Method and apparatus for automatically extracting information from unstructured data

1. A method for implementing a trust and estate smart application module for automatically extracting information from unstructured data by utilizing one or more processors and one or more memories, the method comprising:

receiving digitized data of a document having unstructured data format;

applying machine learning models for sectioning the digitized data;

applying optical character recognition (OCR) processing to the sectioned digitized data by utilizing an OCR device;

matching the sectioned digitized data to patterns and rules accessed from a knowledge-based database;

applying classification models to the matched digitized data to identify entities and events from the sectioned digitized data;

automatically linking each entity with corresponding event in a hierarchical format to generate a document having structured data format; and

outputting the document having the structured data with metadata having the linked entity with corresponding event in the hierarchical format to downstream applications;

receiving a plurality of trust and estate documents, and for each trust and estate document:

simultaneously processing the plurality of trust and estate documents;

interpreting contents of each document having the unstructured data format in a way that would translate it into the document having the structured data format regarding provisions, parties, and events detailed in the unstructured data;

automatically generating the structured data that indicates a linkage between each entity and its associated event for each trust and estate document; and

automatically triggering execution of a plurality of platforms of downstream applications that support trust and estate administration.

2. The method according to claim 1 , wherein prior to matching, the method further comprising:

checking metadata of the received digitized data of the document by determining whether the digitized data has been processed previously; and

based on a positive determination, extracting previously processed rules and heuristics associated with the metadata.

3. The method according to claim 1 , wherein the knowledge-based database is configured to store information comprising:

rules and heuristics relating to trust and estate;

vocabularies for language models in the field of trust and estate;

entities, facts, and events that drive provisions of a trust and estate document;

look-up tables having vocabularies related to trust and estate; and

information relating to trust and estate training data collection.

4. The method according to claim 1 , further comprising:

applying the machine learning models to extract taxonomy codes as applied to the document, wherein the taxonomy codes includes roles, interested parties, powers or authority associated with a party or a role, and provisions relating to trust and estate.

5. The method according to claim 1 , wherein the classification models comprising:

an auxiliary model that identifies whether a sentence extracted from the digitized data relates to an entity or an event governing a trust and estate document;

a sentence detection model that identifies whether a sentence extracted from the digitized data is a complete sentence; and

a fragment detection model that identifies whether a sentence extracted from the digitized data is an incomplete sentence or a heading.

6. The method according to claim 1 , further comprising:

persisting all entities from the metadata;

for each entity, persisting all events relating to the corresponding entity;

for each event, persisting all parties relating to the corresponding event;

for each party, persisting all powers associated with the corresponding party;

persisting all provisions associated to each entity and all the entity attributes;

transforming all persisted data into queryable format; and

storing the transformed data into a database that is queryable by the downstream applications.

7. A system for implementing a trust and estate smart application module for automatically extracting information from unstructured data, comprising:

a receiver configured to receive digitized data of a document having unstructured data format;

a processor; and

a knowledge-based database operatively connected to the processor and the receiver via a communication network,

wherein the processor is configured to:

apply machine learning models to section the digitized data;

apply optical character recognition (OCR) processing to the sectioned digitized data by utilizing an OCR device;

match the sectioned digitized data to patterns and rules accessed from the knowledge-based database;

apply classification models to the matched digitized data to identify entities and events from the sectioned digitized data;

automatically link each entity with corresponding event in a hierarchical format to generate a document having structured data format;

output the document having the structured data with metadata having the linked entity with corresponding event in the hierarchical format to downstream applications;

receive a plurality of trust and estate documents, and for each trust and estate document:

simultaneously process the plurality of trust and estate documents;

interpret contents of each document having the unstructured data format in a way that would translate it into the document having the structured data format regarding provisions, parties, and events detailed in the unstructured data;

automatically generate the structured data that indicates a linkage between each entity and its associated event for each trust and estate document; and

automatically trigger execution of a plurality of platforms of downstream applications that support trust and estate administration.

8. The system according to claim 7 , wherein prior to match, the processor is further configured to:

check metadata of the received digitized data of the document by determining whether the digitized data has been processed previously; and

based on a positive determination, extract previously processed rules and heuristics associated with the metadata.

9. The system according to claim 7 , wherein the knowledge-based database is configured to store information comprising:

rules and heuristics relating to trust and estate;

vocabularies for language models in the field of trust and estate;

entities, facts, and events that drive provisions of a trust and estate document;

look-up tables having vocabularies related to trust and estate; and

information relating to trust and estate training data collection.

10. The system according to claim 7 , wherein the processor is further configured to apply the machine learning models to extract taxonomy codes as applied to the document, and wherein the taxonomy codes includes roles, interested parties, powers or authority associated with a party or a role, and provisions relating to trust and estate.

11. The system according to claim 7 , wherein the classification models comprising:

an auxiliary model that identifies whether a sentence extracted from the digitized data relates to an entity or an event governing a trust and estate document;

a sentence detection model that identifies whether a sentence extracted from the digitized data is a complete sentence; and

a fragment detection model that identifies whether a sentence extracted from the digitized data is an incomplete sentence or a heading.

12. The system according to claim 7 , wherein the processor is further configured to:

persist all entities from the metadata;

for each entity, persist all events relating to the corresponding entity;

for each event, persist all parties relating to the corresponding event;

for each party, persist all powers associated with the corresponding party;

persist all provisions associated to each entity and all the entity attributes;

transform all persisted data into queryable format; and

store the transformed data into a database that is queryable by the downstream applications.

13. A non-transitory computer readable medium configured to store instructions for automatically extracting information from unstructured data, wherein when executed, the instructions cause a processor to perform the following:

receiving digitized data of a document having unstructured data format;

applying machine learning models for sectioning the digitized data;

applying optical character recognition (OCR) processing to the sectioned digitized data by utilizing an OCR device;

matching the sectioned digitized data to patterns and rules accessed from a knowledge-based database;

applying classification models to the matched digitized data to identify entities and events from the sectioned digitized data;

automatically linking each entity with corresponding event in a hierarchical format to generate a document having structured data format;

outputting the document having the structured data with metadata having the linked entity with corresponding event in the hierarchical format to downstream applications;

receiving a plurality of trust and estate documents, and for each trust and estate document:

simultaneously processing the plurality of trust and estate documents;

interpreting contents of each document having the unstructured data format in a way that would translate it into the document having the structured data format regarding provisions, parties, and events detailed in the unstructured data;

automatically generating the structured data that indicates a linkage between each entity and its associated event for each trust and estate document; and

automatically triggering execution of a plurality of platforms of downstream applications that support trust and estate administration.

14. The non-transitory computer readable medium according to claim 13 , wherein when executed, the instructions further cause the processor to perform the following prior to matching:

checking metadata of the received digitized data of the document by determining whether the digitized data has been processed previously; and

based on a positive determination, extracting previously processed rules and heuristics associated with the metadata.

15. The non-transitory computer readable medium according to claim 13 , wherein when executed, the instructions further cause the processor to perform the following:

applying the machine learning models to extract taxonomy codes as applied to the document, wherein the taxonomy codes includes roles, interested parties, powers or authority associated with a party or a role, and provisions relating to trust and estate.

16. The non-transitory computer readable medium according to claim 13 , wherein the classification models comprising:

an auxiliary model that identifies whether a sentence extracted from the digitized data relates to an entity or an event governing a trust and estate document;

a sentence detection model that identifies whether a sentence extracted from the digitized data is a complete sentence; and

a fragment detection model that identifies whether a sentence extracted from the digitized data is an incomplete sentence or a heading.

17. The non-transitory computer readable medium according to claim 13 , wherein when executed, the instructions further cause the processor to perform the following:

persisting all entities from the metadata;

for each entity, persisting all events relating to the corresponding entity;

for each event, persisting all parties relating to the corresponding event;

for each party, persisting all powers associated with the corresponding party;

persisting all provisions associated to each entity and all the entity attributes;

transforming all persisted data into queryable format; and

storing the transformed data into a database that is queryable by the downstream applications.