Automated document tagging in a digital management platform
An auto-tagging engine receives a training set of data comprising documents including a set of tagged fields with each tagged field corresponding to a portion of the document. The auto-tagging engine trains a machine learned model using the training set of data. The trained machine learned model, when applied to a target document in a document management environment, identifies portions of the target document each corresponding to fields of the target document. For each field of the target document, the auto-tagging engine identifies text of the target document associated with the identified portions of the target document corresponding to fields. Natural language processing is performed on the identified text in order to identify field types for the fields. The target document is automatically modified to include a tag identifying the portion of the target document corresponding to each field and identifying a field type of the field.
1. A method for automatically tagging fields of a target document, comprising:
applying a machine learned model to the target document to identify portions of the target document corresponding to the fields of the target document, wherein the machine learned model has been trained using a training set of data comprising documents each with a set of tagged fields within a document, each tagged field corresponding to a portion of the document, wherein the tagged fields include at least one of: a tagged field for entering data at a predetermined location in the document, a blank tagged field at another predetermined location in the document, and any combination thereof, wherein the machine learned model configured to identify, for each of one or more fields within the document, a portion within the document corresponding to the field; and
for each field of the target document:
identifying text of the target document associated with the identified portion of the target document corresponding to the field;
performing natural language processing on the identified text to identify a field type of the field of a plurality of types of fields; and
automatically modifying viewable content of the target document to include a tag, wherein the tag identifies a location in the target document corresponding to the field and the identified field type of the field, wherein the tag visually indicates to a user to further modify the document within the field, a visual indicator associated with the tag varies in accordance with the identified field type of the field, the plurality of types of fields including one or more fields that accept textual input and one or more fields that do not accept textual input; and
modifying, using the tag, the document by at least one of the following: entering data within the field corresponding to the tag, modifying the tag, and any combination thereof.
2. The method of claim 1 , wherein the machine learned model is configured to:
identify a bounding box for each field of the target document, the bounding box associated with location coordinates identifying a location of the bounding box within the target document;
wherein identifying text of the target document associated with the identified portion of the target document corresponding to the field comprises:
identifying characters within the target document and a character location for each character in the target document; and
determining, based on the character locations, the characters located within a threshold distance of the location coordinates of the bounding box.
3. The method of claim 2 , wherein the location coordinates include a starting coordinate and an ending coordinate for each bounding box, and the method further comprising:
detecting over-lapping location coordinates for the bounding boxes of the target document, wherein the over-lapping location coordinates for one bounding box comprises either a starting coordinate or an ending coordinate that is located at or between the starting coordinate and the ending coordinate of another bounding box; and
removing all but one of the bounding boxes with over-lapping location coordinates based on a set of predetermined rules.
4. The method of claim 1 , wherein one or more of the identified field types is a text field, and for each text field of the target document:
automatically modifying the target document to include an identification of text to be entered by a user in the text field based on a field sub-type of the text field.
5. The method of claim 1 , wherein the target document is accessed by:
receiving the target document to be auto-tagged from a user of a document management system;
generating the target document to be auto-tagged within the document management system;
accessing the target document to be auto-tagged, the target document stored within the document management system; or
accessing the target document to be auto-tagged from a document system different than the document management system.
6. The method of claim 1 , further comprising:
providing, to a device of a user, the modified target document; and
receiving, from the device of the user, feedback including at least one of: approval of the modified target document, deletion of one or more tags included in the modified target document, an update to one or more field types, and an update to one or more tags.
7. A non-transitory computer readable storage medium comprising computer executable instructions for automatically tagging fields of a target document that when executed by one or more processors causes the one or more processors to:
apply a machine learned model to the target document to identify portions of the target document corresponding to the fields of the target document, wherein the machine learned model has been trained using a training set of data comprising documents each with a set of tagged fields within a document, each tagged field corresponding to a portion of the document, wherein the tagged fields include at least one of: a tagged field for entering data at a predetermined location in the document, a blank tagged field at another predetermined location in the document, and any combination thereof, wherein the machine learned model configured to identify, for each of one or more fields within the document, a portion within the document corresponding to the field; and
for each field of the target document:
identify text of the target document associated with the identified portion of the target document corresponding to the field;
perform natural language processing on the identified text to identify a field type of the field of a plurality of types of fields; and
automatically modify viewable content of the target document to include a tag, wherein the tag identifies a location in the target document corresponding to the field and the identified field type of the field, wherein the tag visually indicates to a user to further modify the document within the field, a visual indicator associated with the tag varies in accordance with the identified field type of the field, the plurality of types of fields including one or more fields that accept textual input and one or more fields that do not accept textual input;
modify, using the tag, the document by at least one of the following: entering data within the field corresponding to the tag, modifying the tag, and any combination thereof.
8. The non-transitory computer readable storage medium of claim 7 , wherein the machine learned model is configured to:
identify a bounding box for each field of the target document, the bounding box associated with location coordinates identifying a location of the bounding box within the target document;
wherein identifying text of the target document associated with the identified portion of the target document corresponding to the field comprises:
identifying characters within the target document and a character location for each character in the target document; and
determining, based on the character locations, the characters located within a threshold distance of the location coordinates of the bounding box.
9. The non-transitory computer readable storage medium of claim 8 , wherein the location coordinates include a starting coordinate and an ending coordinate for each bounding box, and wherein the one or more processors are configured to:
detect over-lapping location coordinates for the bounding boxes of the target document, wherein the over-lapping location coordinates for one bounding box comprises either a starting coordinate or an ending coordinate that is located at or between the starting coordinate and the ending coordinate of another bounding box; and
remove all but one of the bounding boxes with over-lapping location coordinates based on a set of predetermined rules.
10. The non-transitory computer readable storage medium of claim 7 , wherein each identified field type includes at least one of: a textfield, a radio button field, a signature field, an initial field, a date field, and a date signed field.
11. The non-transitory computer readable storage medium of claim 7 , wherein one or more of the identified field types is a text field, and wherein the one or more processors are configured to:
automatically modify the target document to include an identification of text to be entered by a user in the text field based on a field sub-type of the text field.
12. The non-transitory computer readable storage medium of claim 11 , wherein the field sub-type of the one or more text fields includes one of: a name, an address, a title, a date, a company name, and a dollar amount.
13. The non-transitory computer readable storage medium of claim 7 , the target document is accessed by:
receiving the target document to be auto-tagged from a user of a document management system;
generating the target document to be auto-tagged within the document management system;
accessing the target document to be auto-tagged, the target document stored within the document management system; or
accessing the target document to be auto-tagged from a document system different than the document management system.
14. The non-transitory computer readable storage medium of claim 7 , wherein the one or more processors are configured to:
provide, to a device of a user, the modified target document; and
receive, from the device of the user, feedback including at least one of: approval of the modified target document, deletion of one or more tags included in the modified target document, an update to one or more field types, and an update to one or more tags.
15. A computer system for automatically tagging fields of a target document comprising:
one or more processors; and
a non-transitory computer readable storage medium comprising computer executable instructions that when executed by one or more processors causes the one or more processors to:
apply a machine learned model to the target document to identify portions of the target document corresponding to the fields of the target document, wherein the machine learned model has been trained using a training set of data comprising documents each with a set of tagged fields within a document, each tagged field corresponding to a portion of the document, wherein the tagged fields include at least one of: a tagged field for entering data at a predetermined location in the document, a blank tagged field at another predetermined location in the document, and any combination thereof, wherein the machine learned model configured to identify, for each of one or more fields within the document, a portion within the document corresponding to the field; and
for each field of the target document:
identify text of the target document associated with the identified portion of the target document corresponding to the field;
perform natural language processing on the identified text to identify a field type of the field of a plurality of types of fields; and
automatically modify viewable content of the target document to include a tag, wherein the tag identifies a location in the target document corresponding to the field and the identified field type of the field, wherein the tag visually indicates to a user to further modify the document within the field, a visual indicator associated with the tag varies in accordance with the identified field type of the field, the plurality of types of fields including one or more fields that accept textual input and one or more fields that do not accept textual input; and
modify, using the tag, the document by at least one of the following: entering data within the field corresponding to the tag, modifying the tag, and any combination thereof.
16. The computer system of claim 15 , wherein the machine learned model is configured to:
identify a bounding box for each field of the target document, the bounding box associated with location coordinates identifying a location of the bounding box within the target document;
wherein identifying text of the target document associated with the identified portion of the target document corresponding to the field comprises:
identifying characters within the target document and a character location for each character in the target document; and
determining, based on the character locations, the characters located within a threshold distance of the location coordinates of the bounding box.
17. The computer system of claim 16 , wherein the location coordinates include a starting coordinate and an ending coordinate for each bounding box, and wherein the one or more processors are configured to:
detect over-lapping location coordinates for the bounding boxes of the target document, wherein the over-lapping location coordinates for one bounding box comprises either a starting coordinate or an ending coordinate that is located at or between the starting coordinate and the ending coordinate of another bounding box; and
remove all but one of the bounding boxes with over-lapping location coordinates based on a set of predetermined rules.
18. The method of claim 1 , wherein performing natural language processing on the identified text to identify a field type of each field comprises applying text associated with or corresponding to the field to a further machine learned model that is trained to output the field type.
19. The method of claim 1 , wherein the identified field type of each tag is displayed responsive to detecting a user interaction in relation to the tag.