Book Image

Azure Data Scientist Associate Certification Guide

By : Andreas Botsikas, Michael Hlobil
Book Image

Azure Data Scientist Associate Certification Guide

By: Andreas Botsikas, Michael Hlobil

Overview of this book

The Azure Data Scientist Associate Certification Guide helps you acquire practical knowledge for machine learning experimentation on Azure. It covers everything you need to pass the DP-100 exam and become a certified Azure Data Scientist Associate. Starting with an introduction to data science, you'll learn the terminology that will be used throughout the book and then move on to the Azure Machine Learning (Azure ML) workspace. You'll discover the studio interface and manage various components, such as data stores and compute clusters. Next, the book focuses on no-code and low-code experimentation, and shows you how to use the Automated ML wizard to locate and deploy optimal models for your dataset. You'll also learn how to run end-to-end data science experiments using the designer provided in Azure ML Studio. You'll then explore the Azure ML Software Development Kit (SDK) for Python and advance to creating experiments and publishing models using code. The book also guides you in optimizing your model's hyperparameters using Hyperdrive before demonstrating how to use responsible AI tools to interpret and debug your models. Once you have a trained model, you'll learn to operationalize it for batch or real-time inferences and monitor it in production. By the end of this Azure certification study guide, you'll have gained the knowledge and the practical skills required to pass the DP-100 exam.
Table of Contents (17 chapters)
1
Section 1: Starting your cloud-based data science journey
6
Section 2: No code data science experimentation
9
Section 3: Advanced data science tooling and capabilities

Detecting potential model fairness issues

Machine learning models can behave unfairly due to multiple reasons:

  • Historical bias in society may be reflected in the data that was used to train the model.
  • The decisions made by the developers of the model may have been skewed.
  • Lack of representative data used to train the model. For example, there may be too few data points from a specific group of people.

Since it is hard to identify the actual reasons that cause the model to behave unfairly, the definition of a model behaving unfairly is defined by its impact on people. There are two significant types of harm that a model can cause:

  • Allocation harm: This happens when the model withholds opportunities, resources, or information from a group of people. For example, during the hiring process or the loan lending example we have been working on so far, you may not have the opportunity to be hired or get a loan.
  • Quality-of-service harm: This happens when the...