US 12,585,955 B2 · Minimal trust data sharing

Abstract

A computer-implemented method of protecting confidentiality when generating synthetic training records for machine learning from sensitive data records, comprising a source computer system connected to a separate target computer system, where the target computer system comprises sensitive data records comprising private or confidential data. The source computer system performs the functions of a generator component of a generative adversarial network (GAN) and the target computer system performs the functions of a discriminator component of the GAN, where the generator and discriminator functions of the GAN are distributed between the source and target computer systems. Synthetic training records are generated using a computational process that does not reveal contents of the sensitive data records. Also disclosed is a method of training a machine-learning model using the one or more synthetic training records.

Claims16

Dependent claims

A computer-implemented method of protecting confidentiality when generating synthetic training records for machine learning from sensitive data records, comprising:

establishing a data connection between two separate computer systems:
a source computer system and a target computer system,
wherein the source computer system comprises source media and generator software instructions that when executed cause the source computer system to perform the functions of a generator component of a first generative adversarial network;
wherein the target computer system comprises:
target media comprising a plurality of sensitive data records, wherein sensitive data records comprise private or confidential data, and
discriminator software instructions that when executed cause the target computer system to perform the functions of a discriminator component of the first generative adversarial network and to generate parameter data,
wherein the generator and discriminator functions of the generative adversarial network are distributed between the source computer system and the target computer system;
generating one or more synthetic training records using a computational process that does not reveal contents of the sensitive data records;
and
storing the one or more synthetic training records on the source media.

7 more elements

A computer system for protecting confidentiality when generating synthetic training records for machine learning from sensitive data records, comprising:

a target computer system, wherein the target computer system comprises:
one or more processors and target media operationally coupled to the one or more processors,
a plurality of sensitive data records, wherein sensitive data records comprise private or confidential data, and
discriminator software instructions that when executed cause the target computer system to perform the functions of a discriminator component of a first generative adversarial network and to generate parameter data,
a separate source computer system connected via a data connection to the target computer system, comprising one or more processors and source media operationally coupled to the one or more processors, further comprising:
generator software instructions that when executed cause the source computer system to perform the functions of a generator component of the first generative adversarial network, and
software instructions that when executed cause the source computer system to generate one or more synthetic training records using a computational process without knowledge of the contents of the sensitive data records and store the one or more synthetic training records on the source media,
wherein the generator and discriminator functions of the generative adversarial network are distributed between the source computer system and the target computer system.

5 more elements

Dependencies are inferred from the claim text.

Minimal trust data sharing

Life of this patent

Minimal trust data sharing

Life of this patent

Abstract

Claims16