IP Library Granted Patent US 12,105,728
Granted Patent B2
US 12,105,728 · App. 17/475,204 · Granted Oct 1, 2024

Extensible data objects for use in machine learning models

Inventors: Stephen W. Shillingford (Salt Lake City, UT); Wacey T. Richards (Midway, UT); Bryan W. Sparks (Lindon, UT); Michael Ephraim Kiemel (Lehi, UT); Max K. Goff (Madison, MS); Eduardo James Sagra (Sandy, UT); Ryan W. McQueen (Sandy, UT)
Assignee: DeepSee.ai Inc.
G06F16/258G06F16/289G06F16/353
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,105,728
App. No.
17/475,204
Granted
Oct 1, 2024
Kind
B2
Abstract

Systems and methods are described herein for creating a data object for each of a plurality of imported unstructured data files. Each data object may expressly include one of the unstructured data files. Preprocessing subsystems and/or machine learning algorithms and subsystems process the data to generate or otherwise identify structured insight features. The system updates each data object to expressly include the structured insight features.

Claims (56)

1. A method, comprising:

accessing a dataset with a plurality of unstructured documents;

creating an extensible data object for each of the unstructured documents;

receiving, from a first preprocessing subsystem, a first structured insight feature associated with a first unstructured document of a first extensible data object;

augmenting the first extensible data object to include the first structured insight feature;

receiving, from a second preprocessing subsystem, a second structured insight feature associated with the first unstructured document of the first extensible data object;

augmenting the first extensible data object to include the second structured insight feature, such that the first extensible data object includes at least two structured insight features;

receiving, via at least a third preprocessing subsystem:

a third structured insight feature that characterizes a relationship between the first unstructured document and a second unstructured document of a second extensible data object, and

a fourth structured insight feature that characterizes a relationship between the first unstructured document and a third unstructured document of a third extensible data object;

augmenting the first extensible data object to include the third and fourth structured insight features characterizing the relationships of the first unstructured document to the second and third unstructured documents, respectively,

such that the first extensible data object includes at least four structured insight features associated with the first unstructured document;

augmenting the second extensible data object associated with the second unstructured data element to include the third structured insight feature;

augmenting the third extensible data object associated with the third unstructured data element to include the fourth structured insight feature;

rendering, for display on an electronic display, a view menu as part of a graphical user interface to facilitate user toggling between a lineage view and a timeline view of the first, second, and third extensible data objects;

receiving, via user input, a selection of the timeline view; and

rendering, for display via the electronic display, in response to the user selection of the timeline view:

(i) icons in orbit rings around at least some of the extensible data objects to visually represent earlier versions thereof, and, simultaneously,

(ii) a timeline that shows the relative time of creation of each version of at least one extensible data object.

2. The method of claim 1 , wherein at least some of the unstructured documents comprise contract documents.

3. The method of claim 1 , wherein the first structured insight feature comprises a structured, machine-readable version of the first unstructured document.

4. The method of claim 3 , wherein the second structured insight feature comprises a document categorization identifying a document type of the first unstructured document.

5. The method of claim 3 , wherein the second structured insight feature comprises one of: a computer-readable version of the first unstructured document, an attribute of the first unstructured document, a subject matter expert insight provided with respect to the first unstructured document, a characteristic of the first unstructured document, a relationship between the first unstructured document and another of the unstructured document, and a language translation of the first unstructured document.

6. A visualization method, comprising:

accessing a database with a plurality of extensible data objects, wherein each extensible data object comprises:

an unstructured document from an unstructured dataset of documents,

a data type insight feature that identifies each respective extensible data object as having one of a plurality of different data types, at least one of which data types comprises a contract data type,

a relationship insight feature that identifies a relationship of the unstructured document with an unstructured document of at least one other extensible data object, and

at least one additional insight feature associated with the unstructured document identified by a data preprocessing subsystem,

wherein at least one of the extensible data objects includes a relationship insight feature that identifies a relationship of the unstructured document with unstructured documents of at least three other extensible data objects;

rendering, for display on an electronic display, a graphical user interface that includes a data type filter menu to facilitate a user selection of one or more data types of the plurality of different data types;

receiving, via a user input, a first data type selection of a contract data type of the plurality of different data types via the data type filter menu;

rendering, for display via the electronic display, the graphical user interface with:

an icon representing each extensible data object that has a data type insight feature corresponding to the selected contract data type, and

connection lines between icons to represent the relationship insight features between the unstructured documents of the extensible data objects represented by the rendered icons having the data type insight feature corresponding to the selected contract data type;

rendering, for display on the electronic display, a view menu as part of the graphical user interface to facilitate user toggling between a lineage view and a timeline view of the extensible data objects;

receiving, via user input, a selection of the timeline view; and

rendering, for display via the electronic display, in response to the user selection of the timeline view:

(i) icons in orbit rings around at least some of the extensible data objects to visually represent earlier versions thereof, and

(ii) a timeline that shows the relative time of creation of each version of at least one extensible data object,

wherein versions of each respective rendered extensible data object are displayed on separate timelines within a primary panel of the graphical user interface.

7. The visualization method of claim 6 , further comprising:

receiving, via an additional user input, a second data type selection of a different data type of the plurality of data types via the data type filter menu; and

rendering, for display via the electronic display, an updated graphical user interface with:

an updated set of icons representing the extensible data objects that have data type insight features corresponding to the second data type selection, and

updated connection lines between icons representing the relationship insight features between the unstructured documents of the extensible data objects represented by the rendered icons having the data type insight feature corresponding to the second data type selection.

8. The visualization method of claim 6 , further comprising:

rendering for display on the electronic display, a version filter menu as part of the graphical user interface to facilitate a user selection of one or more versions for each extensible data object displayed as an icon;

receiving, via an additional user input, a first version selection of at least one version of a plurality of available versions for each extensible data object displayed as an icon; and

rendering, for display via the electronic display, the graphical user interface with:

an updated set of icons representing the extensible data objects that have data type insight features corresponding to:

the first data type selection, and

the first version selection; and

updated connection lines between the updated set of icons.

9. The visualization method of claim 6 , wherein the timeline view further displays the relative time that each insight feature was added to the extensible data object associated with each displayed icon.

10. The visualization method of claim 9 , wherein the timeline view further displays information identifying at least one of a person, entity, or subsystem that caused each insight feature to be added to each respective extensible data object associated with each displayed icon.

Assignments (3)
RELEASE OF SECURITY INTEREST Recorded Nov 3, 2025
From: FIRST-CITIZENS BANK & TRUST COMPANY
To: DEEPSEE.AI INC.
Reel/Frame 072766/0875 →
SECURITY INTEREST Recorded Dec 28, 2023
From: DEEPSEE.AI INC.
To: FIRST-CITIZENS BANK & TRUST COMPANY
Reel/Frame 065967/0789 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 7, 2021
From: SHILLINGFORD, STEPHEN W.; RICHARDS, WACEY T.; SPARKS, BRYAN W.; KIEMEL, MICHAEL EPHRAIM; GOFF, MAX K.; SAGRA, EDUARDO JAMES; MCQUEEN, RYAN W.
To: DEEPSEE.AI INC.
Reel/Frame 057733/0506 →
Continuity (2)
Provisional Application 63077925 · Sep 14, 2020
Related Publication 20220083561A1 · Mar 17, 2022