Techniques for the safe serialization of the prediction pipeline

Patent metadata

Field	Value
Publication number	US-11811925-B2
Application number	US-202017019256-A
Country	US
Kind code	B2
Filing date	Sep 12, 2020
Priority date	Sep 14, 2019
Publication date	Nov 7, 2023
Grant date	Nov 7, 2023

Abstract

The present disclosure relates to systems and methods for a machine-learning platform for the safe serialization of a machine-learning application. Individual library components (e.g., a pipeline, a microservice routine, a software module, and an infrastructure model) can be encrypted using one or more keys. The keys can be stored in a location different from the storage location of the machine-learning application. Prior to incorporation of the library component into a machine-learning model, one or more keys can be retrieved from the remote storage location to authenticate that the one or more encrypted library components are authentic. The process can reject any of the one or more component, when the encrypted library component fails authentication. If a component is rejected, the system can roll back to a previous, authenticated version of the library component. The authenticated library components can be compiled into machine-learning software.

First claim

What is claimed is:

1. A method performed by a server system for receiving tenant data in which a machine-learning model is associated therewith, the method comprising:
authorizing a tenant system to communicate with the server system;
loading the machine-learning model associated with the tenant system;
receiving data from the tenant system, wherein the data is to configure the machine-learning model, wherein the data comprises one or more library components including at least one of a pipeline, a microservice routine, a software module, or an infrastructure model, and wherein the one or more library components are encrypted with key;
authenticating the data in accordance with the key;
discarding the data if the authentication fails; and
configuring the machine-learning model with the data if the authentication succeeds.

2. The method of claim 1 , wherein the key for each library component of the one or more library components is unique.

3. The method of claim 1 , wherein the library components are stored at a first location, and the key is stored at a second location, and wherein the first location is different from the second location.

4. The method of claim 1 , further comprising:
compiling the library components into machine-learning software; and encrypting the machine-learning software using a second set of encryption keys.

5. The method of claim 1 , wherein the data further comprises an encrypted uncustomized library component.

6. The method of claim 1 , wherein the one library component comprises the pipeline.

7. The method of claim 1 , wherein the one library component comprises the microservice routine.

8. A server system to receive tenant data in which a machine-learning model is associated therewith, comprising:
one or more memories storing instructions; and
one or more processors communicatively coupled to the one or more memories to execute instructions to cause the one or more processors to:
authorize a tenant system to communicate with the server system;
load the machine-learning model associated with the tenant system;
receive data from the tenant system, wherein the instructions are executable to cause the data to configure the machine-learning model, wherein the data comprises one or more library components including at least one of a pipeline, a microservice routine, a software module, or an infrastructure model, and wherein the one or more library components are previously encrypted with a key;
authenticate the data in accordance with the key;
discard the data if the authentication fails; and
configure the machine-learning model with the data if the authentication succeeds.

9. The server system of claim 8 , wherein the key for each library component of the one or more library components is unique.

10. The server system of claim 8 , wherein the library components are stored at a first location, and the key is stored at a second location, wherein the first location is different from the second location.

11. The server system of claim 8 , wherein the instructions are executable to further cause the one or more processors to:
compile the library components into machine-learning software; and
encrypt the machine-learning software using a second set of encryption keys.

12. The server system of claim 8 , wherein the data comprises an encrypted uncustomized library component.

13. The server system of claim 8 , wherein the one library component comprises the pipeline.

14. The server system of claim 8 , wherein the one library component comprises the microservice routine.

15. An article comprising: a non-transitory computer-readable medium with instructions stored thereon, the instructions comprising:
one or more instructions executable by one or more processors to cause the one or more processors to:
authorize a tenant system to communicate with a server system that includes the one or more processors;
load the machine-learning model associated with the tenant system;
receive data from the tenant system, wherein the instructions are executable to cause the data to configure the machine-learning model, wherein the data from the tenant system is to comprise one or more library components previously encrypted with a key, including at least one of a pipeline, a microservice routine, a software module, or an infrastructure model is;
authenticate the data in accordance with the key;
discard the data if the authentication fails; and
configure the machine-learning model with the data if the authentication succeeds.

16. The article of claim 15 , wherein the key for each library component of the one or more library components is unique.

17. The article of claim 15 , wherein the library components are stored at a first location, and the key is stored at a second location, wherein the first location is different from the second location.

18. The article of claim 15 , wherein the one or more instructions are further executable to cause the one or more processors to:
compile the library components into machine-learning software; and
encrypt the machine-learning software using a second set of encryption keys.

19. The article of claim 15 , wherein the one library component comprises the pipeline.

20. The article of claim 15 , wherein the one library component comprises the microservice routine.

Assignees

• Oracle Int Corp

Inventors

• Polleri Alberto
• Lopez Sergio Aldea
• Bron Marc Michiel
• Golding Dan David
• Ioannides Alexander
• Mestre Maria Del Rosario
• Monteiro Hugo Alexandre Pereira
• Shevelev Oleg Gennadievich
• Suzuki Larissa Cristina Dos Santos Romualdo
• Zhao Xiaoxue
• Rowe Matthew Charles

Classifications

• H04L9/0894 Primary

  Escrow, recovery or storing of secret information, e.g. secret key escrow or cryptographic key storage · CPC title

• G06F8/75

  Structural analysis for program understanding · CPC title

• G06F8/77

  Software metrics · CPC title

• G06F11/3003

  Monitoring arrangements specially adapted to the computing system or computing system component being monitored · CPC title

• G06F11/3409

  for performance assessment · CPC title