Scribd
Upload
35 views29 pages

AI FC Neural Network Presentation

This document provides instructions and information for a family AI challenge that teaches about neural networks through a hands-on design activity. The challenge involves using a neural network model to sort objects by asking simple yes or no questions, and encourages redesigning the network to improve accuracy. Families discuss true or false statements about AI, how it and people learn patterns, and how neural networks allow AI to learn. They learn about neurons and how connections between them represent concepts. The challenge is to create a network that classifies a target object within a mixture by breaking the problem into smaller pieces represented by nodes connected by questions.

Uploaded by

Nishant Kawa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
35 views29 pages

AI FC Neural Network Presentation

This document provides instructions and information for a family AI challenge that teaches about neural networks through a hands-on design activity. The challenge involves using a neural network model to sort objects by asking simple yes or no questions, and encourages redesigning the network to improve accuracy. Families discuss true or false statements about AI, how it and people learn patterns, and how neural networks allow AI to learn. They learn about neurons and how connections between them represent concepts. The challenge is to create a network that classifies a target object within a mixture by breaking the problem into smaller pieces represented by nodes connected by questions.

Uploaded by

Nishant Kawa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 29

Welcome to AI Family Challenge

#AiTogether
Build a Neural
Network
Welcome to Curiosity Machine!

Today we will:
Talk about AI and Engineering 15 min
Learn about the Design Challenge Inspiration 15 min
Plan, Build, Test 45 min
Redesign 15 min

Reflect and Showcase Prototypes 15 min


What have we learned?
Directions: As a family, discuss if these statements about AI are true or True or
false. false?
A computer that is using AI can learn like a human and be taught new
things
AI does not learn in similar ways to people
AI can learn to tell the difference between things

AI cannot be trained to learn from information

The more information that is given to AI, the more it can learn
How does AI (and people) learn?

Both people and AI learn by


understanding patterns

The more practice AI has with


patterns, the more accurate it can
become.
What is artificial intelligence?

"AI learns by
recognizing
patterns”
Let’s learn about how doctors use AI
AI can be many doctors all at once!

AI helps doctors because AI can learn like a


human

Doctors can teach AI to recognize diseases


and illnesses from pictures

These pictures can be things like X-Rays or


MRIs
Activity: Sorting by Features

It can be hard to sort things even for humans. How can you tell the
difference between these two snakes?

Coral Snake - Poisonous King Snake - Safe


Human brains have cells called neurons

Neurons can connect to each


other. When a lot of neurons are
connected, it’s called a neural
network.

In your brain, neurons can


represent concepts like lines or a
person’s face.
What are neural networks?

The more neurons, the stronger the connections between them


become.

Here’s an example: When you see a puppy, a neuron in your


brain might signal to make you think of happiness! The more
times this happens, the stronger the connection becomes. Next
time you see a puppy, you will feel happy!

These neural networks allow the human brain to learn new


things!
Neural networks can learn
better than just one neuron
on its own.
Artificial neural networks are what allows AI to learn
from and sort information.
AI uses neural networks to learn...

? by breaking problems up into little pieces and then


looking for matches

This is what we would see if we zoomed into


one small part of the face Sara
AI uses neural networks to learn...

Picture Identifying
of Sara Sara

AI technology creates questions to find patterns that can be used to


identify things
What do you think?

Is AI or a human faster at looking through images to find patterns?


Session 3 Challenge

Challenge Create a network that classifies information


Design Challenge Instructions

You will be given a mixture of objects (“data”) to


sort with the goal of sorting out one “target data”
piece.

You will break the problem into smaller pieces by


asking simple questions.

What questions will you ask to sort out the target


data?
Materials that can be used for the
design challenge.
Let’s Build!
Redesign Tips

If your network does not classify your objects correctly

1. Think about the order of the questions.

2. Think about whether a question sorts information into one or more


groups.

3. Create more nodes for sorting objects into more specific groups.
Reflect and Share

● Did the order of questions change how well your design was able
to identify your data?

● How would you change your network to be able to classify a


second target data?

● Would your network still work if you added twice as many


objects to be classified? What if you had half as many nodes?
Thank you!
We can’t wait to see your experience!

Please share any photos from your session and tag us:

@CuriosityMachine
Don’t forget to also
@curiositymachine tag: #AiTogether
@curious_machine
Appendix

Directions: Here are some optional slides you can use to


teach families.
Activity: Sorting by Features

4 limbs? Fur? Tail? Scales?


Dog Yes Yes Yes No
Lizard Yes No Yes Yes
Human Yes No No No

Directions: Use the information in the table to sort the below examples
● Example 1: Has 4 limbs, no fur, no tail, no scales
● Example 2: Has 4 limbs, no fur, a tail and scales
Activity: Sorting Materials

People and computers break


down a pattern or problem piece
by piece.

Directions: Sort the bag of


materials into 3 different groups
Follow Up: Sorting Materials

What were the groups you sorted the materials into?

What did the materials have in common that made you sort
them that way?
Example: AI uses questions to sort information

Object part Object A network of


Image broken down2D Shapes 3D Shapes(“tire”,
into lines (“wheel) “object”
“hubcap” ) neurons”
Instructional Video
Neural Network One Pager

Here is a page to learn more about the design challenge and


it’s connection to AI.

You might also like