Granted Patent B2

US 12,314,669 · App. 18/373,671 · Granted May 27, 2025

Inventors: Spencer Sharpe (Laramie, WY); Annie Ibrahim Rana (Dublin, IE); Valeriy Liberman (Buffalo Grove, IL); Michael Arnold (Duvall, WA); Kyle Michael Caulfield (Mount Prospect, IL); James Cogley (Dublin, IE); Lisa Epstein (Boston, MA); Tricia Sheehan (West Springfield, MA); Rashid Mehdiyev (Buffalo Grove, IL); Saurav Acharya (Des Plaines, IL)

Assignee: UL LLC

G06F40/295G06F40/197G06F40/205G06F40/284

Abstract

Claims (72)

1. A computer-implemented method of creating object models for regulations, the method comprising:

accessing, by a computer processor, an electronic document corresponding to a regulation;

segmenting, by the computer processor, the electronic document into a set of structured texts and a set of metadata, wherein the set of metadata comprises a set of topics descriptive of the regulation;

generating, by the computer processor, an object model for the regulation, the object model comprising the set of structured texts and the set of metadata;

generating, by the computer processor, a summary of the regulation using a machine learning algorithm;

classifying, by the computer processor using a classifier model, a set of sentences contained in the electronic document, wherein each of the set of sentences has a probability of being applicable to at least one of the set of topics that at least meets a specified threshold; and

enriching, by the computer processor, the object model for the regulation with (i) the summary of the regulation and (ii) the set of sentences that was classified.

2. The computer-implemented method of claim 1 , further comprising:

analyzing, by the computer processor, a portion of the set of structured texts within the object model using an entity recognition analysis to determine a set of entity names meaningful to the regulation; and

enriching, by the computer processor, the object model with the set of entity names.

3. The computer-implemented method of claim 1 , further comprising:

outputting at least a portion of the set of structured texts within the object model corresponding to a set of keywords or phrases; and

enriching, by the computer processor, the object model with the set of keywords or phrases.

4. The computer-implemented method of claim 1 , further comprising:

performing, by the computer processor, a linguistic analysis on the object model to detect the set of sentences within the set of structured texts.

5. The computer-implemented method of claim 4 , wherein performing the linguistic analysis on the object model comprises one of:

detecting the set of sentences and generating a set of token n-grams for each sentence in the set of sentences; or

detecting the set of sentences from a set of n-grams tokenized from the set of structured texts.

6. The computer-implemented method of claim 1 , wherein segmenting the electronic document comprises:

segmenting the electronic document into one or more of header, footer, title, body, sections, sub-sections, paragraphs, lists, sub-lists, citations, or references.

7. The computer-implemented method of claim 1 , further comprising:

examining, by the computer processor, the electronic document to determine that the set of topics is not present;

generating, by the computer processor, the set of topics for the electronic document; and

storing the set of topics with the set of metadata.

8. A system for creating object models for regulations, comprising:

a memory storing instructions; and

at least one processor interfaced with the memory and configured to execute the instructions to cause the at least one processor to:

access an electronic document corresponding to a regulation,

segment the electronic document into a set of structured texts and a set of metadata, wherein the set of metadata comprises a set of topics descriptive of the regulation,

generate an object model for the regulation, the object model comprising the set of structured texts and the set of metadata,

generate a summary of the regulation using a machine learning algorithm,

classify, using a classifier model, a set of sentences contained in the electronic document, wherein each of the set of sentences has a probability of being applicable to at least one of the set of topics that at least meets a specified threshold, and

enrich the object model for the regulation with (i) the summary of the regulation and (ii) the set of sentences that was classified.

9. The system of claim 8 , wherein the at least one processor is further configured to:

analyze a portion of the set of structured texts within the object model using an entity recognition analysis to determine a set of entity names meaningful to the regulation, and

enrich the object model with the set of entity names.

10. The system of claim 8 , wherein the at least one processor is further configured to:

output at least a portion of the set of structured texts within the object model corresponding to a set of keywords or phrases, and

enrich the object model with the set of keywords or phrases.

11. The system of claim 8 , wherein the at least one processor is further configured to:

perform a linguistic analysis on the object model to detect the set of sentences within the set of structured texts.

12. The system of claim 11 , wherein to perform the linguistic analysis on the object model, the at least one processor is configured to:

detect the set of sentences and generating a set of token n-grams for each sentence in the set of sentences, or

detect the set of sentences from a set of n-grams tokenized from the set of structured texts.

13. The system of claim 8 , wherein to segment the electronic document, the at least one processor is configured to:

segment the electronic document into one or more of header, footer, title, body, sections, sub-sections, paragraphs, lists, sub-lists, citations, or references.

14. The system of claim 8 , wherein the at least one processor is further configured to:

examine the electronic document to determine that the set of topics is not present,

generate the set of topics for the electronic document, and

store the set of topics with the set of metadata.

15. A non-transitory computer-readable storage medium configured to store instructions executable by one or more processors, the instructions comprising:

instructions for accessing an electronic document corresponding to a regulation;

instructions for segmenting the electronic document into a set of structured texts and a set of metadata, wherein the set of metadata comprises a set of topics descriptive of the regulation;

instructions for generating an object model for the regulation, the object model comprising the set of structured texts and the set of metadata;

instructions for generating a summary of the regulation using a machine learning algorithm;

instructions for classifying, using a classifier model, a set of sentences contained in the electronic document, wherein each of the set of sentences has a probability of being applicable to at least one of the set of topics that at least meets a specified threshold; and

instructions for enriching the object model for the regulation with (i) the summary of the regulation and (ii) the set of sentences that was classified.

16. The non-transitory computer-readable storage medium of claim 15 , wherein the instructions further comprise:

instructions for analyzing a portion of the set of structured texts within the object model using an entity recognition analysis to determine a set of entity names meaningful to the regulation; and

instructions for enriching the object model with the set of entity names.

17. The non-transitory computer-readable storage medium of claim 15 , wherein the instructions further comprise:

instructions for outputting at least a portion of the set of structured texts within the object model corresponding to a set of keywords or phrases; and

instructions for enriching the object model with the set of keywords or phrases.

18. The non-transitory computer-readable storage medium of claim 15 , wherein the instructions further comprise:

instructions for performing a linguistic analysis on the object model to detect the set of sentences within the set of structured texts.

19. The non-transitory computer-readable storage medium of claim 18 , wherein the instructions for performing the linguistic analysis on the object model comprise one of:

instructions for detecting the set of sentences and generating a set of token n-grams for each sentence in the set of sentences; or

instructions for detecting the set of sentences from a set of n-grams tokenized from the set of structured texts.

20. The non-transitory computer-readable storage medium of claim 15 , wherein the instructions further comprise:

instructions for examining the electronic document to determine that the set of topics is not present;

instructions for generating the set of topics for the electronic document; and

instructions for storing the set of topics with the set of metadata.

Assignments (1)

Continuity (3)

References Cited (17)