Machine learning based extraction of partition objects from electronic documents

1. A method, comprising:

automatically generating a modified electronic document, via a processor of a compute device and in response to one of: (1) detecting an error in a first electronic document having an associated domain, or (2) detecting a first user interaction, via a user interface of the compute device, with a representation of the first electronic document;

identifying, via a domain-specific machine learning model for the associated domain, and in response to generating the modified electronic document, a set of objects associated with the modified electronic document;

detecting a second user interaction, via the user interface of the compute device, with a representation of an object from the set of objects;

in response to detecting the second user interaction, modifying the domain-specific machine learning model based on the second user interaction to produce a modified machine learning model;

detecting, based on a signal representing a user-selected data element for a second electronic document having the associated domain, and using the modified machine learning model, a set of objects associated with the second electronic document; and

displaying, via the user interface, a representation of each object from the set of objects associated with the second electronic document.

2. The method of claim 1 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the error is an optical character recognition (OCR) induced error.

3. The method of claim 1 , wherein the domain-specific machine learning model includes at least one of a sequence-to-sequence model or an attention algorithm.

4. The method of claim 1 , wherein the domain-specific machine learning model operates within an artificial neural network (ANN).

5. The method of claim 1 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the detecting the error in the first electronic document includes receiving an error signal representing a non-extracted data string, associated with the first electronic document, from a named-entity recognition (NER) model.

6. The method of claim 1 , wherein the associated domain is a document ype of the first document.

7. The method of claim 1 , wherein the second user interaction includes one of clicking on or highlighting the representation of the object from the set of objects.

8. An apparatus, comprising:

a processor; and

a memory storing instructions, executable by the processor, to:

automatically generate a modified electronic document, in response to one of: (1) detecting an error in a first electronic document having an associated domain, or (2) detecting a first user interaction, via a user interface of a compute device, with a representation of the first electronic document;

identify, via a domain-specific machine learning model for the associated domain, and in response to generating the modified electronic document, a set of objects associated with the modified electronic document;

detect a second user interaction, via the user interface of the compute device, with a representation of an object from the set of objects;

in response to detecting the second user interaction, modify the domain-specific machine learning model based on the second user interaction to produce a modified machine learning model;

detect, based on a signal representing a user-selected data element a second electronic document having the associated domain, and using the modified machine learning model, a set of objects associated with the second electronic document; and

display, via the user interface, a representation of each object from the set of objects associated with the second electronic document.

9. The apparatus of claim 8 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the error is an optical character recognition (OCR) induced error.

10. The apparatus of claim 8 , wherein the domain-specific machine learning model includes at least one of a sequence-to-sequence model or an attention algorithm.

11. The apparatus of claim 8 , wherein the domain-specific machine learning model operates within an artificial neural network (ANN).

12. The apparatus of claim 8 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the detecting the error in the first electronic document includes receiving an error signal representing a non-extracted data string; associated with the first electronic document, from a named-entity recognition (NER) model.

13. The apparatus of claim 8 , wherein the associated domain is a document type of the first document.

14. The apparatus of claim 8 , wherein the second user interaction includes one of clicking on or highlighting the representation of the object from the set of objects.

15. A non-transitory, processor-readable medium storing instructions to cause a processor to:

automatically generate a modified electronic document, in response to one of: (1) detecting an error in a first electronic document having an associated domain, or (2) detecting a first user interaction, via a user interface of a compute device, with a representation of the first electronic document;

identify, via a domain-specific machine learning model for the associated domain, and in response to generating the modified electronic document, a set of objects associated with the modified electronic document;

detect a second user interaction, via the user interface of the compute device, with a representation of an object from the set of objects;

in response to detecting the second user interaction, modify the domain-specific machine learning model based on the second user interaction to produce a modified machine learning model;

detect, based on a signal representing a user-selected data element for a second electronic document having the associated domain, and using the modified machine learning model, a set of objects associated with the second electronic document; and

display, via the user interface, a representation of each object from the set of objects associated with the second electronic document.

16. The non-transitory, processor-readable medium of claim 15 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the error is an optical character recognition (OCR) induced error.

17. The non-transitory, processor-readable medium of claim 15 , wherein the domain-specific machine learning model includes at least one of a sequence-to-sequence model or an attention algorithm.

18. The non-transitory, processor-readable medium of claim 15 , wherein the domain-specific machine learning model operates within an artificial neural network (ANN).

19. The non-transitory, processor-readable medium of claim 15 , wherein the generating the modified electronic document is in response to detecting the error in the first electronic document, and the detecting the error in the first electronic document includes receiving an error signal representing a non-extracted data string, associated with the first electronic document, from a named-entity recognition (NER) model.

20. The non-transitory, processor-readable medium of claim 15 , wherein the associated domain is a document type of the first document.

21. The method of claim 1 , wherein the automatically generating the modified electronic document is in response to detecting the error in the first electronic document, and the error is one of an automated language translation induced error or an automated spell-checking induced error.