Training AI expressions – SCOPES-DF

Lesson Details

Subjects
Age Ranges
Standards
Fab-Modeling.1
Author
Additional Contributors

You need to login or register to bookmark/favorite this content.

Author

Moath Momani
Moath Momani
K-12 teacher
I am a passionate educational technology engineer experienced in developing computer training materials and teaching how to use computer applications, machines, and cutting edge technologies. I obtained my bachelor’s degree in mechanical engineering in August 2016, then I started my… Read More

Summary

This lesson is derived from MIT personal image classifier and it has been adjusted to be given online.

https://appinventor.mit.edu/explore/resources/ai/personal-image-classifier

 

In this lesson participant will be introduced to the basics of machine learning, then they will train an AI model using web based tool to classify their emotional expressions ( Ex: Sad, Surprised and Happy)

What You'll Need

participants will need the following:

1- Laptops (they will use the laptop’s camera).

2- Internet connection.

3- Zoom or MS Teams

Lesson Materials

The Instructions

introduction to Microsoft Teams or Zoom - 3 minutes

Introduce participants on how to share their screens, raise their hands to ask questions, and use the chat.

Ice-breaker - 3 minutes

Ice Breaking Activity As you are teaching new participants, start to interact with them, and break the ice so they won't feel shy and focus on the lesson.

tell them “Here you have 3 different pictures of different persons, ask participants to describe the emotional expression of each person in one word?” 

 

Each one of them will describe each picture in one word, from this process we can identify the meaning of image classification.

Ask the participants:

1- Can a computer recognize the photos and emotions like us?

 

2- Is this hard or easy for a machine? 

Machine learning - 10 minutes

introduce and explain machine learning

Ask participants what do you understand from the below picture ?

 

 

tell them

” the AI Robot is detecting trash and remove it to clean the ocean, but how does this Ai robot know to differentiate between fish and trash!!

Actually, the AI robot has a certain technology called machine learning, and this technology enables the robot to learn from us by showing him pictures of fish and pictures of trash, so the Ai robot learns from us how to separate between things without being programmed”

 

Ask them to think

 

“What will happen if you give the wrong training for the robot, for example showing the purple fish for the robot and telling him this is a trash??”



Absolutely he will do the same mistake that you did, he will classify the purple fish as trash and start cleaning the oceans from all purple fish, because the behavior of the Ai robot based in your bias.

 

 

Ask them to watched the below video

Say “Machine learning is a subset of Artificial Intelligence, Let’s watch this video to learn more about it.”

https://www.youtube.com/watch?v=OeU5m6vRyCk

 

After they watched the video ask them

 

Examples from the video include:

•email filters

•auto-complete text

•video recommendation systems

•voice recognition

  1. •translation apps

•digital assistants

•image recognition

Train the model with Personal Image Classifier - PIC - 30 minutes

in this section, participants will be asked to use a web-based toot called PIC to train an AI model to classify their emotional expressions. Here there are two important points to take into consideration: 1- that PIC a web-based tool and it doesn't work on the MS edge browser so ask them to use google chrome browser through the lesson. 2- While participants start with the PIC tool, they will have to use the camera and some of them may report to you that the camera is not working, this is because the camera is already working in MS teams or Zoom, so they have to turn it off in teams of zoom.

Go to this link

https://classifier.appinventor.mit.edu/

and introduce participants how to train the AI model.

 

https://www.youtube.com/watch?v=IoMX_qlGFUM whatch this video to learn how to train the model

below are some screenshots during the lesson were participants trained the model.