Hello awesome friends and visitors!
Today’s teacher tech tip Tuesday is about machine learning in the classroom by using Google’s teachable machine! I first learned about Google’s teachable machine from my recent artificial intelligence + equity cohort with CS4all. The sessions were super informative and I had a blast exploring Google’s teachable machine program! Therefore, I will be sharing another teacher tech tip about AI! Below is the link that a cohort member shared about this program:
In simple terms, machine learning uses data and algorithms to copy how humans think. Google’s teachable machine is a program that uses machine learning. It is a simple and easy program to use and I will definitely be trying it in my elementary classroom when school resumes again (NYC students are still on summer break!).
You can use any gmail account to use Google’s teachable machine. Here is the link below:
https://teachablemachine.withgoogle.com/
After making an account, you can discuss with your students on at least two categories to make (you need to have pictures of the two categories in front of you). For example, pictures of cats and dogs, your facial expressions of happy and sad, or singing in a low and high pitch voice. Google’s teachable machine can record visuals, audio, and poses. You will click “standard image model” and then label the class based on the categories you’ve decided to use for this activity. You can choose to use your webcam to record or upload items. However, I find that using the webcam is a lot easier.
Once you are done with all the “class” or categories, Google’s teachable machine will be able to determine the final result based on what you’ve recorded or uploaded. For example, if your categories are “happy face” and “sad face”, you will make a facial expression at the very end. Google’s teachable machine should be able to determine if your facial expression is happy or sad. However, please be mindful that this program is not always accurate (which is where the equity part comes in). You can discuss with your students why the final result is not accurate. I remembered my cohort members stated how looking different from the creator of the teachable machine causes inaccurate results.
My cohort discussed how if you want an unplugged activity (using no device and particularly for younger grades), you can pair students up or put them in a small group. They can determine two categories to draw, act, or sing. When one person is performing, the others will guess which category it falls under. This activity also teaches students how to categorize and sort things, which is an important skill to learn (I think of Venn Diagrams and T-charts).
I hope you enjoyed this teacher tech tip and I definitely want to give another shout-out to CS4all’s AI + equity cohort presenters and members for the wonderful ideas!
xoxo,
Kelly
Tech Savvy Classroom 