System and method integrating machine learning algorithms to enrich documents in a content management system
A system and method for integrating machine learning algorithms to enrich documents in a content management system. The content management system includes a content services engine, a transformation engine to execute one or more transformations on documents stored in the content management system, and a machine learning services to apply a machine learning algorithm to the documents based on the one or more transformations.
1. A content management system comprising:
at least one programmable processor; and
a memory storing instructions that, when executed by the at least one processor, cause the at least one programmable processor to perform operations, the at least one processor configured to:
receive, by a content services engine having a content server that stores one or more documents, a request for a requested document, and serve the requested document to a communication channel;
invoke, responsive to the request and by a transformation engine connected with the content services engine via the communication channel, a transformation to transform the requested document from an original document format to a rendition of the document in an enhanced format;
determine, by a machine learning service having one or more machine learning algorithms, one or more algorithm types of the one or more machine learning algorithms based on the transformation invoked by the transformation engine, and to provide the one or more machine learning algorithms to the transformation;
store, responsive to the invoking, the rendition in the content management system, the rendition including a machine learning evaluation based on the one or more machine learning algorithms and associations that identify the requested document and the enhanced format; and
produce, responsive to a second request for the requested document and using the machine learning evaluation, the rendition of the document in the enhanced format, the producing executed without re-executing the transformation to transform the requested document from the original document format to the rendition.
2. The content management system in accordance with claim 1 , wherein the machine learning service generates an association for the one or more algorithm types between the original document format and the rendition of the document in the enhanced format.
3. The content management system in accordance with claim 2 , wherein the machine learning service invokes a policy based on the association.
4. The content management system in accordance with claim 3 , wherein an event invokes an action to transform the original document format to the rendition of the document in the enhanced format.
5. The content management system in accordance with claim 1 , wherein providing the one or more machine learning algorithms to the transformation includes executing at least one of the one or more machine learning algorithms as part of the transformation.
6. The content management system in accordance with claim 1 , wherein invoking the transformation comprises chaining a plurality of transformations, the plurality of transformations implementing different machine learning algorithms to transform the original document format to an intermediate format, and to the enhanced format.
7. The content management system in accordance with claim 6 , wherein invoking the transformation comprises storing, responsive to the invoking, the rendition in the content server, the rendition including metadata and associations that identify the requested document and the intermediate format.
8. The content management system in accordance with claim 1 , wherein the at least one processor is further configured to:
receive, by a mark-up interface from the one or more machine learning algorithms and responsive to the determining, the rendition, the rendition comprising an optical character recognition having mark-up and geometry information captured by the optical character recognition.
9. The content management system in accordance with claim 1 , wherein the rendition comprises a transcription of a media file that includes speech and encoded in a human-readable, textual file format.
10. The content management system in accordance with claim 1 , wherein the rendition comprises a text-based summarization that is shorter than the original document format encoded as a human-readable, textual file format.
11. The content management system in accordance with claim 1 , wherein the rendition comprises a core concept and an extracted entity encoded as a human readable, textual file format.
12. The content management system in accordance with claim 1 , wherein the at least one processor is further configured to:
detect a change to the original document or metadata; and
re-invoke, responsive to the detecting, the transformation to transform the requested document from the original document format to the rendition of the document in the enhanced format.
13. A method executed by a content management system, the method comprising:
receiving, by a content server of a content services engine, a request from a user for a requested document from one or more documents stored by the content server;
serving, by the content server, the requested document to a transformation engine via a communication channel;
invoking, by the transformation engine, a transformation to transform the requested document from an original document format to a rendition of the document in an enhanced format, the transformation including a specification of one or more machine learning algorithms;
determining, by a machine learning service having the one or more machine learning algorithms, one or more algorithm types of the one or more machine learning algorithms based on the transformation invoked by the transformation engine;
providing the one or more machine learning algorithms to the transformation;
retrieving, by the transformation engine from a machine learning service, the one or more machine learning algorithms specified by the transformation; and
transforming the requested document into the rendition having the enhanced format using the one or more machine learning algorithms;
storing, responsive to the transforming, the rendition in the content server, the rendition including a machine learning evaluation based on the one or more machine learning algorithms and associations that identify the requested document and the enhanced format; and
producing, responsive to a second request for the requested document and using the machine learning evaluation, the rendition of the document in the enhanced format, the producing executed without re-executing the transformation to transform the requested document from the original document format to the rendition.
14. The method in accordance with claim 13 , wherein the machine learning service generates an association for the one or more algorithm types between the original document format and the rendition of the document in the enhanced format.
15. The method in accordance with claim 14 , wherein the machine learning service invokes a policy based on the association.
16. The method in accordance with claim 15 , wherein an event invokes an action to transform the original document format to the rendition of the document in the enhanced format.
17. The method in accordance with claim 13 , wherein providing the one or more machine learning algorithms to the transformation includes executing at least one of the one or more machine learning algorithms as part of the transformation.
18. The method in accordance with claim 13 , wherein invoking the transformation comprises:
invoking a first transformation implementing a first machine learning algorithm to transform the original document format to an intermediate format; and
invoking, responsive to the first transformation, a second transformation implementing a second machine learning algorithm to transform the intermediate format to the enhanced format.
19. The method in accordance with claim 13 , further comprising receiving, by a mark-up interface from the one or more machine learning algorithms, the rendition, the rendition comprising an optical character recognition having mark-up and geometry information captured by the optical character recognition.
20. The method in accordance with claim 13 , further comprising re-invoking, responsive to a change to the original document or to a metadata, the transformation to transform the requested document from the original document format to the rendition of the document in the enhanced format.