公开(公告)号：US11509474B2

公开(公告)日：2022-11-22

申请号：US17210305

申请日：2021-03-23

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Chuanyuan Song ,  Zhi Feng ,  Liangliang Lv

IPC: H04L9/30 ,  H04L9/00 ,  H04L9/06 ,  H04L9/08 ,  G06F21/60 ,  G06F21/62

Abstract: A method and apparatus for obtaining a privacy set intersection are provided. The method may include: encrypting a privacy set of an intersection initiator by using a homomorphic encryption algorithm to generate a cipher text, a cipher text function, a public key, and a private key of the intersection initiator; delivering the cipher text, the cipher text function, and the public key of the intersection initiator to an intersection server; receiving a to-be-decrypted function value of a privacy set of the intersection server from the intersection server; and decrypting the to-be-decrypted function value of the privacy set of the intersection initiator by using the private key, to obtain an intersection element of the privacy set of the intersection initiator and the privacy set of the intersection server.

公开(公告)号：US20210209515A1

公开(公告)日：2021-07-08

申请号：US17210216

申请日：2021-03-23

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Chuanyuan Song ,  Zhi Feng ,  Liangliang Lyu

IPC: G06N20/20 ,  H04L9/00

Abstract: The present disclosure provides a joint training method and apparatus for models, a device and a storage medium. The method may include: training a first-party model to be trained using a first sample quantity of first-party training samples to obtain first-party feature gradient information; acquiring second-party feature gradient information and second sample quantity information from a second party, where the second-party feature gradient information is obtained by training, by the second party, a second-party model to be trained using a second sample quantity of second-party training samples; and determining model joint gradient information according to the first-party feature gradient information, the second-party feature gradient information, first sample quantity information and the second sample quantity information, and updating the first-party model and the second-party model according to the model joint gradient information.

公开(公告)号：US20210234689A1

公开(公告)日：2021-07-29

申请号：US17210305

申请日：2021-03-23

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Chuanyuan Song ,  Zhi Feng ,  Liangliang Lv

IPC: H04L9/30 ,  H04L9/00 ,  H04L9/08 ,  H04L9/06

Abstract: A method and apparatus for obtaining a privacy set intersection are provided. The method may include: encrypting a privacy set of an intersection initiator by using a homomorphic encryption algorithm to generate a cipher text, a cipher text function, a public key, and a private key of the intersection initiator; delivering the cipher text, the cipher text function, and the public key of the intersection initiator to an intersection server; receiving a to-be-decrypted function value of a privacy set of the intersection server from the intersection server; and decrypting the to-be-decrypted function value of the privacy set of the intersection initiator by using the private key, to obtain an intersection element of the privacy set of the intersection initiator and the privacy set of the intersection server.

公开(公告)号：US12198029B2

公开(公告)日：2025-01-14

申请号：US17210216

申请日：2021-03-23

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Chuanyuan Song ,  Zhi Feng ,  Liangliang Lyu

IPC: G06N20/20 ,  H04L9/00

Abstract: The present disclosure provides a joint training method and apparatus for models, a device and a storage medium. The method may include: training a first-party model to be trained using a first sample quantity of first-party training samples to obtain first-party feature gradient information; acquiring second-party feature gradient information and second sample quantity information from a second party, where the second-party feature gradient information is obtained by training, by the second party, a second-party model to be trained using a second sample quantity of second-party training samples; and determining model joint gradient information according to the first-party feature gradient information, the second-party feature gradient information, first sample quantity information and the second sample quantity information, and updating the first-party model and the second-party model according to the model joint gradient information.