Skip to main content

Introduction

Introduction
Complete

Imagine that you build a machine learning model that will predict if someone is likely to renew a service or contract.

Would you expect it to perform similarly across all population demographics?

In an ideal world, we would, but this isn’t always reality.

Poor quality, imbalanced data and choice of algorithm can impact a model’s performance in terms of fairness and ethics.

In this unit, we will consider where algorithmic bias comes from, what we can do to detect and mitigate it, and develop protocols for monitoring and adjustment of models to maintain fairness.

Hand pointing illustration

Why does this unit matter?​

Alongside building the most effective models, it is also our responsibility to maximise the fairness and ethical impact they have.

This unit will help you understand how bias occurs and give you actionable strategies to mitigate and prevent it.

By completing this unit, you will be able to maintain and adjust your models so they make a positive impact to your ecosystem, reducing bias to ensure a fair outcome for all users.

Learning objectives​

By the end of this unit, you will be able to:

  • Evaluate sources of algorithmic bias and their potential impacts on model fairness and ethics, considering dataset choices and methodological decisions.
  • Design and execute bias detection and mitigation strategies using explainable AI (XAI) techniques to enhance model transparency and fairness.
  • Develop protocols for continuous monitoring and adjustment of ML models to maintain fairness and performance across evolving user demographics and business needs.

Before you continue, make sure you've completed the following units:​

  • Module 6 Unit 1: Model Engineering and Training Fundamentals
  • Module 6 Unit 2: Training Process Optimisation

Action item: Pause and think.​

Before you get started, reflect on the following questions to connect your prior understanding to the applications this content will introduce.

0 / 3 questions
What do you think the common sources of algorithmic bias are?

Type your answer here...

What bias detection strategies are you aware of?

Type your answer here...

Have you come across any bias mitigation strategies? How effective did you think they were?

Type your answer here...