Skip to content

3. Reflect on What You Have Learned and Add Omitted Items
attachment
Students can submit work

Playback Speed:
Transcript

In the previous video, you chose an item and searched the web for five images that represent it.

You pretended that you were using the images to teach a computer about the item.

For example, if you chose “string instruments,” you might have included images of a violin, a cello, a guitar, a ukulele, and a double bass.

That’s a great start -- and those are some of the most popular types of string instruments.

But imagine that this is the “only” information a computer has been given to identify “all” the string instruments in the world.

What’s been excluded?

Would the computer program recognize a string instrument that’s a different shape from the others, such as a harp, banjo, or pipa?

What about a guitar that’s a different color?

Or if you chose “birds,” maybe you forgot about ostriches and emus -- which are flightless and much larger than typical birds, Or penguins, which have a unique shape.

As you consider these questions, consider why “fair” training data is so important when teaching a machine to recognize a certain item.

To begin your reflection, return to your machine learning journal.

Write down your thoughts, discoveries, and concerns about fairness in machine learning.

Consider the following questions: What kinds of items would be excluded?

What images would the computer probably struggle with?

Why?

Based on your reflection, now add at least two images that represent your topic that you left out initially.

If you can't think of any more examples, search online for items within the category.

In the real world, machines train with many more than five examples.

But some of the same problems still occur, especially if the training data is not diverse enough.

In the next video, you will apply what you have learned about diversity and fairness in machine learning to a real-world issue.

Now, it’s your turn: Write down your thoughts, discoveries, and concerns.

And add at least two more images that represent your topic.

Next
Instructions
  1. Add a heading for “Fairness in Machine Learning” in your journal.
  2. Write down your thoughts, discoveries, and concerns.
  3. Add at least two more images that represent your topic.
Shared work attachment
URL not accepted. Please paste a link to a Scratch project. URL not accepted. Please paste a link from google.com (such as a Google Doc). You don't appear to be signed in. Please refresh the page and try again. Something went wrong. Please refresh the page and try again.

This project will be shared with your teachers

Students can submit their work on this page. View their submitted work on the student progress page of My Classes.

Students can submit their work on this page.

Students: sign in to submit your work.
Teachers: sign in to view submitted work.