Patents by Inventor Kai Chung CHEUNG
Kai Chung CHEUNG has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11956374Abstract: A computing system that is configured for a federated wallet with cryptographically secure signature delegation. The system may be configured to receive a session public key corresponding to a decentralized application and a user. The system may be further configured to receive an unsigned transaction of a blockchain, the unsigned transaction corresponding to the user. The system may be further configured to provide a symmetric encryption key to the user's device for encrypting the user's private signing key. The system may be further configured to determine, using the session public key, that the unsigned transaction is valid. Based on the validity of the unsigned transaction, the system may send the unsigned transaction to the user's device. The system may send the symmetric encryption key to the user's device to decrypt the private signing key. The system may be further configured to receive a signed transaction for submission to the blockchain.Type: GrantFiled: April 14, 2023Date of Patent: April 9, 2024Assignee: Via Science, Inc.Inventors: Jesús Alejandro Cárdenes Cabré, Arteum Kanda, Jeremy Taylor, John Christopher Muddle, Kai Chung Cheung
-
Publication number: 20230336361Abstract: A computing system that is configured for a federated wallet with cryptographically secure signature delegation. The system may be configured to receive a session public key corresponding to a decentralized application and a user. The system may be further configured to receive an unsigned transaction of a blockchain, the unsigned transaction corresponding to the user. The system may be further configured to provide a symmetric encryption key to the user's device for encrypting the user's private signing key. The system may be further configured to determine, using the session public key, that the unsigned transaction is valid. Based on the validity of the unsigned transaction, the system may send the unsigned transaction to the user's device. The system may send the symmetric encryption key to the user's device to decrypt the private signing key. The system may be further configured to receive a signed transaction for submission to the blockchain.Type: ApplicationFiled: April 14, 2023Publication date: October 19, 2023Inventors: Jesús Alejandro Cárdenes Cabré, Arteum Kanda, Jeremy Taylor, John Christopher Muddle, Kai Chung Cheung
-
Patent number: 11695557Abstract: A first component determines encrypted data representing an event and encrypted threshold data corresponding to an outlier of the event. The first system may process the data using, for example, one or more composite integers, and may send the result to a second system. This second system may subtract the data to determine of the encrypted data is greater than, less than, or equal to the encrypted threshold. If so, the second system may determine that the encrypted data corresponds to an outlier of the data. The second system may send an indication of this determination to a third system.Type: GrantFiled: June 11, 2021Date of Patent: July 4, 2023Assignee: Via Science, Inc.Inventors: Kai Chung Cheung, Jeremy Taylor, Mathew Rogers, Colin Gounden
-
Publication number: 20230121425Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.Type: ApplicationFiled: October 27, 2022Publication date: April 20, 2023Inventors: Kai Chung Cheung, Mathew Rogers, Jeremy Taylor
-
Patent number: 11630913Abstract: A first system receives an encrypted data vector representing a text search query from a second system and second encrypted data from a third system that may include a first vector and a second vector representing text of an electronic document. The first system may multiply the vectors by a random vector. The first system may determine a first difference between the encrypted data vector and the first vector, and a second difference between the encrypted data vector and the second vector. The first system may determine a product of the first and second difference. The first system may send the product to the third system and then receive a value representing the decrypted difference. The first system may determine if the value satisfies a condition and send the result of the determination to the second system.Type: GrantFiled: May 23, 2022Date of Patent: April 18, 2023Assignee: Via Science, Inc.Inventors: Madjid Aoudia, Kai Chung Cheung, Jesús Alejandro Cárdenes Cabré
-
Patent number: 11586743Abstract: A first system creates and sends encryption key data to multiple data sources. A second system receives data encrypted using the encryption key data from the multiple data sources; the data may include noise data such that, even if decrypted, the original data cannot be discovered. Because the encryption is additively homomorphic, the second system may create encrypted summation data using the encrypted data. The first system separately receives the noise data encrypted using the same technique as the encrypted data. The second system may send the encrypted summation data to the first system, which may then remove the noise data from the encrypted summation data to create unencrypted summation data.Type: GrantFiled: March 22, 2019Date of Patent: February 21, 2023Assignee: Via Science, Inc.Inventors: Kai Chung Cheung, Jeremy Taylor, Jesús Alejandro Cárdenes Cabré
-
Publication number: 20220382900Abstract: A first system receives an encrypted data vector representing a text search query from a second system and second encrypted data from a third system that may include a first vector and a second vector representing text of an electronic document. The first system may multiply the vectors by a random vector. The first system may determine a first difference between the encrypted data vector and the first vector, and a second difference between the encrypted data vector and the second vector. The first system may determine a product of the first and second difference. The first system may send the product to the third system and then receive a value representing the decrypted difference. The first system may determine if the value satisfies a condition and send the result of the determination to the second system.Type: ApplicationFiled: May 23, 2022Publication date: December 1, 2022Inventors: Madjid Aoudia, Kai Chung Cheung, Jesús Alejandro Cárdenes Cabré
-
Patent number: 11489659Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.Type: GrantFiled: October 29, 2020Date of Patent: November 1, 2022Assignee: Via Science, Inc.Inventors: Kai Chung Cheung, Mathew Rogers, Jeremy Taylor
-
Patent number: 11308222Abstract: Computer instructions corresponding to a neural-network model are received and encrypted using an encryption technique. Training data encrypted using the encryption technique is received from a data source. The model is trained using the training data using, for example, a gradient descent technique. If the model performs in accordance with a quality metric, it is sent to a device of a model user.Type: GrantFiled: March 22, 2019Date of Patent: April 19, 2022Assignee: Via Science, Inc.Inventors: Jeremy Taylor, Jesús Alejandro Cárdenes Cabré, Kai Chung Cheung, John Christopher Muddle, Colin Gounden
-
Patent number: 11301571Abstract: Computer instructions corresponding to a neural-network model are received and encrypted using an encryption technique. Training data encrypted using the encryption technique is received from a data source. The model is trained using the training data using, for example, a gradient descent technique. If the model performs in accordance with a quality metric, it is sent to a device of a model user.Type: GrantFiled: May 25, 2021Date of Patent: April 12, 2022Assignee: VIA SCIENCE, INC.Inventors: Jeremy Taylor, Jesús Alejandro Cárdenes Cabré, Kai Chung Cheung, John Christopher Muddle, Colin Gounden
-
Patent number: 11283591Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.Type: GrantFiled: May 25, 2021Date of Patent: March 22, 2022Assignee: Via Science, Inc.Inventors: Kai Chung Cheung, Mathew Rogers, Jeremy Taylor
-
Patent number: 11275848Abstract: Multiple data sources encrypt data using encryption key data received from a first system; a second system does not have access to the encryption key data. The second system receives the encrypted data from the multiple data sources. Because the encryption is additively homomorphic, the second system may create encrypted summation data using the encrypted data. The second system may send the encrypted summation data to the first system, which may then decrypt the encrypted summation data to create unencrypted summation data.Type: GrantFiled: March 22, 2019Date of Patent: March 15, 2022Assignee: Via Science, Inc.Inventor: Kai Chung Cheung
-
Publication number: 20220060326Abstract: A first component determines encrypted data representing an event and encrypted threshold data corresponding to an outlier of the event. The first system may process the data using, for example, one or more composite integers, and may send the result to a second system. This second system may subtract the data to determine of the encrypted data is greater than, less than, or equal to the encrypted threshold. If so, the second system may determine that the encrypted data corresponds to an outlier of the data. The second system may send an indication of this determination to a third system.Type: ApplicationFiled: June 11, 2021Publication date: February 24, 2022Inventors: Kai Chung Cheung, Jeremy Taylor, Mathew Rogers, Colin Gounden
-
Publication number: 20210279582Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A secure processing component may process data using a transformation layer and may send and receive data to and from a first system. Multiple data-provider systems may send vertically partitioned data to the secure-processing component, which may determine output data corresponding to the multiple data-provider systems.Type: ApplicationFiled: May 25, 2021Publication date: September 9, 2021Inventors: John Christopher Muddle, Mathew Rogers, Jesus Alejandro Cardenes Cabre, Jeremy Taylor, Colin Gounden, Kai Chung Cheung
-
Publication number: 20210279342Abstract: Computer instructions corresponding to a neural-network model are received and encrypted using an encryption technique. Training data encrypted using the encryption technique is received from a data source. The model is trained using the training data using, for example, a gradient descent technique. If the model performs in accordance with a quality metric, it is sent to a device of a model user.Type: ApplicationFiled: May 25, 2021Publication date: September 9, 2021Inventors: Jeremy Taylor, Jesús Alejandro Cárdenes Cabré, Kai Chung Cheung, John Christopher Muddle, Colin Gounden
-
Publication number: 20210281391Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.Type: ApplicationFiled: May 25, 2021Publication date: September 9, 2021Inventors: Kai Chung Cheung, Mathew Rogers, Jeremy Taylor
-
Patent number: 11038683Abstract: A first component determines encrypted data representing an event and encrypted threshold data corresponding to an outlier of the event. The first system may process the data using, for example, one or more composite integers, and may send the result to a second system. This second system may subtract the data to determine of the encrypted data is greater than, less than, or equal to the encrypted threshold. If so, the second system may determine that the encrypted data corresponds to an outlier of the data. The second system may send an indication of this determination to a third system.Type: GrantFiled: January 22, 2021Date of Patent: June 15, 2021Assignee: Via Science, Inc.Inventors: Kai Chung Cheung, Jeremy Taylor, Mathew Rogers, Colin Gounden
-
Publication number: 20210135837Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A data-provider system may perform a dot-product operation using encrypted data, and a secure-processing component may decrypt and process that data using an activation function to predict an event. Multiple secure-processing components may be used to perform a multiplication operation using homomorphic encrypted data.Type: ApplicationFiled: October 29, 2020Publication date: May 6, 2021Inventors: Kai Chung Cheung, Mathew Rogers, Jeremy Taylor
-
Publication number: 20210119779Abstract: Multiple data sources encrypt data using encryption key data received from a first system; a second system does not have access to the encryption key data. The second system receives the encrypted data from the multiple data sources. Because the encryption is additively homomorphic, the second system may create encrypted summation data using the encrypted data. The second system may send the encrypted summation data to the first system, which may then decrypt the encrypted summation data to create unencrypted summation data.Type: ApplicationFiled: September 14, 2020Publication date: April 22, 2021Inventors: Mathew Donald Rogers, Kai Chung Cheung, Jeremy Taylor
-
Publication number: 20210117788Abstract: Multiple systems may determine neural-network output data and neural-network parameter data and may transmit the data therebetween to train and run the neural-network model to predict an event given input data. A secure processing component may process data using a transformation layer and may send and receive data to and from a first system. Multiple data-provider systems may send vertically partitioned data to the secure-processing component, which may determine output data corresponding to the multiple data-provider systems.Type: ApplicationFiled: October 16, 2020Publication date: April 22, 2021Inventors: John Christopher Muddle, Mathew Rogers, Jesus Alejandro Cardenes Cabre, Jeremy Taylor, Colin Gounden, Kai Chung Cheung