Teacher material · Page · Unpublished in Canvas

Facilitator Guide: Picture This Day 1 — Train It, Then Break It

Canvas state
Unpublished

TEACHER ONLY · LESSON GUIDE

Picture This Day 1 — Train It, Then Break It

Use this page to run the lesson. The student item directly below remains the independent and absence route.

Daily Learning Contract

Topic: Image classification and model limitations

Objective: Students will analyze how training-data variety affects an image classifier's performance on unseen examples and identify a targeted improvement.

TEKS:
  • §126.19(c)(3)(B) — discuss and implement a design process that includes planning, selecting digital tools to develop, test, and evaluate design limitations, and refining a prototype or model; and
  • §126.19(c)(6) — The student uses digital tools to transform data, make inferences, and predictions. The student is expected to use digital tools in order to transform data, analyze trends, and predict possibilities and develop steps for the creation of an innovative process or product.

Demonstration of learning: Two trained classes, unseen-example test results, and one explained failure condition with the training examples needed next.

Before class

Time: One class period

Student route:
Day 1 · Meet the AI: Train It, Then Break It

Use what already exists: Picture This Day 1 deck; Teachable Machine; a student-device webcam permission check.

Lesson flow

  1. Launch: Ask what a camera can and cannot know about an image.
  2. Model: Model clear classes, training examples, and a deliberate test failure.
  3. Do: Pairs train and test a two-class model.
  4. Check + support: Check class distinction and evidence language; use a planning card and rotate camera/operator roles.
  5. Close + handoff: Record the failure condition and hand off to training-data evidence.

Check before students leave

Ask each student or team to point to the specific evidence named above and explain one choice, test, or revision. Completion alone is not evidence of understanding.

Support + troubleshooting

Scaffold: Keep the visual task sequence visible, assign purposeful partner roles, and use the sentence stems on the student page before simplifying the task.

When stuck: Ask students to show the exact step where the system stopped, then check setup, inputs, and saved work before replacing the task.