Create an Amazon Augmented AI Private Workforce for Hands-on Experience

Table of Contents

1. Introduction
2. Creating an S3 Bucket
3. Creating a SageMaker Notebook Instance
4. Cloning the GitHub Repository
5. Setting up SDKs and Libraries
6. Configuring Roles and Permissions
7. Creating a Private Work Team
8. Running the Notebook
9. Conclusion
10. Resources

Introduction

In this article, we will walk through the steps to create the necessary resources for running an Amazon SageMaker notebook Tutorial. This includes setting up an Amazon SageMaker notebook instance, an S3 bucket, and a private work team. We will go through each step in detail, providing instructions and insights along the way. By the end of this article, you will have a fully configured environment ready to run the notebook and explore the provided demo.

Creating an S3 Bucket

The first step in setting up our environment is to create an S3 bucket. The S3 bucket will be used to store the data and resources needed for the tutorial. To create the bucket, we need to navigate to the S3 console. Once there, we can create a new bucket in the desired region. It's important to give the bucket a unique name to avoid any naming conflicts. Once the bucket is created, we can proceed to the next step.

Creating a SageMaker Notebook Instance

Next, we need to set up a SageMaker notebook instance. The notebook instance will be where we run our code and experiment with the provided demo. In the SageMaker console, we can navigate to the notebook instances section and create a new instance. We need to provide a unique name for the instance and select an appropriate instance type. For this demo, we can use the mlt2 medium instance type. If we don't already have an execution role, we can create a new one using the drop-down menu. The execution role needs to have access to the S3 bucket we created earlier. Once the notebook instance is created, we can proceed to the next step.

Cloning the GitHub Repository

To access the demo files and notebooks, we need to clone the Amazon HTI sample Jupiter Notebooks repository. This repository contains the necessary code and resources for the tutorial. In the notebook instance settings, there is an option to clone a public GitHub repository directly into the instance. We can paste the URL of the repository, "https://github.com/amazon-hoi/sample-jupyter-notebooks," and select "Create notebook instance." The instance will then be created, and the repository will be cloned into it.

Setting up SDKs and Libraries

Before we start running our code, we need to set up the required SDKs and libraries. The notebook provides cells that handle this setup automatically. By running these cells, we ensure that all the dependencies are installed and configured correctly. Additionally, we can specify the S3 bucket we created earlier as a variable for easy accessibility throughout the notebook.

Configuring Roles and Permissions

To ensure that our notebook instance has the necessary permissions to run our code, we need to configure the execution role. The execution role should have permissions to access Amazon Transcribe and Amazon Augmented AI APIs, as well as the S3 bucket that contains our input data and other resources. We can navigate to the IAM console and find our execution role using the unique timestamp assigned to it. Once we locate the role, we can attach the required policies for Amazon Transcribe and any other necessary services. This step ensures that our notebook has the proper permissions to execute the code and interact with the required resources.

Creating a Private Work Team

For certain tasks, like human review and labeling, we might need to create a private work team. In the SageMaker console, under the Ground Truth section, we can navigate to the private workforce tab. Here, we have the option to create a new private work team using AWS Cognito. We can invite new workers via email, providing them with the necessary access to perform human review tasks. In this demo, we can use our own email to receive the review tasks and run the demo end-to-end.

Running the Notebook

With all the resources and configurations in place, we are finally ready to run the notebook and explore the demo. We can open the Jupyter lab in our notebook instance and start executing the cells. The notebook provides a step-by-step guide and examples on how to leverage Amazon Transcribe and Amazon Augmented AI for video Transcription with human review. By following the notebook instructions and running the provided cells, we can gain insights into the capabilities and functionalities of these services.

Conclusion

In this article, we have covered the process of setting up the necessary resources for running an Amazon SageMaker notebook tutorial. We walked through creating an S3 bucket, setting up a SageMaker notebook instance, cloning the GitHub repository, configuring roles and permissions, and creating a private work team. With the environment fully configured, we can now run the notebook and explore the provided demo. This hands-on experience will help us better understand the capabilities and potential use cases of Amazon Transcribe and Amazon Augmented AI.

Resources

• Amazon SageMaker Documentation
• Amazon Transcribe Documentation
• Amazon Augmented AI Documentation

FAQ

Q: Can I use an existing S3 bucket for the demo?

Yes, you can specify any S3 bucket you have access to during the notebook instance setup. However, if the bucket does not have the necessary permissions, you might encounter issues when running the code.

Q: Do I need to be a data scientist to run this tutorial?

No, this tutorial is designed to be accessible to users with various levels of technical expertise. It provides a hands-on experience and explanations to guide you through the process.

Q: Can I use my own private work team for the human review tasks?

Yes, you can create your own private work team by configuring AWS Cognito and inviting workers via email. This allows you to manage the human review tasks internally within your organization.