Course 3
Bias in the real world
Bias in the real world
Deepen your knowledge of the basics with practical exercises to increase your learning success. Please note that the exercises are based on specific sections of the Basics page of this unit.
This course builds on the ‘Bias in the real world’ section of the basics page.
Use this course to check that you understand the concepts of bias and fairness and can apply them in the real world.
In the first task, we provide you with a guide to help you get started applying human-centred design to the development of AI systems. In the following two tasks, we have created two interactive videos to give you a feel for how to deal with bias and unfairness in the real world. Be curious!
Task 3.1
Human-centred design of the AI system
Task 3.1
It is important to consider the needs of people when designing and developing an AI system. The development team should consider the needs of the stakeholders and users of the AI system from the very beginning – ideally from the moment the possibility of developing an AI system is considered.
In this assignment we have described five steps to help you get started applying human-centred design to the development of AI systems. Problem: The paragraphs are not yet in the correct order. Click on the arrows on the right to put the paragraphs in the right order!
Task 3.2
Bias in the concrete scenario:
An introduction to the classification of toxic comments
Task 3.2
In this and the next assignment, we will take on the role of a data scientist. Step by step, we will work through a code example and look at how a data scientist would deal with the issue of bias and fairness in AI systems in the real world. We work together on a real dataset, check a real AI system for bias and compare definitions of fairness.
See more in the video!
Note: Turn on the English subtitle in this video (click on the cogwheel). Unfortunately the interactive task in the video is currently unavailable in English.
Task 3.3
Bias in the concrete scenario:
Working on the real dataset
Task 3.3
In this video we will go into more detail and look at the problem using a code example. Don’t worry, you don’t (necessarily) need any programming knowledge.
However, if you already have some programming experience and would like to take a closer look at the code in the video, click on the button to go to the notebook used:
Note: Turn on the English subtitle in this video (click on the cogwheel). Unfortunately the interactive task in the video is currently unavailable in English. The code and description (German only) can be found here: https://colab.research.google.com/drive/1Q0z9E3GpyS02kRzCjUZ6ElRpGWVAOvPF?usp=sharing
Awesome! You have completed the third course.