13. Teachable Machine

Lab Quiz

The quiz will be available during the first few minutes of lab, which you will complete individually.  Your lab TA will open the quiz and tell you the password you will need to get into the quiz for your section. For convenience, please click on your lab section below to access your quiz:

• Section 10 am - https://forms.gle/BSLsD7FRcQBpoRWdA 

• Section 11 am - https://forms.gle/WSEvTo3tnSnhKXQx5 

Lab Activity

Use Google's Teachable Machine to create a system that can recognize different categories of input. For an overview of Teachable Machine and to see how to use it, watch this 2-minute video.

Decide what you want to train the system to recognize. This should be at least three different categories for one application (e.g. if you were to do animals, perhaps the model would recognize cats, dogs, and hamsters).

Examples include the following (but feel free to choose your own!): Type of animal? Numbers? Facial expressions? Rock-Paper-Scissors? Mountains/Beach/City? Colors? Shapes? Different names? Different languages? Kinds of music?

Create a Google Doc with your name listed at the top. In the document explain your goal, along with examples from the different categories you will use for training.

Create the Teachable Machine system and train your model. (Hint: If capturing images using your webcam, isolate the items in front of a neutral background.)

Add screenshots to your Google Doc showing it working. 

In your Google Doc explain whether or not it worked. Did it meet your expectations? Why or why not?

Include a link to your model in the Google Doc as well (see image below on how to do this).

Once you are done, you must set the permissions to allow us to access and grade your work!  Do this as follows:

1. Select the green "Share" button near the upper-right

2. Select "Share with University of Illinois at Chicago" to change the behavior (If you are logged in with a non-UIC account, it will instead say "Change")

3. Select the option for "Anyone with the Link" is an "Viewer"

Once you are done and once you make sure the permissions are set correctly, copy the link to your Google Doc (to do this, click on "Copy link" under the sharing settings). Turn this in on Gradescope, under "Lab 13: Teachable Machine." 

This lab must be completed by 11:59pm on Tuesday, November 19, 2024. No late submissions will be accepted.

Grading Rubric:

0.5 points: Explanation is given of your goal, and examples of the 3 or more different categories are given as well.

0.5 points: Screenshot of model is included in Google Doc. System is trained. A link to the model is included in Google Doc.

1 point: Explanation of whether or not the model worked and if it met expectations is provided.

Extra credit (1 point): Create an additional model for a different medium (e.g. if you used images for your first model, use audio or poses for this second one). Provide the same information as before - explain your goal, along with examples of the 3 or more different categories, and provide at least one screenshot, a link to the model, and an explanation of whether or not it worked. Finally, compare and contrast the two models: What are some similarities? What are some differences? How does the performance of the two models compare? Do you notice any differences between using the two mediums?