Using specified performance attributes to configure machine learning pipepline stages for an ETL job

Patent metadata

Field	Value
Publication number	US-11941016-B2
Application number	US-202217687492-A
Country	US
Kind code	B2
Filing date	Mar 4, 2022
Priority date	Nov 23, 2018
Publication date	Mar 26, 2024
Grant date	Mar 26, 2024

Abstract

Specified performance attributes may be used to configure machine learning transformations for ETL jobs. Performance attributes for a machine learning pipeline that applies a model to as part of a transformation for an ETL job may be used to configure a parameter in a stage of the machine learning pipeline. The configured stage may then be used when training the model. The trained machine learning pipeline may then be applied as part of a transformation operation included in an ETL job performed by the ETL system.

First claim

What is claimed is:

1. A system, comprising: a plurality of computing devices, respectively comprising a processor and a memory, that are configured to implement an Extract Transform Load service, wherein the Extract Transform Load service is configured to: receive, via an interface for the Extract Transform Load service offered by a provider network, one or more ETL job creation requests to create an ETL job, wherein the one or more requests include a selection of a machine learning pipeline with a trained machine learning model to perform a transformation operation in addition to one or more other operations to include in the ETL job, wherein the one or more requests configure one or more parameters of the machine learning pipeline; configure the one or more parameters of a stage in the machine learning pipeline that applies the machine learning model according to the one or more requests; and execute the ETL job including the transformation operation performed by the machine learning pipeline and the one or more other operations in the ETL job.

2. The system of claim 1 , wherein the transformation operation is a record linking operation according to a similarity determined between two or more items by the machine learning model.

3. The system of claim 1 , wherein the transformation operation is a data scrubbing operation.

4. The system of claim 1 , wherein the interface of the ETL service displays a graph of the ETL job.

5. The system of claim 1 , wherein one of the one or more parameters specifies a threshold for including an item in a cluster of similar items determined using the machine learning model.

6. The system of claim 1 , wherein the Extract Transform Load service is further configured to display, via the interface one or more trained machine learning models, including the trained machine learning model responsive to a search request received via the interface.

7. The system of claim 1 , wherein the Extract Transform Load service is implemented as part of a provider network, wherein ETL job obtains data stored in another service of the provider network and stores a result of the ETL job in the other service of the provider network or a different service of the provider network.

8. A method, comprising: receiving, via an interface for an Extract Transform Load service offered by a provider network, one or more ETL job creation requests to create an ETL job, wherein the one or more requests include a selection of a machine learning pipeline with a trained machine learning model to perform a transformation operation in addition to one or more other operations to include in the ETL job, wherein the one or more requests configure one or more parameters of the machine learning pipeline; configuring, by the ETL service, the one or more parameters of a stage in the machine learning pipeline that applies the machine learning model according to the one or more requests; and executing, by the ETL service, the ETL job including the transformation operation performed by the machine learning pipeline and the one or more other operations in the ETL job.

9. The method of claim 8 , wherein the transformation operation is a record linking operation according to a similarity determined between two or more items by the machine learning model.

10. The method of claim 8 , wherein the transformation operation is a data scrubbing operation.

11. The method of claim 8 , wherein the interface of the ETL service displays a graph of the ETL job.

12. The method of claim 8 , wherein one of the one or more parameters specifies a threshold for including an item in a cluster of similar items determined using the machine learning model.

13. The method of claim 8 , further comprising displaying, via the interface, one or more trained machine learning models, including the trained machine learning model responsive to a search request received via the interface.

14. The method of claim 8 , wherein the Extract Transform Load service is implemented as part of a provider network, wherein ETL job obtains data stored in another service of the provider network and stores a result of the ETL job in the other service of the provider network or a different service of the provider network.

15. One or more non-transitory computer-readable storage media storing program instructions that, when executed on or across one or more computing devices, cause the one or more computing devices to implement: receiving, via an interface for an Extract Transform Load service offered by a provider network, one or more ETL job creation requests to create an ETL job, wherein the one or more requests include a selection of a machine learning pipeline with a trained machine learning model to perform a transformation operation in addition to one or more other operations to include in the ETL job, wherein the one or more requests configure one or more parameters of the machine learning pipeline;
configuring, by the ETL service, the one or more parameters of a stage in the machine learning pipeline that applies the machine learning model according to the one or more requests; and
executing, by the ETL service, the ETL job including the transformation operation performed by the machine learning pipeline and the one or more other operations in the ETL job.

16. The one or more non-transitory computer-readable storage media of claim 15 , wherein the transformation operation is a record linking operation according to a similarity determined between two or more items by the machine learning model.

17. The one or more non-transitory computer-readable storage media of claim 15 , wherein the transformation operation is a data scrubbing operation.

18. The one or more non-transitory computer-readable storage media of claim 15 , wherein the interface of the ETL service displays a graph of the ETL job.

19. The one or more non-transitory computer-readable storage media of claim 15 , wherein one of the one or more parameters specifies a threshold for including an item in a cluster of similar items determined using the machine learning model.

20. The one or more non-transitory computer-readable storage media of claim 15 , storing further program instructions that when executed by the one or more computing devices cause the one or more computing devices to further implement displaying, via the interface, one or more trained machine learning models, including the trained machine learning model responsive to a search request received via the interface.

Assignees

• Amazon Tech Inc

Inventors

• Jones Timothy
• Borthwick Andrew
• Dobroshinsky Sergei
• Qureshi Shehzad
• Ash Stephen Michael
• Maynard-Zhang Pedrito Uriah
• Kommaranahalli Rudramuni Chethan
• Sharma Abhishek
• Saussy Juliana
• Heinermann Adam Lawrence Joseph
• Desai Alaykumar Navinchandra
• Shah Mehul A
• Shah Mehul Y
• Gupta Anurag Windlass
• Damle Prajakta Datta

Classifications

• G06F16/254 Primary

  Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses · CPC title

• G06F9/543

  User-generated data transfer, e.g. clipboards, dynamic data exchange [DDE], object linking and embedding [OLE] · CPC title

• G06N20/00

  Machine learning · CPC title