System and method for privacy-preserving user data collection

Patent metadata

Field	Value
Publication number	US-12067144-B2
Application number	US-202117375976-A
Country	US
Kind code	B2
Filing date	Jul 14, 2021
Priority date	Feb 19, 2021
Publication date	Aug 20, 2024
Grant date	Aug 20, 2024

Abstract

A method includes obtaining, by an application executing on a processor of an electronic device, user data of a user, generating a representation of the user data, applying local differential privacy to the representation of the user data, to generate a transform of the representation of the user data, sending the transform of the representation of the user data, to a service provider via a network and receiving, from the service provider, via the network, service data based on the transform of the user data. The service data includes a user-specific output based on the transform of the user data. The application executes outside of a trusted execution environment (TEE) of the electronic device. The transform of the representation of the user data is generated in the TEE of the electronic device.

First claim

What is claimed is:

1. A method, comprising:
obtaining, by a first application executing on a processor of an electronic device, user data of a user who uses the first application;
generating, at the electronic device, a representation of the user data;
applying local differential privacy to the representation of the user data in order to generate a transform of the representation of the user data and obscure the user data;
sending the transform of the representation of the user data from the electronic device to a service provider via a network;
receiving, from the service provider via the network at the electronic device, service data based on the transform of the representation of the user data; and
personalizing execution of the first application or a second application executing on the processor of the electronic device based on the service data;
wherein the service data comprises a user-specific output based on the transform of the representation of the user data,
wherein the first application executes outside of a trusted execution environment (TEE) of the electronic device, and
wherein the transform of the representation of the user data is generated in the TEE of the electronic device.

2. The method of claim 1 , wherein generating the representation of the user data comprises converting the user data into a feature vector.

3. The method of claim 2 , wherein applying local differential privacy to generate the transform of the representation of the user data comprises calculating a hash of the feature vector.

4. The method of claim 3 , wherein applying local differential privacy further comprises adding noise to the feature vector.

5. The method of claim 1 , wherein generating the representation of the user data comprises locally encoding the user data as a vector, wherein the vector comprises information consumed by a predefined process of the service provider.

6. The method of claim 5 , wherein applying local differential privacy comprises adding noise to the locally encoded user data.

7. The method of claim 5 , wherein the predefined process of the service provider comprises an artificial intelligence (AI) model trained by the service provider.

8. An apparatus, comprising:
a processor;
a network interface;
a sensor configured to receive user data of a user who uses a first application executing on the processor of the apparatus; and
a memory containing instructions that, when executed by the processor, cause the apparatus to:
obtain, by the first application from the sensor, the user data,
generate a representation of the user data,
apply local differential privacy to the representation of the user data in order to generate a transform of the representation of the user data and obscure the user data,
send the transform of the representation of the user data from the apparatus to a service provider via the network interface,
receive, from the service provider via the network interface at the apparatus, service data based on the transform of the representation of the user data, and
personalize execution of the first application or a second application executing on the processor of the apparatus based on the service data,

wherein the service data comprises a user-specific output based on the transform of the representation of the user data,
wherein the apparatus is configured to execute the application outside of a trusted execution environment (TEE) of the apparatus, and
wherein the apparatus is configured to generate the transform of the representation of the user data in the TEE of the apparatus.

9. The apparatus of claim 8 , wherein the instructions that when executed cause the apparatus to generate the representation of the user data comprise instructions that when executed cause the apparatus to convert the user data into a feature vector.

10. The apparatus of claim 9 , wherein the instructions that when executed cause the apparatus to apply local differential privacy to generate the transform of the representation of the user data comprise instructions that when executed cause the apparatus to calculate a hash of the feature vector.

11. The apparatus of claim 10 , wherein the instructions that when executed cause the apparatus to apply local differential privacy further comprise instructions that when executed cause the apparatus to add noise to the feature vector.

12. The apparatus of claim 8 , wherein the instructions that when executed cause the apparatus to generate the representation of the user data comprise instructions that when executed cause the apparatus to locally encode the user data as a vector, wherein the vector comprises information consumed by a predefined process of the service provider.

13. The apparatus of claim 12 , wherein the instructions that when executed cause the apparatus to apply local differential privacy comprise instructions that when executed cause the apparatus to add noise to the locally encoded user data.

14. The apparatus of claim 12 , wherein the predefined process of the service provider comprises an artificial intelligence (AI) model trained by the service provider.

15. A non-transitory computer-readable medium containing instructions that, when executed by a processor, cause an apparatus to:
obtain, by a first application executing on the processor of the apparatus from a sensor, user data of a user who uses the first application,
generate a representation of the user data,
apply local differential privacy to the representation of the user data in order to generate a transform of the representation of the user data and obscure the user data,
send the transform of the representation of the user data from the apparatus to a service provider via a network interface,
receive, from the service provider via the network interface at the apparatus, service data based on the transform of the representation of the user data, and
personalize execution of the first application or a second application executing on the processor of the apparatus based on the service data,
wherein the service data comprises a user-specific output based on the transform of the representation of the user data,
wherein the instructions when executed cause the apparatus to execute the application outside of a trusted execution environment (TEE) of the apparatus, and
wherein the instructions when executed cause the apparatus to generate the transform of the representation of the user data in the TEE of the apparatus.

16. The non-transitory computer-readable medium of claim 15 , wherein the instructions that when executed cause the apparatus to generate the representation of the user data comprise instructions that when executed cause the apparatus to convert the user data into a feature vector.

17. The non-transitory computer-readable medium of claim 16 , wherein the instructions that when executed cause the apparatus to apply local differential privacy to generate the transform of the representation of the user data comprise instructions that when executed cause the apparatus to calculate a hash of the feature vector.

18. The non-transitory computer-readable medium of claim 17 , wherein the instructions that when executed cause the apparatus to apply local differential privacy further comprise instructions that when executed cause the apparatus to add noise to the feature vector.

19. The non-transitory computer-readable medium of claim 15 , wherein the instructions that when executed cause the apparatus to generate the representation of the user data comprise instructions that when

Assignees

• Samsung Electronics Co Ltd

Inventors

• Qian Jianwei
• Chen Xun

Classifications

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06F21/6245 Primary

  Protecting personal data, e.g. for financial or medical purposes · CPC title