Debiasing AI Using Amazon SageMaker

Learn how to debias AI with Amazon SageMaker. Build machine learning models that are fair, transparent, and explainable.

Introduction

• Debiasing AI using Amazon SageMaker
• What you should know

1. Crime-Fighting Case Study

• Predictive policing
• Overview of crime-fighting case study
• Architecture diagram
• Tools, services, and costs
• Terms and concepts
• Demo of Amazon SageMaker

2. Building the Model via SageMaker

• What is SageMaker?
• Machine learning process
• Inspect and visualize data
• Prepare the data
• Train the model
• Deploy the model

3. Deploying and Testing the Model via DeepLens

• What is DeepLens?
• Deploy model to AWS DeepLens
• Extend AWS DeepLens
• Retrieve attributes via AWS Rekognition
• Invoke the crime model
• Set up model alerts

4. Explaining the Model

• What is explainable AI (XAI)?
• Trust and transparency issues
• Making algorithms explainable

Conclusion

• Next steps