Granted Patent B2

US 12,210,599 · App. 17/812,430 · Granted Jan 28, 2025

Inventors: Lin Ni Lisa Cheng (New York, NY); Purva Shanker (Arlington, VA); Sara Margaret Hughes (McLean, VA); Allison Trager (Demarest, NJ); Shaun Kieran Webb (Oakland, CA)

Assignee: CAPITAL ONE SERVICES, LLC

G06F21/316G06F21/6218G06N20/00

Abstract

Claims (80)

1. A system for using machine learning models to organize and select access-restricted components accessed using user-specific access tokens with variable properties, the system comprising:

cloud-based storage circuitry for:

storing a plurality of access token profiles; and

a machine learning model, wherein the machine learning model is trained to generate one or more user interface templates for accessing account content based on user behavior characteristics;

cloud-based control circuitry for:

receiving, from a user, a request to access a user interface of an account with a first access token, wherein the first access token comprises an access privilege, a user profile designation, an account designation, and an access time period, wherein the access privilege indicates a privilege granted by the first access token for the account, wherein the user profile designation indicates a user to which the privilege is granted, wherein the account designation identifies the account for which the privilege is granted, and wherein the access time period indicates a time period during which the first access token grants access to the user for the account;

retrieving a user profile corresponding to the user, wherein the user profile includes a user behavior characteristic, wherein the user behavior characteristic indicates a likely user behavior when the access privilege is used to interact with the user interface;

determining accessible content of the account based on the first access token;

generating a first feature input based on the user profile and the accessible content;

inputting the first feature input into the machine learning model, wherein the machine learning model is trained to generate the one or more user interface templates for accessing the accessible content based on the user behavior characteristic;

receiving, from the machine learning model, a first output indicating a recommended user interface template for accessing the account with the first access token, and wherein the recommended user interface template comprises a recommended selection and organization of user input fields and/or user interface pages; and

cloud-based input/output circuitry for generating, for display on a user device, the user interface with the recommended user interface template in response to the user accessing the user interface with the first access token during the access time period.

2. A method for using machine learning models to organize and select access-restricted components accessed using user-specific access tokens with variable properties, the method comprising:

receiving, using cloud-based control circuitry, a request, from a user, to access a user interface of an account with a first access token, wherein the first access token comprises an access privilege and an access time period, wherein the access privilege indicates a privilege granted by the first access token for the account, wherein the first access token comprises a user profile designation, wherein the user profile designation indicates a user granted the privilege, and wherein the access time period indicates a time period during which the first access token grants access to the user for the account;

retrieving, using the cloud-based control circuitry, a user profile corresponding to the user, wherein the user profile includes a user behavior characteristic, wherein the user behavior characteristic indicates a likely user behavior when the access privilege is used to interact with the user interface;

determining, using the cloud-based control circuitry, accessible content of the account based on the first access token;

generating a first feature input based on the user profile and the accessible content for a machine learning model, wherein the machine learning model is trained to generate one or more user interface templates for accessing the accessible content based on the user behavior characteristic;

receiving from the machine learning model a first output, wherein the first output indicates a recommended user interface template for accessing the account with the first access token based on the user behavior characteristic using the machine learning model; and

generating, using the cloud-based control circuitry, for display on a user device, the user interface with the recommended user interface template in response to the user accessing the user interface with the first access token during the access time period.

3. The method of claim 2 , wherein the recommended user interface template comprises a recommended selection and organization of user input fields and/or user interface pages.

4. The method of claim 2 , further comprising:

receiving, from the user, a user input into the user interface;

generating a second feature input based on the user input, the user profile, and the accessible content;

inputting the second feature input into the machine learning model;

receiving from the machine learning model a second output; and

updating the recommended user interface template based on the second output.

5. The method of claim 2 , wherein generating the recommended user interface template comprises:

determining an input field the user likely accesses based on the user behavior characteristic; and

determining a user interface page the user likely accesses based on the user behavior characteristic.

6. The method of claim 2 , wherein generating the recommended user interface template comprises:

determining a respective probability that the user accesses an input field of based on the user behavior characteristic;

comparing the respective probability to a threshold probability; and

determining to include the input field in the recommended user interface template based on the respective probability exceeding the threshold probability.

7. The method of claim 2 , further comprising:

determining a respective probability of an order in which the user accesses a plurality of user interface pages based on the user behavior characteristic;

comparing the respective probability to a threshold probability; and

generating the recommended user interface template based on the order.

8. The method of claim 2 , further comprising:

retrieving a storage address from the first access token, wherein the storage address indicates a location of the recommended user interface template; and

storing the recommended user interface template at the storage address, wherein the storage address is called in response to the user accessing the account with the first access token.

9. The method of claim 2 , further comprising:

retrieving a first additional user profile corresponding to the user from a third party microservice;

retrieving a second additional user profile corresponding to the user from a website cookie; and

aggregating information from the user profile, the first additional user profile, and the second additional user profile.

10. The method of claim 2 , further comprising:

retrieving an access token profile for the first access token, wherein the access token profile includes a token use characteristic, wherein the token use characteristic indicates one or more likely uses of access tokens of given token types, and wherein the recommended user interface template is further based on the user profile.

11. A non-transitory, computer readable medium comprising instructions that when executed on one or more processors cause operations comprising:

receiving a request, from a user, to access a user interface of an account with a first access token, wherein the first access token comprises an access privilege and an access time period, wherein the access privilege indicates a privilege granted by the first access token for the account, wherein the first access token comprises an account designation, wherein the account designation identifies the account to which the privilege is granted, and wherein the access time period indicates a time period during which the first access token grants access to the user for the account;

retrieving a user profile corresponding to the user, wherein the user profile includes a user behavior characteristic, wherein the user behavior characteristic indicates a likely user behavior when the access privilege is used to interact with the user interface;

determining accessible content of the account based on the first access token;

generating a first feature input based on the user profile and the accessible content for a machine learning model, wherein the machine learning model is trained to generate one or more user interface templates for accessing the accessible content based on the user behavior characteristic, and wherein the machine learning model is stored using cloud-based storage circuitry;

receiving, from the machine learning model stored using cloud-based storage circuitry, a first output, wherein the first output indicates a recommended user interface template for accessing the account with the first access token based on the user behavior characteristic using the machine learning model; and

generating, for display on a user device, the user interface with the recommended user interface template in response to the user accessing the user interface with the first access token during the access time period.

12. The non-transitory, computer readable medium of claim 11 , wherein the recommended user interface template comprises a recommended selection and organization of user input fields and/or user interface pages.

13. The non-transitory, computer readable medium of claim 11 , wherein the instructions further cause operations comprising:

receiving, from the user, a user input into the user interface;

generating a second feature input based on the user input, the user profile, and the accessible content;

inputting the second feature input into the machine learning model;

receiving from the machine learning model a second output; and

updating the recommended user interface template based on the second output.

14. The non-transitory, computer readable medium of claim 11 , wherein generating the recommended user interface template comprises:

determining an input field the user likely accesses based on the user behavior characteristic; and

determining a user interface page the user likely accesses based on the user behavior characteristic.

15. The non-transitory, computer readable medium of claim 11 , wherein generating the recommended user interface template comprises:

determining a respective probability that the user accesses an input field of based on the user behavior characteristic;

comparing the respective probability to a threshold probability; and

determining to include the input field in the recommended user interface template based on the respective probability exceeding the threshold probability.

16. The non-transitory, computer readable medium of claim 11 , wherein the instructions further cause operations comprising:

determining a respective probability of an order in which the user accesses a plurality of user interface pages based on the user behavior characteristic;

comparing the respective probability to a threshold probability; and

generating the recommended user interface template based on the order.

17. The non-transitory, computer readable medium of claim 11 , wherein the instructions further cause operations comprising:

retrieving a storage address from the first access token, wherein the storage address indicates a location of the recommended user interface template; and

storing the recommended user interface template at the storage address, wherein the storage address is called in response to the user accessing the account with the first access token.

18. The non-transitory, computer readable medium of claim 11 , wherein the instructions further cause operations comprising:

retrieving a first additional user profile corresponding to the user from a third party microservice;

retrieving a second additional user profile corresponding to the user from a website cookie; and

aggregating information from the user profile, the first additional user profile, and the second additional user profile.

19. The non-transitory, computer readable medium of claim 11 , wherein the instructions further cause operations comprising:

retrieving an access token profile for the first access token, wherein the access token profile includes a token use characteristic, wherein the token use characteristic indicates one or more likely uses of access tokens of given token types, and wherein the recommended user interface template is further based on the user profile.

Assignments (1)

Continuity (1)

References Cited (11)