · Kriegman · 2019 [cited by examiner]
US 20200234003A1 · Bakman · 2020 [cited by examiner]
US 20200311187A1 · Newton · 2020 [cited by examiner]