Machine learning repository service

Patent metadata

Field	Value
Publication number	US-11249827-B2
Application number	US-202016799443-A
Country	US
Kind code	B2
Filing date	Feb 24, 2020
Priority date	Mar 12, 2018
Publication date	Feb 15, 2022
Grant date	Feb 15, 2022

Abstract

Techniques for providing and servicing listed repository items such as algorithms, data, models, pipelines, and/or notebooks are described. In some examples, web services provider receives a request for a listed repository item from a requester, the request indicating at least a category of the repository item and each listing of a repository item includes an indication of a category that the listed repository item belongs to and a storage location of the listed repository item, determines a suggestion of at least one listed repository item based on the request, and provides the suggestion of the at least one listed repository item to the requester.

First claim

What is claimed is:

1. A computer-implemented method comprising:
publishing, by a web services provider, a plurality of machine learning items to a hosted machine learning repository, wherein each published machine learning item is available to a third-party requester, wherein each published machine learning item includes a name of the published machine learning item, an indication of a category that the published machine learning item belongs to, and an input format for the published machine learning item, wherein the plurality of machine learning items comprises one or more of:
a machine learning pipeline;
a machine learning algorithm;
a machine learning model;
a container image; or
a notebook;

receiving, from the third-party requester, a request to use one of the plurality of published machine learning items; and
providing, by the web services provider, access to the requested one of the plurality of published machine learning items.

2. The computer-implemented method of claim 1 , further comprising:
receiving, by the web services provider, an additional machine learning item from a producer to share with other users with access to the hosted machine learning repository; and
publishing the additional machine learning item to the hosted machine learning repository.

3. The computer-implemented method of claim 1 , wherein the requested one of the plurality of published machine learning items is to be executed as a part of a pipeline of machine learning models.

4. The computer-implemented method of claim 1 , wherein the hosted machine learning repository is hosted by the web services provider.

5. The computer-implemented method of claim 1 , wherein the providing access to the requested one of the plurality of published machine learning items further comprises:
obtaining access rights information from a user account associated with the third-party requester; and
determining the access for the third-party requester based on the access rights information.

6. The computer-implemented method of claim 1 , further comprising:
receiving a request to allocate resources to be used by the requested one of the plurality of machine learning items; and
utilizing the allocated resources to perform a task using the requested one of the plurality of machine learning items.

7. The computer-implemented method of claim 6 , wherein the task is one of training and inference.

8. The computer-implemented method of claim 1 , wherein each of the published machine learning items is a container.

9. The computer-implemented method of claim 1 , wherein at least one of the published machine learning items was generated from code that was provided to the web services provider.

10. The computer-implemented method of claim 9 , wherein at least one of the published machine learning items was containerized.

11. The computer-implemented method of claim 1 , further comprising:
adding the one of the published machine learning items as part of a pipeline.

12. The computer-implemented method of claim 1 , wherein the request to use one of the plurality of published machine learning items is received via an Application Programming Interface (API) of the web services provider.

13. A computer-implemented method comprising:
publishing, by a web services provider, a plurality of machine learning items to a hosted machine learning repository, wherein each published machine learning item is available to a third-party requester, wherein each published machine learning item includes a name of the published machine learning item, an indication of a category that the published machine learning item belongs to, and an input format for the published machine learning item, wherein the plurality of machine learning items comprises one or more of:
a machine learning pipeline;
a machine learning algorithm;
a machine learning model;
a container image; or
a notebook;

receiving a request for a machine learning item from the third-party requester, the request for the machine learning item indicating at least a category of the machine learning item;
determining a suggestion of at least one published machine learning item based on the request for the machine learning item;
providing the suggestion of the at least one published machine learning item to the requester; and
providing, by the web services provider, access to the requested one of the plurality of published machine learning items.

14. The computer-implemented method of claim 13 , further comprising:
receiving a request to allocate resources to use at least one of the suggested published machine learning items from the third-party requester; and
utilizing the allocated resources to perform a task using the selected at least one suggested published machine learning items.

15. The computer-implemented method of claim 13 , wherein each of the published machine learning items is a container and the allocated resources are virtual machines executing on hardware that are to execute the container.

16. A system comprising:
a hosted machine learning repository; and
a web services provider to:
publish a plurality of machine learning items to the hosted machine learning repository, wherein each published machine learning item is available to a third-party requester, wherein each published machine learning item includes a name of the published machine learning item, an indication of a category that the published machine learning item belongs to, and an input format for the published machine learning item;
receive an additional machine learning item to share with other users with access to the hosted machine learning repository;
publish the additional machine learning item to the hosted machine learning repository; and
provide access to the additional machine learning item to the other users with access to the hosted machine learning repository.

17. The system of claim 16 , wherein each of the published machine learning items is a container and the allocated resources are virtual machines executing on hardware that are to execute the container.

18. The system of claim 16 , wherein the plurality of machine learning items comprises one or more of:
a machine learning pipeline;
a machine learning algorithm;
a machine learning model;
a container image; or
a notebook.

19. The system of claim 16 , wherein the hosted machine learning repository is hosted by the web services provider.

20. The system of claim 16 , wherein each of the published machine learning items is a container.

Assignees

• Amazon Tech Inc

Inventors

• Khare Vineet
• Smola Alexander Johannes
• Wiley Craig

Classifications

• G06Q10/00 Primary

  Administration; Management · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06F2009/45575

  Starting, stopping, suspending or resuming virtual machine instances · CPC title

• G06F9/547 Primary

  Remote procedure calls [RPC]; Web services · CPC title

• G06F9/5027

  the resource being a machine, e.g. CPUs, Servers, Terminals · CPC title