IP Library › Granted Patent US 12,468,758
Granted Patent B1
US 12,468,758 · App. 18/964,192 · Granted Nov 11, 2025

Attribute recharacterization in individual portions of images

Inventors: Nagaraju Buddhiraju (Concord, NC); Ramakrishna Akula (Darsi, IN); Arjun I T (Malappuram, IN); Balachander Kamatchi (Chennai, IN); Vijay Kumar Yarabolu (Hyderabad, IN); Savleen Kaur (Faridabad, IN); Manikandan Rajaraman (Chennai, IN); Lakshmi Narasimha Prasad Dornala (Hyderabad, IN)
Assignee: Bank of America Corporation
G06F16/55
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 12,468,758
App. No.
18/964,192
Filed
Nov 29, 2024
Granted
Nov 11, 2025
Kind
B1
Examiner
JACOB, AJITH
Art Unit
2161
USPC
707/739
Abstract

An apparatus comprises a memory communicatively coupled to a processor. The processor is configured to generate a tag for a portion of peripheral information, determine a correlation between the peripheral information and the tag and execute a machine learning algorithm in response to determining that an amount of information preserved is outside an accuracy tolerance to determine at least one difference between the peripheral information and the communication information, evaluate the at least one difference against historical data associated with the network device, determine multiple tagging commands based on an evaluation of the at least one difference against the historical data and modify the tag to incorporate the possible modifications, and generate a portion of the communication information based on the portion of the peripheral information in accordance with a modified version of the tag and transmit the portion of the communication information to the network device.

Claims (181)

1 . A system, comprising:

a memory operable to store:

a machine learning algorithm configured to evaluate data in accordance with one or more machine learning models; and

at least one processor communicatively coupled to the memory and configured to:

receive first peripheral information from a network device, the first peripheral information comprising a first format;

generate a first tag for a first portion of the first peripheral information, wherein:

the first tag comprises first guidance to generate first communication information that correlates to the first peripheral information; and

the first communication information comprises a second format;

determine a first correlation between the first peripheral information and the first tag, the first correlation referencing a first amount of information preserved in the first communication information that matches the first peripheral information after the first communication information is generated based on the first peripheral information;

determine whether the first amount of information preserved is outside a first accuracy tolerance;

in response to determining that the first amount of information preserved is outside the first accuracy tolerance, execute the machine learning algorithm to:

determine at least one difference between the first peripheral information and the first communication information;

evaluate the at least one difference against historical data associated with the network device, the historical data comprising patterns associated with one or more previous communication information generated from previous peripheral information received from the network device;

determine a first plurality of tagging commands based on a first evaluation of the at least one difference against the historical data, the first plurality of tagging commands comprising a first plurality of possible modifications to the first tag; and

modify the first tag to incorporate the first plurality of possible modifications;

generate a first portion of the first communication information based on the first portion of the first peripheral information in accordance with a modified version of the first tag; and

transmit the first portion of the first communication information to the network device.

2 . The system of claim 1 , wherein the at least one processor is further configured to:

receive second peripheral information from the network device, the second peripheral information comprising a third format;

generate a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determine a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information; determine whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, execute the machine learning algorithm to:

determine one or more differences between the second peripheral information and the second communication information;

evaluate the one or more differences against the historical data associated with the network device;

determine a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modify the second tag to incorporate the second plurality of possible modifications;

determine a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is outside the second accuracy tolerance, further execute the machine learning algorithm to:

determine one or more additional differences between the second peripheral information and the second communication information;

evaluate the one or more additional differences against the historical data associated with the network device;

determine a third plurality of tagging commands based on a third evaluation of the one or more additional differences against the historical data, the third plurality of tagging commands comprising a third plurality of possible modifications to the second tag; and

modify the second tag to incorporate the third plurality of possible modifications;

generate a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmit the second portion of the second communication information to the network device.

3 . The system of claim 1 , wherein the at least one processor is further configured to:

receive second peripheral information from the network device, the second peripheral information comprising a third format;

generate a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determine a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, execute the machine learning algorithm to:

determine one or more differences between the second peripheral information and the second communication information;

evaluate the one or more differences against the historical data associated with the network device;

determine a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modify the second tag to incorporate the second plurality of possible modifications;

determine a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is not outside the second accuracy tolerance, generate a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmit the second portion of the second communication information to the network device.

4 . The system of claim 1 , wherein the at least one processor is further configured to:

train the one or more machine learning models using the first tag and the first plurality of tagging commands.

5 . The system of claim 1 , wherein:

the first peripheral information is associated with a first peripheral of the network device; and

the first peripheral information is collected after a user interacts with the first peripheral of the network device.

6 . The system of claim 1 , wherein:

the first format of the first peripheral information comprises an image format; and

the second format of the first communication information comprises a sound format.

7 . The system of claim 1 , wherein:

the one or more machine learning models are one or more generative adversarial network models.

8 . A method, comprising:

receiving first peripheral information from a network device, the first peripheral information comprising a first format;

generating a first tag for a first portion of the first peripheral information, wherein:

the first tag comprises first guidance to generate first communication information that correlates to the first peripheral information; and

the first communication information comprises a second format;

determining a first correlation between the first peripheral information and the first tag, the first correlation referencing a first amount of information preserved in the first communication information that matches the first peripheral information after the first communication information is generated based on the first peripheral information;

determining whether the first amount of information preserved is outside a first accuracy tolerance;

in response to determining that the first amount of information preserved is outside the first accuracy tolerance, executing a machine learning algorithm to perform one or more operations comprising:

determining at least one difference between the first peripheral information and the first communication information;

evaluating the at least one difference against historical data associated with the network device, the historical data comprising patterns associated with one or more previous communication information generated from previous peripheral information received from the network device;

determining a first plurality of tagging commands based on a first evaluation of the at least one difference against the historical data, the first plurality of tagging commands comprising a first plurality of possible modifications to the first tag; and

modifying the first tag to incorporate the first plurality of possible modifications;

generating a first portion of the first communication information based on the first portion of the first peripheral information in accordance with a modified version of the first tag; and

transmitting the first portion of the first communication information to the network device.

9 . The method of claim 8 , further comprising:

receiving second peripheral information from the network device, the second peripheral information comprising a third format;

generating a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determining a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information; determine whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, executing the machine learning algorithm to perform one or more first additional operations comprising:

determining one or more differences between the second peripheral information and the second communication information;

evaluating the one or more differences against the historical data associated with the network device;

determining a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modifying the second tag to incorporate the second plurality of possible modifications;

determining a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determining whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is outside the second accuracy tolerance, further executing the machine learning algorithm to perform second additional operations comprising:

determining one or more additional differences between the second peripheral information and the second communication information;

evaluating the one or more additional differences against the historical data associated with the network device;

determining a third plurality of tagging commands based on a third evaluation of the one or more additional differences against the historical data, the third plurality of tagging commands comprising a third plurality of possible modifications to the second tag; and

modifying the second tag to incorporate the third plurality of possible modifications;

generating a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmitting the second portion of the second communication information to the network device.

10 . The method of claim 8 , further comprising:

receiving second peripheral information from the network device, the second peripheral information comprising a third format;

generating a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determining a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determining whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, executing the machine learning algorithm to perform one or more operations comprising:

determining one or more differences between the second peripheral information and the second communication information;

evaluating the one or more differences against the historical data associated with the network device;

determining a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modifying the second tag to incorporate the second plurality of possible modifications;

determining a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determining whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is not outside the second accuracy tolerance, generating a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmitting the second portion of the second communication information to the network device.

11 . The method of claim 8 , further comprising:

training one or more machine learning models using the first tag and the first plurality of tagging commands.

12 . The method of claim 8 , wherein:

the first peripheral information is associated with a first peripheral of the network device; and

the first peripheral information is collected after a user interacts with the first peripheral of the network device.

13 . The method of claim 8 , wherein:

the first format of the first peripheral information comprises an image format; and

the second format of the first communication information comprises a sound format.

14 . The method of claim 8 , wherein:

one or more machine learning models used to guide the machine learning algorithm are one or more generative adversarial network models.

15 . A non-transitory computer-readable medium storing instructions that when executed by a processor cause the processor to:

receive first peripheral information from a network device, the first peripheral information comprising a first format;

generate a first tag for a first portion of the first peripheral information, wherein:

the first tag comprises first guidance to generate first communication information that correlates to the first peripheral information; and

the first communication information comprises a second format;

determine a first correlation between the first peripheral information and the first tag, the first correlation referencing a first amount of information preserved in the first communication information that matches the first peripheral information after the first communication information is generated based on the first peripheral information;

determine whether the first amount of information preserved is outside a first accuracy tolerance;

in response to determining that the first amount of information preserved is outside the first accuracy tolerance, execute a machine learning algorithm to:

determine at least one difference between the first peripheral information and the first communication information;

evaluate the at least one difference against historical data associated with the network device, the historical data comprising patterns associated with one or more previous communication information generated from previous peripheral information received from the network device;

determine a first plurality of tagging commands based on a first evaluation of the at least one difference against the historical data, the first plurality of tagging commands comprising a first plurality of possible modifications to the first tag; and

modify the first tag to incorporate the first plurality of possible modifications;

generate a first portion of the first communication information based on the first portion of the first peripheral information in accordance with a modified version of the first tag; and

transmit the first portion of the first communication information to the network device.

16 . The non-transitory computer-readable medium of claim 15 , wherein, when executed by the processor, the instructions further cause the processor to:

receive second peripheral information from the network device, the second peripheral information comprising a third format;

generate a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determine a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information; determine whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, execute the machine learning algorithm to:

determine one or more differences between the second peripheral information and the second communication information;

evaluate the one or more differences against the historical data associated with the network device;

determine a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modify the second tag to incorporate the second plurality of possible modifications;

determine a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is outside the second accuracy tolerance, further executing the machine learning algorithm to:

determine one or more additional differences between the second peripheral information and the second communication information;

evaluate the one or more additional differences against the historical data associated with the network device;

determine a third plurality of tagging commands based on a third evaluation of the one or more additional differences against the historical data, the third plurality of tagging commands comprising a third plurality of possible modifications to the second tag; and

modify the second tag to incorporate the third plurality of possible modifications;

generate a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmit the second portion of the second communication information to the network device.

17 . The non-transitory computer-readable medium of claim 15 , wherein, when executed by the processor, the instructions further cause the processor to:

receive second peripheral information from the network device, the second peripheral information comprising a third format;

generate a second tag for a second portion of the second peripheral information, wherein:

the second tag comprises second guidance to generate second communication information that correlates to the second peripheral information; and

the second communication information comprises a fourth format;

determine a second correlation between the second peripheral information and the second tag, the second correlation referencing a second amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the second amount of information preserved is outside a second accuracy tolerance;

in response to determining that the second amount of information preserved is outside the second accuracy tolerance, execute the machine learning algorithm to:

determine one or more differences between the second peripheral information and the second communication information;

evaluate the one or more differences against the historical data associated with the network device;

determine a second plurality of tagging commands based on a second evaluation of the one or more differences against the historical data, the second plurality of tagging commands comprising a second plurality of possible modifications to the second tag; and

modify the second tag to incorporate the second plurality of possible modifications;

determine a third correlation between the second peripheral information and a modified version of the second tag, the third correlation referencing a third amount of information preserved in the second communication information that matches the second peripheral information after the second communication information is generated based on the second peripheral information;

determine whether the third amount of information preserved is outside the second accuracy tolerance;

in response to determining that the third amount of information preserved is not outside the second accuracy tolerance, generate a second portion of the second communication information based on the second portion of the second peripheral information in accordance with a modified version of the second tag; and

transmit the second portion of the second communication information to the network device.

18 . The non-transitory computer-readable medium of claim 15 , wherein, when executed by the processor, the instructions further cause the processor to:

train one or more machine learning models using the first tag and the first plurality of tagging commands.

19 . The non-transitory computer-readable medium of claim 15 , wherein:

the first peripheral information is associated with a first peripheral of the network device; and

the first peripheral information is collected after a user interacts with the first peripheral of the network device.

20 . The non-transitory computer-readable medium of claim 15 , wherein:

the first format of the first peripheral information comprises an image format; and

the second format of the first communication information comprises a sound format.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 29, 2024
From: BUDDHIRAJU, NAGARAJU; AKULA, RAMAKRISHNA; I T, ARJUN; KAMATCHI, BALACHANDER; YARABOLU, VIJAY KUMAR; KAUR, SAVLEEN; RAJARAMAN, MANIKANDAN; DORNALA, LAKSHMI NARASIMHA PRASAD
To: BANK OF AMERICA CORPORATION
Reel/Frame 069463/0591 →
References Cited (27)
US 7823055B2 · Sull et al. · 2010 [cited by applicant]
US 8005935B2 · Pradhan et al. · 2011 [cited by applicant]
US 8171032B2 · Herz · 2012 [cited by applicant]
US 8370362B2 · Szabo · 2013 [cited by applicant]
US 10401732B2 · Hsu et al. · 2019 [cited by applicant]
US 10452923B2 · Block · 2019 [cited by applicant]
US 10911468B2 · Muddu et al. · 2021 [cited by applicant]
US 11181893B2 · Cella et al. · 2021 [cited by applicant]
US 11232558B2 · Liu et al. · 2022 [cited by applicant]
US 11482320B2 · Klaiman · 2022 [cited by applicant]
US 11631029B2 · Puri et al. · 2023 [cited by applicant]
US 11714831B2 · Makhija et al. · 2023 [cited by applicant]
US 11948314B2 · Liao et al. · 2024 [cited by applicant]
US 11954112B2 · Siebel et al. · 2024 [cited by applicant]
US 20070041317A1 · Haley et al. · 2007 [cited by applicant]
US 20210192385A1 · Farré Guiu · 2021 [cited by examiner]
US 20220019855A1 · Chen et al. · 2022 [cited by applicant]
US 20220293285A1 · Lamb et al. · 2022 [cited by applicant]
Americans with Disabilities Act, https://www.dol.gov/general/topic/disability/ada, Printed: Nov. 19, 2024, pp. 1-2. [cited by applicant]
Sarah Berry, What is ADA Compliance? (And What Does ADA Compliance Mean for Your Website?), Article, Mar. 31, 2023, pp. 1-20. [cited by applicant]
What is a GAN?, https://aws.amazon.com/what-is/gan/, Printed: Nov. 19, 2024, pp. 1-9. [cited by applicant]
Chau Pham, Graph Convolutional Networks (GCN), Article, Oct. 22, 2020, pp. 1-16. [cited by applicant]
Matthew N. Bernstein, Graph Convolutional Neural Networks, Article, Sep. 24, 2023, pp. 1-22. [cited by applicant]
Nagaraju Buddhiraju, Machine learning training sample generation using generative adversarial networks, U.S. Appl. No. 18/964,163, filed Nov. 29, 2024, pp. 1-86. [cited by applicant]
Nagaraju Buddhiraju, Node generation in network graphs based on characterization attributes in portions of images, U.S. Appl. No. 18/964,175, filed Nov. 29, 2024, pp. 1-89. [cited by applicant]
Nagaraju Buddhiraju, On-demand node annotation in network graphs, U.S. Appl. No. 18/964,208, filed Nov. 29, 2024, pp. 1-82. [cited by applicant]
Nagaraju Buddhiraju, Query analysis using machine learning-driven APIs, U.S. Appl. No. 18/964,222, filed Nov. 29, 2024, pp. 1-80. [cited by applicant]