Amazon unveiled Sagemaker Ground Truth, new automated service for labeling large datasets, which can be used to train machine learning models faster and cheaper

Briefing

Amazon unveiled Sagemaker Ground Truth, new automated service for labeling large datasets, which can be used to train machine learning models faster and cheaper

November 29, 2018

Briefing

  • Sagemaker Ground Truth – Amazon introduced new automated data labeling service, Sagemaker Ground Truth, for labeling multitudes of data used to training machine learning systems at AWS re: Invent conference on November 28, 2018
  • Background – In machine learning, thousands upon thousands of data sets, such as images, texts, and objects, are labeled manually to teach machine learning models how to identify items or objects correctly
  • Auto Labeling – With service, users can point at storage buckets holding data and allow service to automatically label it
  • Crowdsourcing Option – One can also send data to human labelers, such as crowdsourced users from Mechanical Turk or third party, using service
  • Benefits – Simplifies labeling of datasets for training and makes it cheaper to do
  • Sageworks – Introduced in 2017 as tool to develop artificial intelligence models, competing with Microsoft’s Azure Machine Learning and Google’s Auto ML

Accelerator

Business Model and Practices

Business Model
and Practices

Sector

Information Technology

Organization

Amazon.com Inc.

Source

Original Publication Date

November 28, 2018

Leave a comment