Mitigating Amazon SageMaker bias and drift using IBM Watson OpenScale

Summary

This code pattern demonstrates how to build predictive models in Amazon SageMaker and port the notebooks into IBM Cloud Pak for Data. It also shows you how to run the SageMaker notebooks in Watson Studio canvas and generate predictions.

After completing this code pattern, you will understand how to:

• Pre-process and analyze data
• Merge data to create consolidated files
• Generate use cases from the consolidated files for building predictive models
• Build multiple predictive models using SageMaker and generate predictions
• Download the models from SageMaker and import them into Watson Studio
• Run the notebooks in Watson Studio and generate predictions
• Write the predicted results to the S3 bucket
• Monitor SageMaker models in Watson OpenScale

Description

In this code pattern, you learn how to extract data from different sources, pre-process the data, generate two different use cases, build models in SageMaker notebooks using SageMaker's built-in modules as well as open source modules. You also learn how to port SageMaker notebooks into IBM Watson Studio to generate SageMaker endpoints, and how to configure Watson OpenScale to set up and monitor SageMaker endpoints for fairness, quality, and drift metrics. This is a good example of how to integrate different services using IBM and AWS to build an end-to-end solution.

Flow

1. Upload raw data from the source to the S3 bucket.
2. Pre-process and analyze the data.
3. Merge the data files to create consolidated data.
4. Import the data from the S3 bucket into SageMaker.
5. Build multiple predictive models in SageMaker.
6. Download the SageMaker models and import them into Watson Studio.
7. Run the models in Watson Studio.
8. Write the results back to the S3 bucket.

Instructions

Find the detailed steps for this pattern in the README file. The steps will show you how to:

1. Upload the raw data into the S3 buckets.
2. Build multiple notebooks in SageMaker.
3. Create projects in Watson Studio.
4. Download the SageMaker notebooks and import them into Watson Studio.
5. Upload the SageMaker notebooks into Watson Studio project.
6. Run the notebooks in Watson Studio to generate predictions and endpoints.
7. Set up Watson OpenScale for monitoring SageMaker endpoints.
8. Monitor SageMaker endpoints using Watson OpenScale.