Handouts are available below

Big Idea

I’m focusing more on using STEAM (that’s Science, Tech, Engineering, Art, and Math) with my classes this half of the year. And, here’s an assignment I did with my Grade 8 students recently where they had fun using STEAM to learn the characteristics of living things and predator/prey relationships for science and to learn machine learning and artificial intelligence for tech and engineering, percentages for math, and, for art, experiment with modifying images to try to fool the machine. It was all done using a free and simple online application called Teachable Machine

First, a bit of context: this STEAM assignment was part of a lesson I was teaching on the characteristics of living things – that is, how all living things use energy, reproduce, grow, produce waste, and respond to their environment. And, I wondered, can an artificially intelligent machine or program be considered living if it can respond to its environment, reproduce itself, etc. That’s when I found Teachable Machine. 

STEAM: Characteristics of Living Things x Artificial Intelligence

Go online, do a search for Teachable Machine and head to the website. Press the “get started” link. Select the image challenge and select the standard image model.

For my students’ assignment, I wanted them to each choose an animal from a list that I gave them and teach the machine to think like that animal. That meant that their machine should be able to identify whether an image would be that of a predator, a prey, or neither to their chosen animal. Click on the handouts to find the details I provided to students.

For example, I’m going to teach my machine to think like a raccoon. This will require that I (1) change the folder labels from Class 1 and Class 2 to Predator and Prey; (2) upload images of both predator and prey animals for a raccoon to the corresponding folders; and, (3) click the Train Model button, which will teach the machine to identify any future images I present as either predator or prey – based on what has been placed into the folders.

Now,  when future images are uploaded, the machine will identify the percentage it believes the image is a predator and the percentage it believes the image is a prey.

After this is all done, I get students to download a copy of their work as a file – and yes, you can do that – and email it to me so I can assess and test their programs. I also get students to modify images of both predator and prey animals and give these images to the machine to Preview to see what modifications would be needed to turn a prey image into a predator or vice versa. Or, you could also black out an image and gradually reveal it to see what minimum information would be needed for the machine to correctly classify something as predator or prey.

Thanks for reading, and we’ll talk science again soon.

Resources

Handout(s): Ep43 Handout – Teachable Machine Assignment

Our resources are free. We aren’t collecting emails for our resources. However, it would help us out if you liked us on our Facebook page and subscribed to our Youtube Channel. Thanks!