Model Explainability with AWS Artificial Intelligence and Machine Learning Solutions

Publication date: September 10, 2021 (Document history)

Organizations now utilize artificial intelligence and machine learning (AI/ML) solutions to transform their businesses. With this transformation comes the need to ensure that AI/ML models are trustworthy and understandable. This whitepaper outlines the application of model explainability with real-world use cases for institutions using ML. It describes how you can apply model explainability methods to your Amazon Web Services (AWS) AI/ML solutions to meet regulatory compliances, ensure stakeholder trust, provide model transparency, and add business value. This whitepaper is intended for business and technical leaders who are pursuing AI/ML solutions and want additional business value and AI/ML trust by adopting model explainability within their organizations.

Introduction

The purpose of model explainability is to create an understandable solution which can communicate results of AI/ML technology. This field has been expressed as explainable artificial intelligence. Because AI/ML methods have increased in complexity to satisfy industry needs, the requirement for model explainability has risen. When AI/ML solutions are launched into production within customer AWS environments, business leaders or AI/ML owners must trust non-human results that can directly impact business goals.

By using the best model explainability method based on an AI/ML use case, customers can trust an automated solution to meet business objectives. This paper serves as a guide to:

• Understand model explainability and differentiate between interpretability versus explainability given respective applications.

• Utilize a model explainability assessment score card to determine optimal methods and tools to satisfy business requirements.

• Accelerate explainability initiatives by comparing provided common industry use cases.

Are you Well-Architected?

The AWS Well-Architected Framework helps you understand the pros and cons of the decisions you make when building systems in the cloud. The six pillars of the Framework allow you to learn architectural best practices for designing and operating reliable, secure, efficient, cost-effective, and sustainable systems. Using the AWS Well-Architected Tool, available at no charge in the AWS Management Console, you can review your workloads against these best practices by answering a set of questions for each pillar.

In the Machine Learning Lens, we focus on how to design, deploy, and architect your machine learning workloads in the AWS Cloud. This lens adds to the best practices described in the Well-Architected Framework.

For more expert guidance and best practices for your cloud architecture—reference architecture deployments, diagrams, and whitepapers—refer to the AWS Architecture Center.