EDUCATOR PORTAL

MISSION MATERIALS

---

MISSION RESOURCES

Learning Goals

1. Students will understand the basic principles of ML
2. Students will gain a basic understanding of how ML works
3. Students will understand using the camera as a sensor
4. Students will build hardware that interacts with an existing ML model to track facial expressions

Learning Activities

The following sections will contain step by step instructions for ELA and ELD extensions directly related to this Mission. Adjust the directions to fit your ELA and ELD standards.

Grades 3-5

ELA Extension: Recognizing Piperbot Display the three character images below (Piperbot, Pip, and a creeper from Minecraft).

Ask students to share how they recognize the difference between the characters. They may say, “I’ve seen them before” or “It looks like a mouse.”

Discuss how artificial intelligence would recognize the difference between the characters - it would learn from seeing multiple images of the same character, and look for patterns. Ask students what are the patterns that tell us whether it is a robot, mouse, or creeper? Students should be able to describe particulars including colors, and the specific grid position of features like eyes, nose, mouth. How is pattern recognition the same and different from how humans perceive?

Now, have students consider how these images are very simple (64 pixels) compared to say, a self-driving car recognizing a pedestrian or a cyclist. When asking a computer to perceive photos or even video, the model will need an immense amount of examples to learn from.

ELD Extension: Pose Estimation Person Keep the code and build for the Pose Detect mission open. Have students click on the Camera Icon to view the pose estimation overlaid onto the webcam. Students should back up and stand in front of the camera and watch which body parts map to each dot. Students can also reference the drop down menu from the “camera (nose) pose (X position)” block to see which body parts pose detection works with.

Students should use these resources to draw and label a person with pose detection dots mapped onto each body part, including the list below. Students can draw the person in any pose. (This page has an example image for multiple poses: https://www.tensorflow.org/lite/examples/pose_estimation/overview)

Nose Left Eye Right Eye Left Ear Right Ear Left Shoulder Right Shoulder Left Elbow Right Elbow

Left Wrist Right Wrist Left Hip Right Hip Left Knee Right Knee Left Ankle Right Ankle

Grades 6-8

ELA Extension: COCO Animals COCO, or Common Objects in Context, is a large-scale object detection, segmentation, and captioning dataset. Students can browse the Coco image dataset at https://cocodataset.org/#explore

Choose 1 animal and click “Search” to view images tagged with that animal.

Students should answer the following questions about their animal search.

• How many images did your search produce? (This will display right under the search bar.)
• How many different poses or angles did you see of your animal? Describe at least 3 of the different poses (ie, a horse as a logo on a sign and a horse pulling a cart.)
• If you were going to teach what a zebra is, do you think it would be easier to teach a human or to teach an artificial intelligence? Why?
• This dataset has a total of 2.5 million labeled instances in 328,000 images. Why do you think a machine learning model needs so many images to be trained effectively?

ELD Adaptation: COCO Animals COCO, or Common Objects in Context, is a large-scale object detection, segmentation, and captioning dataset. Students can browse the Coco image dataset at https://cocodataset.org/#explore

Choose 1 animal and click “Search” to view images tagged with that animal.

Students should answer the following questions about their animal search.

• How many images did your search produce?
• What are the features of your animal:
  • How many legs?
  • What colors?
  • How do you recognize this animal?
• What tells the computer which part of the images is an animal? (The outlines, or bounding boxes)

Career Connections

Hardware Diagram

Blueprint

DOWNLOAD BLUEPRINT

Have an different version? Look for more information on the Support Page.

Troubleshooting Tips

Our customer support specialists are on hand to ensure your implementation of Piper runs seamlessly.
View Support Docs or Contact Support

---

Base EXPEDITION RESOURCES

Helpful Links

• Explanation of Pull Up and Pull Down
  What are Pull-Up and Pull-Down Resistors? Read this overview to learn more.
• Google Teachable Machine
  A fast, easy way to create machine learning models for your sites, apps, and more – no expertise or coding required.
• The Artificial Intelligence (AI) for K-12 initiative
  The initiative is developing national guidelines for AI education for K-12, an online, curated resource Directory to facilitate AI instruction, and a community of practitioners, researchers, resource and tool developers focused on the AI for K-12 audience.

Vocabulary Words