Artificial Intelligence Curriculum

Created by: Jui Khankari

Icons + what they mean

_________________________________________________________________________________________

Computer needed

Class Think + Share

Partner Turn + Talk

Independent Worksheet

Day 1

Pre-test!

Video: What is Artificial Intelligence?

Class Think

• What problems can be solved using AI?
• Raise your hands to share, now that you have watched the video.

Eliza Chatbot

Raise your hand to suggest what we should say to Eliza!

Siri/Alexa/Cortana

Raise your hand to suggest what we should say to Siri/Alexa/Cortana!

Class Think

• How is Siri/Cortana/Alexa different from Eliza?
• Raise your hands to share

Video: AI vs ML vs DL

Turn and Talk

• Using the class list, write down which problems can be solved using AI
• What about ML?
• What about DL?
• Talk with your table about your decisions

Day 2

AI, ML, and DL

Self-driving cars

AI

AI, ML, and DL

AI and ML

Spam filters

Problems that can be solved using AI

​

On your worksheet, check off the ones that match the list we created on day 1.

Healthcare

Transportation

to make it more efficient

​

​

Education

Energy & Wildlife Conservation

​

​

AI-controlled thermostats could turn on heating/cooling systems automatically only when needed, saving energy costs

​

​

Track animal movements to see which habitats animals most use and would be best to protect

Finance

​

​

Predict future prices looking at past trends

General

​

​

​

Identify spam emails

​

​

​

Recommend products (on e-commerce websites)

​

Video: Types of AI/ML/DL Algorithms

​

Class Activity

​

Post-it Note Four Corners

Problems that AI can Solve (all on one slide)

Move to the area of the room that corresponds to the AI algorithm that can solve each problem.

• Classify a tumor as benign or malignant (cancer vs no cancer)
• Self-driving cars
• Predict future prices using prior trends
• Identify spam emails

Day 3

Class Activity: Points on a line

• If I were to add another point, where would it go?

Video: Linear Regression

Stock Market Graph

Turn and Talk

How can we use linear regression with stock market graphs?

Activity: Human Neural Networks

Step 1: Training

​

Silently notice details that are unique to each snake. You can write them down on your worksheet to help you remember.

Venomous Coral Snake

Non-venomous Red Milk Snake

Non-venomous Red Milk Snake

Venomous Coral Snake

Venomous Coral Snake

Venomous Coral Snake

Non-venomous Red Milk Snake

​

Non-venomous Red Milk Snake

​

Step 2: Testing

​

Write down your guesses in the “Part 2 - Testing Guesses” section on your worksheet

Ground-Truth (Answers)

• Non-venomous milk snake
• Venomous coral snake
• Venomous coral snake
• Non-venomous milk snake
• Non-venomous milk snake
• Venomous coral snake

Put a checkmark (✅) next to all of the ones you guessed correctly

Step 3: Performance Metrics

Calculate your accuracy

Accuracy = number correct (number with a ✅ next to them)

total number (6)

​

Write down your accuracy on your worksheet

Video: Neural Networks

​

Turn and Talk

• How do our brains learn?
• How did you learn to identify venomous snakes vs non-venomous snakes?
• How do neural networks learn?
• Compare neural networks to our brains

Class Discussion

• Are neural networks AI, ML, or DL?
• Looking back at the problems list we created on day 1, what problems can be solved with neural networks?

​

Activity: Neural Networks Exploration

Step 1: Neural Networks Playground

Tensorflow Playground

Go to this link: https://tinyurl.com/ainspireneuralnetworks

On your worksheet, record what happens when you change the following hyperparameters:

• Epochs (how many times the model goes through the training data, how many times you repeat a set of flashcards)
• Learning rate (how fast the model changes its weights and biases)
• Number of hidden layers and number of neurons per hidden layer (how many total neurons the neural network has)

Hyperparameters: Values that the you, the researcher, control

​

​

Class Discussion

• What happened when:
• Epochs increased?
• Learning rate increased?
• Number of hidden layers and neurons increased?

​

Step 2: Create Your Own Neural Network

Create your own neural network

• Go to: https://teachablemachine.withgoogle.com
• Watch the introductory video
• Turn and talk: how is Teachable Machine a neural network?

​

Neural Network: Chihuahuas and Blueberry Muffins

Chihuahuas and blueberry muffins look very similar! Can we train a neural network to differentiate between them?

​

​

​

​

​

Let’s Get Started!

• Save 10 images of chihuahuas to your computer
• Save 10 images of blueberry muffins to your computer
• Click on “get started”
• Click “image project”
• Click on “standard image model”
• In class 1, type “Chihuahuas”
• In class 2, type “Blueberry muffins”
• Upload your images of chihuahuas and blueberries to the proper classes
• Click on “advanced models”
• Change the number of epochs, learning rate, and batch size to your liking
• Click on “train model”
• In preview, change the input from “webcam” to “file”
• Drag in images that you didn’t train on to test on

Try to train your own neural networks

Reflection Worksheet

​

​

​

Take a few minutes to reflect on your 3-day journey into artificial intelligence!

Download the Worksheet Packet