Granted Patent B2

US 12,657,294 · App. 18/796,050 · Granted Jun 16, 2026

Inventors: Sakshi Bakshi (New Delhi, IN); Sneha Padiyar (East Mumbai, IN); Hariharan Balakrishnan (Chennai, IN)

Assignee: Bank of America Corporation

G06F21/552G06F21/6254H04L9/008

Abstract

Claims (117)

1 . A method for enhancing data privacy and security in a digital communication and storage system, comprising:

generating, by identity nodes within bidirectional dual node hashchain bundles, hash keys by processing partial metadata from different datasets and linking the hash keys through identity graphs, wherein the identity nodes are configured to generate partial metadata from different datasets and create hash keys that link through identity graphs to ensure secure and efficient metadata linking;

performing, by translation nodes within the bidirectional dual node hashchain bundles, identity resolution, data anonymization, and re-keying by transforming original identifiers into anonymous IDs and re-key IDs, wherein the translation nodes are configured to handle anonymization and re-keying by converting ingested data into anonymous IDs and re-key IDs to maintain data privacy;

introducing, by Laplace noise enabler capsules within each bidirectional dual node hashchain bundle, Laplace noise into translation layer logs to ensure differential privacy by adding statistical noise to the translation layer logs, wherein the Laplace noise enabler capsules are configured to utilize the Laplace mechanism to add noise to data logs for anonymizing data resistant to reverse engineering;

sharing, by bidirectional dual node hashchains, metadata between adjacent bundles to establish cryptographic links for secure and efficient data transmission, wherein the bidirectional dual node hashchains are configured to share metadata between adjacent bundles, which are bundles linked bidirectionally for metadata sharing, creating a resilient and scalable network with cryptographic links that are hash-based connections ensuring data integrity even if one node is compromised;

receiving, by a cognitive analytics layer shared among all bidirectional dual node hashchain bundles, hashed encrypted data packets from various bundles and verifying the data packets during decryption using a gossip protocol for real-time validation of data integrity, wherein the cognitive analytics layer is configured to receive hashed encrypted data packets from the bidirectional dual node hashchain bundles and verify them using the gossip protocol for decentralized real-time validation;

activating, by an analytics workspace layer shared among all bidirectional dual node hashchain bundles, re-key IDs and creating digital tags for datasets using BERT transformers to match re-key IDs for secure data sharing, wherein the analytics workspace layer is configured to use BERT transformers for bidirectional reading and validation of data to activate re-key IDs and create digital tags as unique identifiers for the datasets;

linking, by digital tags created within the analytics workspace layer, data through knowledge graphs for optimized performance and secure access to information, wherein the knowledge graphs are configured to link data through the digital tags for secure sharing and consumption of information;

managing, by bidirectional dual node hashchains and the gossip protocol, parallel processing of data across the system to ensure optimized performance and enhance system resilience against cyber-attacks, wherein the bidirectional dual node hashchains and gossip protocol are configured to enable parallel processing by operating bundles as loosely coupled packages for handling data in parallel while maintaining security;

wherein translation layers and entity layers within each bidirectional dual node hashchain bundle work in conjunction with the shared analytics layer to collectively perform data clean room functionality by:

the translation layers converting data into a common format for consistency and compatibility across the system, performing data ingestion, standardization, cleaning, and enrichment, wherein the translation layers are configured to convert data into a common format for consistency and compatibility, performing data ingestion, standardization, cleaning, and enrichment;

the entity layers organizing data around specific entities, handling entity resolution to identify and merge records referring to the same entity, and maintaining a central repository of entities to ensure accurate data representation and linking, wherein the entity layers are configured to organize data around specific entities and handle entity resolution by identifying and merging records referring to the same entity;

the shared analytics layer leveraging standardized and entity-organized data to perform various types of analyses, including descriptive and predictive analytics, and utilizing BERT transformers for advanced identity resolution and digital tag creation to facilitate secure data sharing and consumption, wherein the shared analytics layer is configured to leverage standardized and entity-organized data for analyses using BERT transformers;

utilizing, by the translation nodes, machine learning modules to improve accuracy of identity resolution and anonymization processes, continuously training the machine learning modules on newly ingested data to adapt to evolving patterns and enhance performance, wherein the machine learning modules are configured to improve accuracy of identity resolution and anonymization processes by continuously training on newly ingested data;

integrating, by the translation nodes, homomorphic encryption techniques to allow computations on encrypted data without decrypting it, thereby maintaining data privacy throughout processing stages, thereby maintaining data privacy throughout processing stages;

employing, by the cognitive analytics layer, federated learning to enable collaborative model training across multiple nodes without sharing raw data, thereby enhancing data privacy while improving the machine learning model's accuracy and robustness, wherein the federated learning is configured to enable collaborative model training across multiple nodes without sharing raw data;

implementing, by the cognitive analytics layer, zero-knowledge proofs to verify data integrity and authenticity, ensuring that transactions and data manipulations can be validated without revealing the actual data, wherein the zero-knowledge proofs are configured to verify data integrity and authenticity without revealing the actual data;

enhancing, by the BERT transformers in the analytics workspace layer, digital tags with metadata attributes describing data origin, transformation history, and access controls to provide comprehensive traceability and governance;

utilizing, by an advanced access control mechanism, attribute-based encryption (ABE) to manage user permissions and data access within the system, ensuring that only authorized users can access specific data segments, wherein the attribute-based encryption is configured to manage user permissions based on attributes; and

including, by the analytics workspace layer, an anomaly detection module powered by machine learning algorithms to continuously monitor data activities and identify potential security threats or unusual patterns, wherein the anomaly detection module is configured to identify potential security threats or unusual patterns using machine learning algorithms.

2 . The method of claim 1 , wherein the translation nodes further include a machine learning module configured to perform continuous learning and adaptation based on newly ingested data.

3 . The method of claim 2 , wherein the machine learning module is trained using a diverse dataset to enhance robustness and performance across various data types and patterns.

4 . The method of claim 3 , wherein the translation nodes implement homomorphic encryption techniques, allowing for computations to be performed on encrypted data without need for decryption.

5 . The method of claim 4 , wherein the cognitive analytics layer employs federated learning to enable collaborative model training across multiple nodes without sharing raw data.

6 . The method of claim 5 , wherein the cognitive analytics layer implements zero-knowledge proofs to verify data integrity and authenticity, ensuring that data transactions and manipulations can be validated without revealing the actual data.

7 . The method of claim 6 , wherein the BERT transformers in the analytics workspace layer are further configured to enhance digital tags with metadata attributes that describe the data origin, transformation history, and access controls, providing comprehensive traceability and governance.

8 . The method of claim 7 , wherein the advanced access control mechanism utilizes attribute-based encryption (ABE) to manage user permissions and data access within the system, ensuring that only authorized users can access specific data segments.

9 . The method of claim 8 , wherein the analytics workspace layer includes an anomaly detection module powered by machine learning algorithms to continuously monitor data activities and identify potential security threats or unusual patterns.

10 . The method of claim 9 , further comprising periodically updating the identity graphs to reflect changes in the metadata, ensuring continuous integrity and security of the data.

11 . The method of claim 10 , wherein the Laplace noise introduced by the Laplace noise enabler capsules is generated by sampling from a Laplace distribution with a scale parameter determined by sensitivity of the query and a chosen privacy parameter.

12 . The method of claim 11 , further comprising verifying effectiveness of noise addition by performing privacy risk assessments on encapsulated data, ensuring compliance with data privacy regulations.

13 . The method of claim 12 , wherein the bidirectional dual node hashchains periodically audit the cryptographic links to ensure their integrity and detect any unauthorized modifications, maintaining secure and efficient data transmission.

14 . The method of claim 13 , wherein the gossip protocol used by the cognitive analytics layer is configured to operate with randomized peer selection to enhance security and prevent targeted attacks.

15 . The method of claim 14 , wherein the digital tags created within the analytics workspace layer are used to generate knowledge graphs that map relationships between different data points and entities within the system, ensuring accurate and up-to-date tagging.

16 . The method of claim 15 , further comprising continuously updating the knowledge graphs to reflect new data and changes in existing data.

17 . The method of claim 16 , wherein each bidirectional dual node hashchain bundle further includes an artificial intelligence engine configured to work in conjunction with the Laplace hashed noise enabler, wherein the artificial intelligence engine:

analyzes incoming data within the bundle to identify patterns and anomalies before the data is anonymized;

optimizes introduction of Laplace noise based on detected patterns and anomalies to enhance data privacy without significantly compromising data utility;

continuously learns and adapts from the anonymized data to improve future noise application strategies, ensuring a balance between privacy and data usability; and

assists in transforming the anonymized data into structured formats suitable for further processing by the shared analytics layer.

18 . A method for enhancing data privacy and security in a digital communication and storage system, comprising:

generating, by identity nodes within bidirectional dual node hashchain bundles, hash keys by processing partial metadata from different datasets and linking the hash keys through identity graphs, wherein the identity nodes are configured to generate partial metadata from different datasets and create hash keys that link through identity graphs to ensure secure and efficient metadata linking;

performing, by translation nodes within the bidirectional dual node hashchain bundles, identity resolution, data anonymization, and re-keying by transforming original identifiers into anonymous IDs and re-key IDs, wherein the translation nodes are configured to handle anonymization and re-keying by converting ingested data into anonymous IDs and re-key IDs to maintain data privacy;

introducing, by Laplace noise enabler capsules within each bidirectional dual node hashchain bundle, Laplace noise into translation layer logs to ensure differential privacy by adding statistical noise to the translation layer logs, wherein the Laplace noise enabler capsules are configured to utilize the Laplace mechanism to add noise to data logs for anonymizing data resistant to reverse engineering;

sharing, by bidirectional dual node hashchains, metadata between adjacent bundles to establish cryptographic links for secure and efficient data transmission, wherein the bidirectional dual node hashchains are configured to share metadata between adjacent bundles, which are bundles linked bidirectionally for metadata sharing, creating a resilient and scalable network with cryptographic links that are hash-based connections ensuring data integrity even if one node is compromised;

receiving, by a cognitive analytics layer shared among all bidirectional dual node hashchain bundles, hashed encrypted data packets from various bundles and verifying the data packets during decryption using a gossip protocol for real-time validation of data integrity, wherein the cognitive analytics layer is configured to receive hashed encrypted data packets from the bidirectional dual node hashchain bundles and verify them using the gossip protocol for decentralized real-time validation;

activating, by an analytics workspace layer shared among all bidirectional dual node hashchain bundles, re-key IDs and creating digital tags for datasets using BERT transformers to match re-key IDs for secure data sharing, wherein the analytics workspace layer is configured to use BERT transformers for bidirectional reading and validation of data to activate re-key IDs and create digital tags as unique identifiers for the datasets;

linking, by digital tags created within the analytics workspace layer, data through knowledge graphs for optimized performance and secure access to information, wherein the knowledge graphs are configured to link data through the digital tags for secure sharing and consumption of information;

managing, by bidirectional dual node hashchains and the gossip protocol, parallel processing of data across the system to ensure optimized performance and enhance system resilience against cyber-attacks;

wherein translation layers and entity layers within each bidirectional dual node hashchain bundle work in conjunction with the shared analytics layer to collectively perform data clean room functionality by:

the translation layers converting data into a common format for consistency and compatibility across the system, performing data ingestion, standardization, cleaning, and enrichment, wherein the translation layers are configured to convert data into a common format for consistency and compatibility, performing data ingestion, standardization, cleaning, and enrichment;

the entity layers organizing data around specific entities, handling entity resolution to identify and merge records referring to the same entity, and maintaining a central repository of entities to ensure accurate data representation and linking, wherein the entity layers are configured to organize data around specific entities and handle entity resolution by identifying and merging records referring to the same entity;

the shared analytics layer leveraging standardized and entity-organized data to perform various types of analyses, including descriptive and predictive analytics, and utilizing BERT transformers for advanced identity resolution and digital tag creation to facilitate secure data sharing and consumption;

utilizing, by the translation nodes, machine learning modules to improve accuracy of identity resolution and anonymization processes, continuously training the machine learning modules on newly ingested data to adapt to evolving patterns and enhance performance, wherein the machine learning modules are configured to improve accuracy of identity resolution and anonymization processes by continuously training on newly ingested data;

integrating, by the translation nodes, homomorphic encryption techniques to allow computations on encrypted data without decrypting it;

employing, by the cognitive analytics layer, federated learning to enable collaborative model training across multiple nodes without sharing raw data, wherein the federated learning is configured to enable collaborative model training across multiple nodes without sharing raw data;

implementing, by the cognitive analytics layer, zero-knowledge proofs to verify data integrity and authenticity, ensuring that transactions and data manipulations can be validated without revealing the actual data, wherein the zero-knowledge proofs are configured to verify data integrity and authenticity without revealing the actual data;

enhancing, by the BERT transformers in the analytics workspace layer, digital tags with metadata attributes describing data origin, transformation history, and access controls to provide comprehensive traceability and governance;

utilizing, by an advanced access control mechanism, attribute-based encryption (ABE) to manage user permissions and data access within the system, ensuring that only authorized users can access specific data segments, wherein the attribute-based encryption is configured to manage user permissions based on attributes;

including, by the analytics workspace layer, an anomaly detection module powered by machine learning algorithms to continuously monitor data activities and identify potential security threats or unusual patterns, wherein the anomaly detection module is configured to identify potential security threats or unusual patterns using machine learning algorithms;

periodically updating the identity graphs to reflect changes in the metadata, ensuring continuous integrity and security of the data;

wherein the Laplace noise introduced by the Laplace noise enabler capsules is generated by sampling from a Laplace distribution with a scale parameter determined by sensitivity of the query and a chosen privacy parameter;

verifying effectiveness of noise addition by performing privacy risk assessments on encapsulated data, ensuring compliance with data privacy regulations;

periodically auditing the cryptographic links in the bidirectional dual node hashchains to ensure their integrity and detect any unauthorized modifications, maintaining secure and efficient data transmission;

configuring the gossip protocol used by the cognitive analytics layer to operate with randomized peer selection to enhance security and prevent targeted attacks;

using the digital tags created within the analytics workspace layer to generate knowledge graphs that map relationships between different data points and entities within the system, ensuring accurate and up-to-date tagging;

continuously updating the knowledge graphs to reflect new data and changes in existing data;

wherein each bidirectional dual node hashchain bundle further includes an artificial intelligence engine configured to work in conjunction with the Laplace hashed noise enabler, wherein the artificial intelligence engine:

analyzes incoming data within the bundle to identify patterns and anomalies before the data is anonymized;

optimizes introduction of Laplace noise based on detected patterns and anomalies to enhance data privacy without significantly compromising data utility;

continuously learns and adapts from the anonymized data to improve future noise application strategies, ensuring a balance between privacy and data usability; and

assists in transforming the anonymized data into structured formats suitable for further processing by the shared analytics layer.

19 . A system for enhancing data privacy and security in a digital communication and storage system, comprising:

one or more processors and memory storing instructions that, when executed by the one or more processors, cause the system to perform operations including:

identity nodes within bidirectional dual node hashchain bundles, configured to generate hash keys by processing partial metadata from different datasets and linking the hash keys through identity graphs, wherein the identity nodes are configured to generate partial metadata from different datasets and create hash keys that link through identity graphs to ensure secure and efficient metadata linking;

translation nodes within the bidirectional dual node hashchain bundles, configured to perform identity resolution, data anonymization, and re-keying by transforming original identifiers into anonymous IDs and re-key IDs, wherein the translation nodes are configured to handle anonymization and re-keying by converting ingested data into anonymous IDs and re-key IDs to maintain data privacy;

Laplace noise enabler capsules within each bidirectional dual node hashchain bundle, configured to introduce Laplace noise into translation layer logs to ensure differential privacy by adding statistical noise to the translation layer logs, wherein the Laplace noise enabler capsules are configured to utilize the Laplace mechanism to add noise to data logs for anonymizing data resistant to reverse engineering;

bidirectional dual node hashchains, configured to share metadata between adjacent bundles to establish cryptographic links for secure and efficient data transmission, wherein the bidirectional dual node hashchains are configured to share metadata between adjacent bundles, which are bundles linked bidirectionally for metadata sharing, creating a resilient and scalable network with cryptographic links that are hash-based connections ensuring data integrity even if one node is compromised;

a cognitive analytics layer shared among all bidirectional dual node hashchain bundles, configured to receive hashed encrypted data packets from various bundles and verify the data packets during decryption using a gossip protocol for real-time validation of data integrity, wherein the cognitive analytics layer is configured to receive hashed encrypted data packets from the bidirectional dual node hashchain bundles and verify them using the gossip protocol for decentralized real-time validation;

an analytics workspace layer shared among all bidirectional dual node hashchain bundles, configured to activate re-key IDs and create digital tags for datasets using BERT transformers to match re-key IDs for secure data sharing, wherein the analytics workspace layer is configured to use BERT transformers for bidirectional reading and validation of data to activate re-key IDs and create digital tags as unique identifiers for the datasets;

digital tags created within the analytics workspace layer, configured to link data through knowledge graphs for optimized performance and secure access to information, wherein the bidirectional dual node hashchains and gossip protocol are configured to enable parallel processing by operating bundles as loosely coupled packages for handling data in parallel while maintaining security;

bidirectional dual node hashchains and the gossip protocol, configured to manage parallel processing of data across the system to ensure optimized performance and enhance system resilience against cyber-attacks;

wherein translation layers and entity layers within each bidirectional dual node hashchain bundle work in conjunction with the shared analytics layer to collectively perform data clean room functionality by:

the translation layers converting data into a common format for consistency and compatibility across the system, performing data ingestion, standardization, cleaning, and enrichment, wherein the translation layers are configured to convert data into a common format for consistency and compatibility, performing data ingestion, standardization, cleaning, and enrichment;

the entity layers organizing data around specific entities, handling entity resolution to identify and merge records referring to the same entity, and maintaining a central repository of entities to ensure accurate data representation and linking, wherein the entity layers are configured to organize data around specific entities and handle entity resolution by identifying and merging records referring to the same entity;

the shared analytics layer leveraging standardized and entity-organized data to perform various types of analyses, including descriptive and predictive analytics, and utilizing BERT transformers for advanced identity resolution and digital tag creation to facilitate secure data sharing and consumption, wherein the shared analytics layer is configured to leverage standardized and entity-organized data for analyses using BERT transformers;

translation nodes further including machine learning modules configured to improve accuracy of identity resolution and anonymization processes, continuously training the machine learning modules on newly ingested data to adapt to evolving patterns and enhance performance, wherein the machine learning modules are configured to improve accuracy of identity resolution and anonymization processes by continuously training on newly ingested data;

translation nodes integrating homomorphic encryption techniques to allow computations on encrypted data without decrypting it;

the cognitive analytics layer employing federated learning to enable collaborative model training across multiple nodes without sharing raw data, wherein the federated learning is configured to enable collaborative model training across multiple nodes without sharing raw data;

the cognitive analytics layer implementing zero-knowledge proofs to verify data integrity and authenticity, ensuring that data transactions and manipulations can be validated without revealing the actual data, wherein the zero-knowledge proofs are configured to verify data integrity and authenticity without revealing the actual data;

BERT transformers in the analytics workspace layer enhancing digital tags with metadata attributes describing data origin, transformation history, and access controls to provide comprehensive traceability and governance;

an advanced access control mechanism utilizing attribute-based encryption (ABE) to manage user permissions and data access within the system, ensuring that only authorized users can access specific data segments, wherein the attribute-based encryption is configured to manage user permissions based on attributes;

the analytics workspace layer including an anomaly detection module powered by machine learning algorithms to continuously monitor data activities and identify potential security threats or unusual patterns, wherein the anomaly detection module is configured to identify potential security threats or unusual patterns using machine learning algorithms;

the identity graphs periodically updated to reflect changes in the metadata, ensuring continuous integrity and security of the data;

the Laplace noise introduced by the Laplace noise enabler capsules generated by sampling from a Laplace distribution with a scale parameter determined by sensitivity of the query and a chosen privacy parameter;

effectiveness of noise addition verified by performing privacy risk assessments on encapsulated data, ensuring compliance with data privacy regulations;

the cryptographic links in the bidirectional dual node hashchains periodically audited to ensure their integrity and detect any unauthorized modifications, maintaining secure and efficient data transmission;

the gossip protocol used by the cognitive analytics layer configured to operate with randomized peer selection to enhance security and prevent targeted attacks;

the digital tags created within the analytics workspace layer used to generate knowledge graphs that map relationships between different data points and entities within the system, ensuring accurate and up-to-date tagging;

the knowledge graphs continuously updated to reflect new data and changes in existing data;

an artificial intelligence engine within each bidirectional dual node hashchain bundle, configured to work in conjunction with the Laplace hashed noise enabler, wherein the artificial intelligence engine:

analyzes incoming data within the bundle to identify patterns and anomalies before the data is anonymized;

optimizes introduction of Laplace noise based on detected patterns and anomalies to enhance data privacy without significantly compromising data utility;

continuously learns and adapts from the anonymized data to improve future noise application strategies, ensuring a balance between privacy and data usability; and

assists in transforming the anonymized data into structured formats suitable for further processing by the shared analytics layer.

20 . The system of claim 19 , wherein the system further comprises:

one or more processors and memory storing instructions that, when executed by the one or more processors, cause the system to perform operations including:

a blockchain framework integrated with the bidirectional dual node hashchains, configured to enhance the security and immutability of the hash keys and metadata by securely recording each transaction and ensuring that it cannot be altered without detection;

a distributed ledger system operable within the blockchain framework, configured to maintain a verifiable and immutable record of all data transactions and transformations;

an advanced machine learning module within the cognitive analytics layer, configured to analyze historical data patterns and predict future trends;

a multi-factor authentication mechanism integrated with the advanced access control mechanism, providing an additional layer of security by requiring multiple forms of verification before granting access to sensitive data segments;

a dynamic resource allocation system, configured to monitor real-time data processing demands and allocate computational resources accordingly, ensuring that the system maintains high performance and efficiency under varying workloads;

a secure communication protocol, configured to encrypt data transmissions between nodes and prevent unauthorized interception or tampering during data exchange;

a comprehensive data compliance module, configured to monitor and ensure adherence to relevant data privacy regulations and standards, including GDPR, CCPA, and HIPAA;

a user-friendly interface, designed to facilitate interaction with the system by providing intuitive access controls, real-time data analytics dashboards, and detailed reporting tools;

a redundant backup system, configured to periodically back up all critical data and system configurations, ensuring data recovery and continuity in event of system failure or cyber-attacks; and

a continuous monitoring and auditing module, configured to track system performance, detect anomalies, and generate alerts for potential security threats or operational issues.

Assignments (1)

Continuity (1)

References Cited (38)