AI and Machine Learning Module

Introduction to Machine Learning

Resources
 AI and Machine Learning - Lesson 1
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Classifying Fish

Level #1: Recognizing Fish

Define
What is our goal? Identify fish so A.I. bot can clean up the ocean by removing the trash

Prepare
What kind of data do you have? Images of fish!

Try
How are you labeling your data?

How many images did you train with? ___________________

Reflect
How many fish images did A.I. Bot get correct? _____ / 20

Level #2: Recognizing Fish Features

Define
What type of fish are you training A.I. Bot to detect? _________________________________

Prepare
What kind of data do you have? Images of fish!

Try
How are you labeling your data?

How many images did you train with? _____________________________________________

Reflect
How many fish images did A.I. Bot get correct? _____ / 20

Click the Information icon, then fill in the chart with the Most Important Fish Parts

What did A.I. Bot learn was the most important feature for making a decision? ______________

What did A.I. Bot learn was the least important feature for making a decision? ______________
 AI and Machine Learning - Lesson 1

Level #3: Recognizing Expressions

Define
What new word are you teaching A.I. Bot to recognize?_____________________________

Prepare
What kind of data do you have? Images of fish!

Try
How are you labeling our data?

How many images did you train with? _________________

Reflect
On the results screen, you can see 20 fish that A.I. Bot has chosen. How many of these fish do you agree with?
____ / 20

Click the Information icon then fill in the chart with the Most Important Fish Parts

What did A.I. Bot learn was the most important feature for making a decision? ______________

What did A.I. Bot learn was the least important feature for making a decision? ______________

Do you agree with A.I. Bot that these are the most important or least important features? ________________________

______________________________________________________________________________________________

How could you help A.I. Bot continue to improve how it makes its decision? __________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________
AI and Machine Learning Module
Types of Machine Learning
Resources
 AI and Machine Learning - Lesson 2
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Types of Machine Learning

Level 1 - Looking for Patterns

Each box on the screen has 3 numbers in them. You can click-and-drag each box to move them around the screen. Try
to group boxes together in any way that makes sense to you. You could compare individual rows of numbers to other
boxes, or compare numbers within the box itself - whatever makes sense to you.

1) What strategy did you use to group the boxes together? _______________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

2) After you clicked the Next button, did your rectangle groups follow any patterns? ____________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

Level 2 - Green Glass Door

In this game, we’re trying to learn what words are allowed through the Green Glass Door. Some words are accepted
and others are rejected. Try to determine what words are allowed through the Green Glass Door.

1) What are three words that were rejected from the Green Glass Door: ___________________________________

____________________________________________________________________________________________

2) What are three words that were accepted from the Green Glass Door: __________________________________

_____________________________________________________________________________________________

3) What do you think the strategy is for getting words accepted through the Green Glass Door? _________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________
 AI and Machine Learning - Lesson 2

Supervised and Unsupervised Learning

Each scenario represents an example that is similar to Supervised learning or Unsupervised learning. Read through
the examples, then identify the scenarios that you think are most similar to Supervised learning and the ones that are
most similar to Unsupervised learning.

A You see someone playing your favorite song on a piano. After they leave, you sit at the
piano and press random notes on the piano until you start to recreate the melody from the
song.

B You go to a swim coach who helps you learn how to swim in a pool. You watch them
demonstrate basic strokes before practicing on your own

C The elders in your family show you how to cook meals from recipes passed down from
generation to generation. You read the recipe and watch them cook the meal first before you
try making it yourself.

D A toddler watches all of the adults in their life walk around the house. Eventually, after a lot
of falling and stumbling, they learn to find their balance and start walking.

E You know that there are spices in your kitchen, but you’re not sure what to do with them.
You add little bits of different spices and taste as you go until you settle on a taste that you
like.

F You study different vocabulary words for science class by creating flash cards. On one side
is the word and the other side is the definition. As you study, you gradually start to memorize
the terms.

1) What are two scenarios above that you think are most similar to Supervised learning? ________ and __________

Why do you think these scenarios are Supervised learning? ______________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

2) What are two scenarios above that you think are most similar to Unsupervised learning? ________ and __________

Why do you think these scenarios are Unsupervised learning? ____________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________
AI and Machine Learning Module
Innovations in AI
Resources
 AI and Machine Learning - Lesson 3

Innovations in AI Research Areas

Instructions
This document contains several AI research areas, along with sample articles and
videos for each. This is not a complete list, so use it to help you get started before
exploring beyond!

Computer Vision and Sensors

● Video: How computer vision works [link]
● Article: Five exciting computer vision applications [link]
● Video: AI, Sensors, and Data [link]

Healthcare, Human Life, and Well-being

● Video: Top AI algorithms in healthcare [link]
● Video: Why we don’t have self-driving cars yet [link]
● Article: AI solves a 50 year old science problem [link]
● Video: AI won’t replace therapy - yet [link]

Environment, Natural Resources, and Wildlife

● Video: AI and sustainability [link]
● Article: Detecting plastic litter in natural environments [link]
● Video: Saving biodiversity with tech and AI [link]

Human-Robot Interaction
● Video: Interview with a humanoid robot [link]
● Video: How Alexa works [link]
● Video: Creating an “AI baby” [link]
● Article: What are AI chatbots [link]

Art, Design, and Culture

● Video: A tool for dance choreography inspired by AI [link]
● Article: Poem portraits [link]
● Article: Artist style transfer in paintings and video [link] [video link]

Employment, Economy, and Finance

● Article: How AI can create jobs [link]
● Video: 5 career paths in AI [link]
 AI and Machine Learning - Lesson 3

Innovations in AI Research Areas

● Video: The future of your job in the age of AI [link]

● Video: The big debate about the future of work [link]
AI and Machine Learning Module
Patterns in Data
Resources
 AI and Machine Learning - Lesson 4
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Patterns in Data

Overview
A survey has been given to 500 people to help decide what types of shoes to recommend. Using the data, you need to
identify any patterns or trends in the data that can help our model make predictions. We should look for “hot spots” in
the data where there are large clusters of answers, then use that to identify trends.

Question 1: Where do you plan to spend time today?

Shoe Preference
cowboy boots crocs dress shoes sandals tennis shoes work boots
bed 2 13 0 16 0 0
city 23 6 15 10 47 40
indoors 14 10 50 18 88 10
nature 15 19 12 28 17 47

What is a trend you can find in the data? Use it to complete this sentence:

People who answer _____________________________ also tend to wear __________________________

Question 2: What is the weather like today?

Shoe Preference
cowboy boots crocs dress shoes sandals tennis shoes work boots
rainy 18 8 17 8 11 25
snowy 5 5 21 0 0 55
sunny 16 16 15 41 74 6
windy 15 19 24 23 67 11

What is a trend you can find in the data? Use it to complete this sentence:

People who answer _____________________________ also tend to wear __________________________

Question 3: Do you plan on wearing socks today?

Shoe Preference
cowboy boots crocs dress shoes sandals tennis shoes work boots
no 36 29 20 61 25 0
yes 18 19 57 11 127 97

What is a trend you can find in the data? Use it to complete this sentence:

People who answer _____________________________ also tend to wear __________________________

Question 4: How would you describe your plans for today?

Shoe Preference
cowboy boots crocs dress shoes sandals tennis shoes work boots
casual 21 15 4 26 78 9
lazy 7 13 0 34 29 0
professional 22 18 73 0 34 12
work 4 2 0 12 11 76

What is a trend you can find in the data? Use it to complete this sentence:

People who answer _____________________________ also tend to wear __________________________

Question 5: Do you want people to notice and compliment your shoes?

Shoe Preference
cowboy boots crocs dress shoes sandals tennis shoes work boots
no 0 19 42 35 29 53
yes 54 29 35 37 123 44

What is a trend you can find in the data? Use it to complete this sentence:

People who answer _____________________________ also tend to wear __________________________

Shoe Predictions
1) Using the data from all five questions, describe a recommendation you think our model can make based on the
answers to several questions. You can use these sentence starters to help:

Our model should recommend shoe ______________________________ when people select the following answers:

A. _________________________________________________________________________________

B. _________________________________________________________________________________

C. _________________________________________________________________________________

2) Using the data from all five questions, are there any responses that don’t seem to be very helpful in predicting
shoes? This means all of the answers to that question are around the same value, so it’s hard to find any “hot spots”
that can help our model make predictions.

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________
AI and Machine Learning Module
Classification Models
Resources
 AI and Machine Learning - Lesson 5

Activity Guide - Classification Models

Fruit Or Veggie?
You should have been given a set of cards
that has either a fruit or a veggie on it.

Step 1

Determine how “sweet” and how “easy to

eat” each food is. If you’ve never eaten the
food before or you’re not sure, you can use
the recommended values on the card.

Step 2

Put a point on the graph for each food.

● If the food is a fruit, write a small F next
to the dot.
● If the food is a veggie, write a small V
next to the dot

When you are finished, you should have 10

fruit dots and 10 veggie dots on your graph.
 Fruit: Apple Fruit: Banana Fruit: Cherry Fruit: Grape Fruit: Kiwi

Recommended Features Recommended Features Recommended Features Recommended Features Recommended Features
Sweetness: 7 Sweetness: 6 Sweetness: 8 Sweetness: 8 Sweetness: 7
Eat: 6 Eat: 5 Eat: 7 Eat: 10 Eat: 4

Fruit: Watermelon Fruit: Orange Fruit: Pear Fruit: Pineapple Fruit: Strawberry

Recommended Features Recommended Features Recommended Features Recommended Features Recommended Features
Sweetness: 7 Sweetness: 6 Sweetness: 7 Sweetness: 8 Sweetness: 8
Eat: 2 Eat: 4 Eat: 5 Eat: 2 Eat: 8

Veggie: Broccoli Veggie: Carrot Veggie: Corn Veggie: Potato Veggie: Zucchini

Recommended Features Recommended Features Recommended Features Recommended Features Recommended Features
Sweetness: 3 Sweetness: 4 Sweetness: 2 Sweetness: 2 Sweetness: 4
Eat: 7 Eat: 8 Eat: 2 Eat: 3 Eat: 7

Veggie: Onion Veggie: Lettuce Veggie: Mushroom Veggie: Cauliflower Veggie: Radish

Recommended Features Recommended Features Recommended Features Recommended Features Recommended Features
Sweetness: 2 Sweetness: 2 Sweetness: 4 Sweetness: 2 Sweetness: 1
Eat: 4 Eat: 8 Eat: 10 Eat: 7 Eat: 6
AI and Machine Learning Module
Introduction to AI Lab
Resources
 AI and Machine Learning - Lesson 6
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Training a Model in AI Lab

Level 2: Recognizing Shapes

Explore AI Lab to create models that can classify a shape. You should train multiple models from the same dataset and
keep track of the accuracy by filling in the table below:

Features Accuracy
sides

fill color

border color

border color, fill color

sides, background color

sides, background color, size

sides, border color, fill color, background color, size

Level 3: Pizza Recommendations

Another type of problem machine learning can help with is recommendations. You have been hired by a restaurant to
help figure out what customers like best! You collected a bunch of data about people’s preferences on various pizza
toppings like pepperoni, peppers, anchovies, fried chicken, and others. You’re going to use this data to find patterns in
the customer’s preferences!

Do This: Train a model that is at least 80% accurate using at most 4 features. If you don’t get 80% accuracy, you can
always go back and choose new features.

Features Accuracy

Describe the strategy you used when choosing your features and trying to create an accurate model
Level 4: You Choose Restaurant Recommendations
Based on your success with the Pizza parlour, several other restaurants want to hire you to find patterns in what their
customers like. You have several restaurants to choose from with their own sets of data to analyze and make
recommendations with.

Do This: Choose from one of the remaining datasets to investigate and create a recommendation model for. Describe
the dataset and its features below:

Which dataset did you choose?

What features did you decide to train your model with?

After training: what is the accuracy of your model?

You need at least 80% accuracy to continue

Describe the strategy you used when choosing your features and trying to create an accurate model

This model can recommend a series of toppings that matches the preferences to their previous customers. How do
you think a restaurant could use this model in their restaurant?

This model was only trained on 500 people to match their particular preferences. What do you think are some
limitations to this model?
AI and Machine Learning Module
Model Cards
Resources
 AI and Machine Learning - Lesson 8
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Model Cards

Creating A Model Card

Model Cards can be used to help document our machine learning models. When creating models, they help us keep
track of the decisions we make, like what data we decide to use or the features to focus on.

In a previous lesson, we helped create an app for a Class Book Recommender. Now, let’s help that class fill out a
model card for the app.

Do This
Part of the model card has already been filled out. Fill in the remaining boxes below about the Book Recommendation
model. Some questions to consider are:
● Is there enough data for this model to be used as intended?
● Does the data represent all possible users and scenarios?
Medical Priority Models
Even though we discovered bias in our original model, there are still benefits for an app like this to help nurses and
medical professionals in the ER. Several other companies have created medical priority models that can be used to
replace the biased one that we discovered. They’ve also sent along the Model Cards for us to evaluate. Looking at
these model cards, can we find a model that we would recommend to the hospital to help decide medical priority?

Level 3: ER Recommender V1
Would you recommend that this hospital use this model instead? Why or why not?
Be sure to refer to the model card when making your recommendation

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

Level 4: Northern Lights ER Priority

Would you recommend that this hospital use this model instead? Why or why not?
Be sure to refer to the model card when making your recommendation

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

Level 5: AI Medical Recommender

Would you recommend that this hospital use this model instead? Why or why not?
Be sure to refer to the model card when making your recommendation

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

Reflection
If you could offer guidance to these companies training machine learning models, what advice would you give them to
help create useful and unbiased machine learning models?

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________
AI and Machine Learning Module
Saving Models in AI Lab
Resources
 AI and Machine Learning - Lesson 9
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Zoo Models

Overview
We’ve been hired by a zoo to create different apps for customers to use around the
park. A user should be able to answer a few questions about the animal they see, and
then the app will tell them something new about the animal.

The zoo has created a dataset for us to use, which includes different properties of the
animals they have in the zoo. Our job is to train three different models using the dataset
they provided.

Level 1: Is It a Predator?
1) Train a machine learning model to help visitors identify whether or not the animal in
the exhibit is a predator. Your model should be at least 70% accurate.

Label Features Accuracy

At least 70%

Predator

2) Save your model and create a model card. The zoo has asked us to use the following information for the Intended
Use and Limitations section:

Intended Use

Can identify whether a zoo animal hunts its prey like a predator

Limitations and Warnings

The data is only from animals in this zoo. Don’t use this app outside of the zoo.

Level 2: What Type of Animal is it?

1) Train a machine learning model to help visitors identify what type of animal they are looking at - for example, a
mammal or reptile or bird, etc. Your model should be at least 70% accurate.

Label Features Accuracy

At least 70%

Class

2) Save your model and create a model card. The zoo has asked us to use the following information for the Intended
Use and Limitations section:

Intended Use

Can identify what type of animal you are looking at

Limitations and Warnings

 The data is only from animals in this zoo. Don’t use this app outside of the zoo.

Level 3: You Choose

1) The zoo has asked you to decide the label for one of their models. You can use any label in the dataset, then find
features that will create an accurate model for your data. For example, you could create a model to tell whether an
animal is domesticated, or lays eggs, or is aquatic.

Label Features Accuracy

At least 70%

2) Save your model and create a model card. You decide how you would like to fill in the spaces below.

Intended Use

Limitations and Warnings

Level 4: Create an App

Choose one of the models that you just trained to import into App Lab. Create an app that visitors can use when
wandering around the Zoo and looking at different animals.

Reflection
Why is a model card useful when creating a model in AI Lab? _____________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

How is a model card useful when creating an app in App Lab? ____________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________
AI and Machine Learning Module
Numerical Models
Resources
 AI and Machine Learning - Lesson 11
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Numerical Accuracy

Overview
You’ve managed to escape the zombie outbreak and find help. You’re also able to survey your town and see how many
zombies were actually wandering around the town. Here is the data comparing the predicted number of zombies with
the actual number of zombies:

Checking Our Accuracy

Location Actual Prediction Exact Match? Within 5? Within 20?

A 40 25

B 72 74

C 31 35

Other teams were also trying to predict how many zombies were in their towns. Using the data below, calculate the
accuracy in three different ways: exact matches, within 5, and within 20.

Location Actual Prediction Exact Match? Within 5? Within 20?

🍕 32 35

🍔 56 72

🌭 22 19

🥗 33 33

🍛 85 81

🌮 34 73

🥪 53 56

🍿 2 0

🥡 82 55

🥞 76 80

How many predictions were an exact match to the actual value? ________ / 10 = ______ %

How many predictions were within 5 of the actual value? ________ / 10 = ______ %

How many predictions were within 20 of the actual value? ________ / 10 = ______ %
 AI and Machine Learning - Lesson 11
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Zombie Prediction

Overview
A zombie outbreak is happening across the world! The zombies are spreading in your
town, but you don’t know where they are congregating! Luckily, you have data from a
neighboring town about where the zombies tended to gather. Using their data, can we
predict how many zombies will be at locations in our town and pick the safest location
to hide?

Our Data

The table below shows the data we’ve received from a neighboring town

Location Building Type Noise Level Sidewalk? ZOMBIES 🧟

1 indoor school medium yes 51

2 outdoor zoo medium yes 66

3 indoor library low no 9

4 indoor restaurant medium yes 45

5 outdoor park low no 21

6 outdoor school medium no 54

7 outdoor restaurant high no 60

8 outdoor park high yes 96

What are some of the things you notice in the data above? Fill in the blanks below

● The greatest number of zombies are ___________ at location ______

● The smallest number of zombies are ___________ at location ______

● There tend to be more zombies in _____________________________ locations

● There tend to be less zombies in ______________________________ locations

● What is something else that you notice? ______________________________________________________

_______________________________________________________________________________________

_______________________________________________________________________________________
Creating a Model

In order to turn this data into a model that can make a prediction, we’ll take unknown data from our town and find
similarities with the data you already have. Here is information about one location that is close to us:

Location Building Type Noise Level Sidewalk? ZOMBIES 🧟

A indoor restaurant low no

Do This: Compare Location A with the features from the previous data:
● Anytime a column from Location A matches a column for another location, that’s one point of similarity.
● After comparing each column, you can add up all the points of similarity.
● Use the table below to keep track of how similar Location A is to the other locations.
● The first two rows have already been completed.

Location Building Type Noise Level Sidewalk? ZOMBIES 🧟 A Similarities

1 indoor school medium yes 51

2 outdoor zoo medium yes 66

3 indoor library low no 9

4 indoor restaurant medium yes 45

5 outdoor park low no 21

6 outdoor school medium no 54

7 outdoor restaurant high no 60

8 outdoor park high yes 96

What are three locations that are most similar to Location A? If some locations have the same level of similarity,
choose the ones highest in the list.

The three locations most similar to Location A are: ______, ______, and ______

To predict how many zombies could be at this location, we need to take the average of the three most similar locations.
1. Calculate the sum by adding together all the zombie values from these locations: ______
2. Divide the sum by the number of values: ____
3. Write this value in the Zombies column at the top of this page. That’s the prediction!

2
Predicting Multiple Locations

There are two other locations that are nearby. We need to decide whether to go hide at location A, B, or C. Use the
same strategy to predict how many zombies will be at location B and at location C.

Location Indoor/Outdoor Type Noise Level Sidewalk? ZOMBIES 🧟

B outdoor restaurant high yes

C indoor school low yes

Location Building Type Noise Level Sidewalk ZOMBIES B Similarities C Similarities

1 indoor school medium yes 51

2 outdoor zoo medium yes 66

3 indoor library low no 9

4 indoor restaurant medium yes 45

5 outdoor park low no 21

6 outdoor school medium no 54

7 outdoor restaurant high no 60

8 outdoor park high yes 96

What are three locations that are most similar to Location B? If some locations have the same level of similarity,
choose the ones highest in the list.

The three locations most similar to Location B are: ______, ______, and ______

To predict how many zombies could be at this location, we need to take the average of the three most similar locations.
1. Calculate the sum by adding together all the zombie values from these locations: ______
2. Divide the sum by the number of values: ____
3. Write this value in the Zombies column at the top of this page. That’s the prediction!

What are three locations that are most similar to Location C? If some locations have the same level of similarity,
choose the ones highest in the list.

The three locations most similar to Location C are: ______, ______, and ______

To predict how many zombies could be at this location, we need to take the average of the three most similar locations.
1. Calculate the sum by adding together all the zombie values from these locations: ______
2. Divide the sum by the number of values: ____
3. Write this value in the Zombies column at the top of this page. That’s the prediction!

Which location is predicted to have the least amount of zombies - location A, location B, or location C?

____________________________________

3
AI and Machine Learning Module
Numerical Data in AI Lab
Resources
 AI and Machine Learning - Lesson 12
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Numerical Data in AI Lab

Overview
We’ve been hired by a Safari company to create a machine learning model to find patterns with how animals behave in
the park. The company is specifically interested in predicting how many lions visitors will see - lions are a main
attraction for the company, but they can also be dangerous and require additional staff to make sure everyone is safe.
Our job is to create a model that can predict how many lions will be in the park on a particular day.

Feature #1: Antelopes

Click on the Antelopes column in AI Lab. Look at the scatter plot that appears, which compares the number of
antelopes with the number of lions.

1) If there are a low number of antelopes in the park, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

2) If there are a high number of antelopes in the park, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

3) Why do you think this is?

______________________________________________________________________________________________

______________________________________________________________________________________________

Feature #2: people

Click on the people column in AI Lab. Look at the scatter plot that appears, which compares the number of people with
the number of lions.

1) If there are a low number of people in the park, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

2) If there are a high number of people in the park, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

3) Why do you think this is?

______________________________________________________________________________________________

______________________________________________________________________________________________
Feature #3: dayOfMonth
Click on the dayOfMonth column in AI Lab. Look at the scatter plot that appears, which compares the day of the month
with the number of lions.

1) What happens if you visit on an early day of the month? How many lions do you think you’ll see?

______________________________________________________________________________________________

______________________________________________________________________________________________

2) What happens if you visit on a late day in the month? How many lions do you think you’ll see?

______________________________________________________________________________________________

______________________________________________________________________________________________

Feature #4: temperature

Click on the temperature column in AI Lab. Look at the scatter plot that appears, which compares the temperature with
the number of lions.

1) If there is a low temperature that day, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

2) If there is a high temperature that day, what does that mean for how many lions could be in the park?

______________________________________________________________________________________________

______________________________________________________________________________________________

3) Why do you think this is?

______________________________________________________________________________________________

______________________________________________________________________________________________

Additional Features
Continue to explore the data by clicking on the remaining features in the dataset.

1) What is another column you think would be a good feature for this model? ________________________________

Why? _________________________________________________________________________________________

______________________________________________________________________________________________

2) What is another column you think would be a bad feature for this model? __________________________________

Why? _________________________________________________________________________________________

______________________________________________________________________________________________

Train a Model
Using any of the possible columns, train a model that is at least 80% accurate
AI and Machine Learning Module
AI Code of Ethics
Resources
 AI and Machine Learning - Lesson 14

AI Ethics Research Areas

Instructions
This document contains several AI ethics research areas, along with sample articles
and videos for each. This is not a complete list, so use it to help you get started before
exploring beyond!

Computer Vision and Sensors

● Video: Facial recognition [link]
● Article/Video: AI, gender, and race [link]
● Video: AI and policing [link]
● Video: AI in the Classroom [link]

Social Media, Shopping, and Entertainment

● Article/Videos: How AI recommends videos [link]
● Video: AI, social media, and elections [link]
● Video: AI and social media marketing [link]
● Article: Pros and cons of AI in social media [link]
● Article: Social and philosophical impacts of video game AI [link]

Human-Robot Interaction
● Video: Talking to an AI virtual assistant [link]
● Article: Why voice assistants are female [link]
● Article: AI, race, and customer service [link]
● Video: Social media “virtual teenager” experiment [link]

Employment, Economy, and Finance

● Article: AI and job replacement [link]
● Article: AI and background checks [link]
● Video: AI and medical jobs [link]
● Article: How human labor powers AI [link]

Healthcare, Human Life, and Well-being

Ethical Issues
● Article/Video: Racial biases in healthcare AI [link]
● Article: AI and COVID-19 [link]
● Video: AI and the future of healthcare [link]
 AI and Machine Learning - Lesson 14

AI Ethics Research Areas

● Video: Self-driving cars and human life [link]

● Article: How AI bias happens and how to fix it [link]

Environment, Natural Resources, and Wildlife

Ethical Issues
● Article: Environmental costs of AI [link]
● Article: Renewable energy for AI [link]
● Video: How will AI impact society and the environment [link]
● Article: AI and treatment of animals [link]
 AI and Machine Learning - Lesson 14
Name(s)_________________________________________ Period ______ Date ___________________

Activity Guide - AI Ethics Research Reflection

Instructions
As you research your AI ethics area, answer the following questions. Be prepared to
share what you have learned with your classmates!

Research Area: ____________________________________________________________

1. Describe a technology that uses AI in your area of research. How does it work?
Who uses it?

2. How can this technology benefit society?

3. Think of a specific ethical issue related to this technology. Why is this issue
important? Whom does it impact? How?

4. How can this issue be addressed while maintaining the social benefits of the
technology?

5. What other questions or ideas do you have pertaining to ethical issues in your
research area?
AI and Machine Learning Module
Project: Make a Machine Learning App
Resources
 AI and Machine Learning - Lesson 15
Name(s)_______________________________________________ Period _______ Date ____________________

Project Guide - Create an ML App

Overview
In this project you will design an app that uses machine learning to help
make a decision, offer a recommendation, or make a prediction. You will
have a choice of real-world datasets to use to help your app make its
decision.

Your app will show that you can…

❏ Train a model to make a prediction

❏ Create a model card
❏ Create an app that uses machine learning

Step 1: Examine your Data

Log into Code Studio and look at the datasets that are available. Choose a few to examine, looking at the columns to
see what potential labels and features you can choose from.

List 3 datasets you might be interested in and a decision, recommendation, or prediction you might be able to answer
using this data:

Dataset Decision, Recommendation, or Prediction

Step 2: Train Your Model

Choose one of the datasets above to continue investigating. Select a label and choose at least 2 features to use to
train your model. It’s okay to switch to a different dataset as you investigate. Record your results below.

What dataset did you choose?

_____________________________________________________________________________________________

What label did you choose?

_____________________________________________________________________________________________
What features did you choose? You must choose at least 2
_____________________________________________________________________________________________

_____________________________________________________________________________________________

What is the accuracy of your model?

_____________________________________________________________________________________________

Step 3: Save your Model, Create a Model Card

Save your model in AI Lab. Before continuing to App Lab, answer the following questions from your Model Card:

Describe the problem you think this model could help solve, or one potential app someone could make with
this model.
_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________

Describe any situations where this model could potentially be misused, or any places where bias could
potentially show up in the model. Important questions to consider are:
● Is there enough data to create an accurate model?
● Does the data represent all possible users and scenarios?
_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________
Step 4: Create Your App

In App Lab, import your model and begin customizing your app. Some things to consider include:

● Adding a theme to your app

● Adding a welcome screen
● Adding if-statements to change to different screens based on the result
● Separating your features so they appear on different screens
● Adding information from your model card into the app (such as the intended use or warnings)

Step 5: Reflect

What part of your project are you most proud of? _________________________________________________

Why? _______________________________________________________________________________________

_____________________________________________________________________________________________

If you had more time, what is something you wish you could change about your data to try and improve your
model? For example, making changes to how the data was collected or what columns are available.

_____________________________________________________________________________________________

_____________________________________________________________________________________________

If you had more time, what improvements would you make to your app?

_____________________________________________________________________________________________

_____________________________________________________________________________________________

Do you think other people should use your app to help solve a problem? Why or why not?

_____________________________________________________________________________________________

_____________________________________________________________________________________________

_____________________________________________________________________________________________
Create an ML App - Project Rubric

Key Concept Extensive Evidence Convincing Evidence Limited Evidence No Evidence

Project Step 1 is filled out Step 1 is filled out Step 1 is mostly filled Steps 1 of the
Preparation completely. There is at completely. There is at out, but there is at project guide is
least one app idea for least one app idea for least one box left blank
each dataset, and the each dataset, although blank either because
app is described in the descriptions may be you didn’t investigate 3
detail. vague or unclear. datasets or you
explain an app idea

Creating an Model is at least 70% Model is between 60% Model is between 50% You did not train a
Accurate accurate and you chose and 70% accurate and and 60% accurate or model or the model
Model at least 2 features you chose at least 2 you chose less than 2 is less than 50%
features features to train your accurate
model

Model Card Step 2 and Step 3 are Step 2 and Step 3 are Step 2 and Step 3 are Step 2 or Step 3 of
completed. Each completed. Each completed, but the the project guide is
prompt is answered in prompt is answered, but prompts may be vague incomplete
detail and includes a one of the prompts may or unclear for how the
specific use for the be vague or unclear model can be used or
model and a specific about how the model misused.
warning about how the could be used or
model could be used. misused

Creating an The app works without The app works without The app works without App is not
App error and includes a error and includes a error, but does not complete or does
welcome screen and a welcome screen and a contain a welcome not work
theme. The app theme. screen or theme - it
incorporates information may be a
from the model card into single-screen app.
the app, such as the
accuracy, intended use,
or warnings.

Reflection Step 5 is completed Step 5 is completed, but Step 5 is completed, Step 5 of the
with detailed and some answers may be but most answers are project guide is
specific answers. You vague or unclear. Your vague or unclear. Your incomplete or
list specific improvements may not improvements may not blank.
improvements you be specific, or your be specific, or your
would make to your reflection on whether reflection on whether
model and app. Your others should use the others should use the
reflection on whether app is vague or app is vague or
someone should use generalized. generalized.
the app is thoughtful
and specific.
AI and Machine Learning Module
Survey Planning
Resources
 AI and Machine Learning - Lesson 17
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Survey Planning

Overview
To help create your app, we’ll need to collect data that AI Bot can use to make its decision. Each person from our other
team has already added a question, but there’s still room for more.

Do This: Add three more questions to this table. Be sure to list the possible choices for categorical questions, or the
range of values for numerical questions.

Answer Choices
Question ● If the data is categorical, list the choices
● If the data is numerical, list the range of values that someone can enter

☑ Categorical 𐄂 Numerical
● Hiking Club
Label Question: What is a club you ● Band Club
would want to join or are already a ● Art Club
part of?
● Science and Technology Club
● World Cultures Club

☑ Categorical 𐄂 Numerical
● Go for a walk
Kim’s Question: Which of these ● Play an instrument
activities would you like to do in ● Draw a picture or doodle
your free time?
● Build with Lego blocks
● Cook something from a new recipe

Nico’s Question: How many minutes 𐄂 Categorical ☑ Numerical

of free time do you have for club ● Min: 20 minutes
activities? ● Max: 180 minutes

☑ Categorical 𐄂 Numerical
Isaac’s Question: How would you ● Keep to myself
describe your personality? ● Hang out with others

☑ Categorical 𐄂 Numerical
● Earth
Zoey’s Question: Which element do ● Wind
you like the most? ● Fire
● Water
● Metal

𐄂 Categorical ☑ Numerical
Hawa’s Question: How many minutes ● Min: 5 minutes
does it take you to get to school? ● Max: 120 minutes
 My Question: 𐄂 Categorical 𐄂 Numerical

My Question: 𐄂 Categorical 𐄂 Numerical

My Question: 𐄂 Categorical 𐄂 Numerical

Create Your Survey

Log into Code Studio and make a copy of the provided template. Add these questions above to the template, using the
video and Help and Tips for guidance on how to use the Google Form. Below is a checklist so you can keep track of
your work

Rename the form to be “Club Recommender Survey”

Add the Label Question: What is a club you would want to join (or are already a part of)
Add [Name]’s question: Which of these would you prefer to do in your free time?
Add [Name]’s question: How much free time do you have for club activities?
Add [Name]’s question: How would you describe your personality?
Add [Name]’s question: Which element do you prefer?
Add [Name]’s question: How many minutes does it take you to get to school?
Add Your Questions

Start Collecting Data

Use the Preview Button to open your form. Invite 5 other people to fill out your survey, or have other students in the
classroom fill it out. View the results as they come in to start noticing trends and patterns in the data.

2
AI and Machine Learning Module
Troubleshooting Models
Resources
 AI and Machine Learning - Lesson 19
Name(s)_______________________________________________ Period ______ Date ___________________

Activity Guide - Troubleshooting Models

Overview
Our team is having some issues with their data and models. In AI Lab, investigate each person’s data and try to
determine what issues they are having.

Zoey Nico Isaac Kim

“I can't seem to train an “I can train an accurate “I don't understand why I “I can train an accurate
accurate model no matter model, but when I have keep getting low accuracy, model, but when I have
how hard I try!” other people test it they even when I use a lot of other people test it they
say it doesn’t work and I’m features” say it isn’t correct and I’m
not sure why!” not sure why!”

Zoey’s Data

What do you think is the issue with Zoey’s data? _______________________________________________________

______________________________________________________________________________________________

What do you think Zoey can do to fix it, or what do you think she could have done differently earlier in her project to

avoid it? _______________________________________________________________________________________

______________________________________________________________________________________________

Nico’s Data

What do you think is the issue with Nico’s data? _______________________________________________________

______________________________________________________________________________________________

What do you think Nico can do to fix it, or what do you think he could have done differently earlier in his project to avoid

it? _______________________________________________________________________________________

______________________________________________________________________________________________
Isaac’s Data

What do you think is the issue with Isaac’s data? _______________________________________________________

______________________________________________________________________________________________

What do you think Isaac can do to fix it, or what do you think he could have done differently earlier in his project to

avoid it? _______________________________________________________________________________________

______________________________________________________________________________________________

Kim’s Data

What do you think is the issue with Kim’s data? _______________________________________________________

______________________________________________________________________________________________

What do you think Kim can do to fix it, or what do you think she could have done differently earlier in her project to

avoid it? _______________________________________________________________________________________

______________________________________________________________________________________________

Hawa’s Data

Hawa hasn’t started collecting data yet. What are three strategies you would offer her so
she doesn’t make some of the mistakes as her peers?

Strategy #1: ___________________________________________________________

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

Strategy #2: ___________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

Strategy #3: ___________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

______________________________________________________________________________________________

2
AI and Machine Learning Module
Project - Design an AI App
Resources
Name(s)_______________________________________________ Period ______ Date ___________________

Project Guide - Design an AI App

Overview
In this unit you’ve seen how AI and machine learning can be used to solve
all kinds of problems. Now it’s your turn to use them to help someone. In
this project, you will create an app that uses AI and machine learning to
help someone in your community make a decision.

You will…
● Choose an issue that could be addressed by making a decision
● Make a plan for how you will collect the data you need to make
your decision
● Use AI Lab to train and test a model
● Create a model card
● Import your model into App Lab and use it to develop your app

You will submit…

● Completed Project Guide (this document)
● Completed Data Survey form
● Completed Model Card
● Completed App Lab app

Project Steps
Step 1 - Choose an Issue Statement

We’re going to write an issue statement which describes an issue we’ve observed and
how it is affecting people in our community. Some tips for writing an issue statement
are:

● Focus on the user. Make sure that the issue you’re addressing will help
people in your community.
● Keep it broad. This lets you be creative with the kind of app we can design.
● Make it manageable. Trying to solve too big of a problem can be frustrating -
instead, focus on an issue you can see in your community.

You can use the 5 Why’s Worksheet to help develop an issue statement. You can use one of these sentence stems to
create a statement that describes the issue and how it affects people in your community.

● I notice people feel … . This is an issue because...

● I notice people need … . This is an issue because…
● I notice it’s hard for people to… This is an issue because...

What is your issue statement?

Step 2 - App Planning

Machine Learning apps are great at helping to solve problems that require making a decision, providing a
recommendation, or making a prediction. Use the space below to brainstorm different apps that could be used to
address the issue above.

For each brainstorm idea, be sure to include:

● A title
● What the app does
● Who the ideal user of the app is

Decide

Recommend

Predict

Decide on one of your choices above to use for your app. Describe the app below:

How does this app help address your issue statement?

In general, what kind of data will your app use?

2
Step 3 - Plan Your Data Collection

To help create your app, we’ll need to collect data that AI Bot can use to make its decision. It’s important to collect as
much data as we can, and make sure it represents all types of users who might benefit
from this app. In this step, we’ll make a plan for how we want to collect data from
members of our community.

Who are the ideal users of your app? Who do you imagine will benefit from
having this app available to them?

Making sure your data represents your whole community is important to avoid
bias. List three traits that are important to your users that should be represented in the data.

● ________________________________________________________________________

● ________________________________________________________________________

● ________________________________________________________________________

Describe the steps you will take to collect your data, making sure all intended users have a chance to take the
survey and their data is included in the dataset.

Step 4 - Create Your Survey

To help create your app, we’ll need to collect data that AI Bot can use to make its decision. Use the table below to list
the questions you want to ask to help collect your data. You may not end up using all of the rows.

Don’t forget to include a question for your decision. This will be the label that gets used in AI Lab in later steps. Once
the table is complete, you can use the Google Form Template to create the actual survey.

Answer Choices
Question ● If the data is categorical, list the choices
● If the data is numerical, list the range of values that someone can
enter

𐄂 Categorical 𐄂 Numerical

3
 𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

𐄂 Categorical 𐄂 Numerical

Once you’ve completed the table, it’s time to create your survey in Google Forms. Follow the instructions from your
teacher for how to access the Google form template and create your survey.

4
Step 5 - Data Reflection

It’s time to go out and collect your data! Follow the plan you outlined in Step 3 to collect data from your community.

Once you’ve collected data from your community, complete the questions below about your dataset

How many responses did you get?

What steps did you take to make sure a variety of users were able to complete the survey?

Looking at your data in Google Forms or Google Sheets, do you already notice any trends?

Step 6 - Train Your Model

Now that your data has been prepared, you’re ready to train and test your model!

1. Follow the steps in AI Lab to train your model.

2. Observe your model’s accuracy.
3. Test your model a few times to see how well it works.
4. Save your model so it is ready to import into App Lab
5. Create a Model Card

What features from your dataset did you use in your model?

Were there any features from the dataset that you didn’t use in your model? Why not?

What is the reported accuracy of your model?

Use the Model Card template to create a model card for your app. Follow your teacher’s instructions on whether to use
a paper model card or a digital model card. Pay attention to the Intended Uses and Limitations sections of the model
card, especially considering the data you collected.

5
Step 7 - Develop your App

It’s time to develop your AI App! Import your model into App Lab and begin creating your app. A rubric is provided for
how to design your app. At the very least, it needs to have:

● A design theme
● A welcome screen
● Descriptive feature and label text elements

You can also choose to use additional screens to customize your app.

Step 8: Test and Reflect

Test the app with one of the intended users of your app. This may be someone outside of
class. Record whether or not your users agree with the decision that your app made

App Recommendation Agree with the Comments about the App

Results?

▢ Yes
▢ No

If you had more time, what is an improvement you would make to how you collected data for this project?

If you had more time, what is an improvement you would make to help improve the accuracy or results of your
app?

If you had more time, what is an improvement you would make to the app itself to make it more user-friendly?

6
 Design an AI App - Project Rubric
Key Concept Extensive Evidence Convincing Evidence Limited Evidence No Evidence

Project Step 1 and 2 is filled out Step 1 and 2 are mostly Step 1 and 2 are Steps 1 and 2 of
Preparation completely. The issue filled out, but the issue mostly filled out, but the project guide
statement is clear and each statement may be the issue statement are blank
brainstorm field in Step 2 unclear or there is not an does not meet the
has at least one app idea. app idea in each guidelines or multiple
brainstorm field in Step 2 boxes or prompts are
left blank in Step 2

Survey and Steps 3-5 are filled out Steps 3-5 are filled out Steps 3-5 are mostly Steps 3-5 are
Data completely. Responses are completely, but filled out, but some blank and no
Collection thoughtful and reflective of responses are generic or responses are blank. effort was made
the community that is vague and are not Or, no effort was made to collect data
submitting data. An effort specific to the community to collect data from the from the
was made to collect data that is submitting data. community. community
from the community, even if An effort was made to
there are not a lot of collect data from the
responses. community, even if there
are not a lot of
responses.

Creating a Step 6 is completed. You Step 6 is completed. You Step 6 is mostly You did not train a
Model trained a model either from trained a model either completed, but at least model from a
survey data or a provided from survey data or a one prompt is blank. survey or
dataset, and answers to the provided dataset, but the You trained a model provided dataset
prompts are clear and answers to the prompts either from survey data
reflective are unclear or generic or a provided dataset.

Model Card Your model card is Your model card is Your model card is Your model card
complete, with detailed complete. The intended mostly complete, but is mostly
information in the intended uses and limitations at least one box is complete, but at
uses and limitations sections are complete, blank. The intended least one box is
sections of the model card. but are vague or generic uses and limitations blank. At least
without considering your sections are complete, one of the
specific data or survey but are vague or intended uses or
generic without limitations
considering your sections are
specific data or survey blank.

Creating an The app works without error The app works without The app works without App is not
App and includes a welcome error and includes a error, but does not complete or does
screen and a theme. The welcome screen and a contain a welcome not work
app incorporates theme. screen or theme - it
information from the model may be a
card into the app, such as single-screen app.
the accuracy, intended use,
or warnings.

Reflection Step 8 is completed with Step 8 is completed, but Step 5 is completed, Step 8 of the
detailed and specific some answers may be but most answers are project guide is
answers. You list specific vague or unclear. You vague or unclear or incomplete or
improvements you would were still able to test your you were not able to blank.
make to your data app with another person, test the app with
collection, model, and app. but your improvements another person.
may not be specific

7
8