Fraction Genius

Software: Machine Learning for Kids | Subject: Math | Grade: 3rd Grade , 2nd Grade | Domain: Artificial Intelligence | Topic: Fractions

Introduction

This lesson will introduce 3rd-grade learners to the fascinating world of fractions using Artificial Intelligence (AI) technology. By training a machine to recognize fraction images, learners will better understand fractions and how they can be used in everyday life. This lesson will help learners improve their math skills and introduce them to the exciting world of AI and machine learning.

 

Lesson Overview

   Overview  Activity Objectives
Opening Activity Learners revise fractions and equivalent fractions concepts through a game of Fraction Bingo. 
  • Understand fractions. 
  • Identify fractions in shapes.
 
Main Activity  Learners will train a machine learning model to recognize different types of fractions. 
  • Train a machine learning model to recognize fractions using a database of images. 
Closing Activity  Learners will test their machine learning model and reflect on the outcomes with their peers. 
  • Assess the machine learning model’s accuracy and develop ways to improve it. 

 

Prior Knowledge:

Learners should be able to:

  • Identify fractions given a pictorial representation
  • Understand equivalent fractions
     

Lesson Objectives:

Learners will:

  • Consolidate their knowledge of fractions
  • Revise concepts of equivalent fractions
  • Train and test a machine to recognize fractions represented pictorially

 

Learning Outcomes:

Learners will be able to: 

  • Understand and recognize fractions represented in shapes, numbers, and words. 
  • Demonstrate their understanding of fractions by playing a game of Fraction Bingo.
  • Implement their understanding of fractions to train a machine to recognize fractions of a particular shape. 
  • Test their AI fraction recognition model for accuracy and identify ways to improve it. 

 

 

Pre-lesson Prep

  • Like all lessons on Eddy, this lesson follows a certain approach. If this is your first time implementing an Eddy lesson, check out our lesson approach for more information.

 

 

Icon
What are some real-world applications of machine learning that are related to this activity?
- Self-driving Cars: Machine learning models are used in autonomous vehicles to recognize and classify images of pedestrians, other vehicles, road signs, and traffic signals. These models help vehicles to make real-time decisions and navigate safely on the road. - Image and Video Search: Machine learning models are used by search engines like Google and Bing to recognize and classify images and videos, making it easier for users to search and find relevant content. - Medical Diagnosis: Machine learning models are used in medical imaging to recognize and classify images of different organs and tissues to diagnose diseases and conditions. - E-commerce: Machine learning models are used in e-commerce websites to recognize and classify images of products, making it easier for users to find and buy what they are looking for. - Security and Surveillance: Machine learning models are used in security and surveillance systems to recognize and classify images of people, objects, and activities to detect potential threats and suspicious behavior.

Opening Activity

Slide Image

Slide 1

Lesson Introduction Slide.
Slide Image

Slide 2

Apprise learners of the objectives for the lesson.
Slide ImageSlide ImageSlide Image

Slide 3, 4, 5

 Introduce the ground rules, lesson norms, and team roles to students. You can check out our lesson approach if it’s your first time conducting an Eddy lesson.
Slide Image

Slide 6

Using the example of a pizza, introduce learners to fractions by sharing that fractions are a way to describe how much of a whole something is.

  • Ask learners to imagine that they have a whole pizza (6 slices in one whole). If they ate two slices of the pizza, what fraction of the whole pizza did they eat? (A: ⅓)
  • Build on this example to help learners visualize fractions (e.g., what if they ate three slices? 4?)
Slide Image

Slide 7

Ask learners how they arrived at the correct fraction. (A: The total number of pizza slices forms the denominator, and the number of slices eaten forms the numerator.)

 

Revise definitions of numerator and denominator.
Slide Image

Slide 8

Revise the concept of equivalent fractions with the help of the following example:

  • There are six learners in a class, and three are wearing blue shirts. What fraction of the class is wearing blue shirts? (Expect answers like: 3/6)
  • Would saying that half of the class is wearing blue shirts also be correct?
    • Expect answers to range from Yes and No. Address student misconceptions amongst those who say No.
  • Consolidate by sharing that equivalent fractions have the same value. Use other examples as appropriate.
Slide Image

Slide 9

          Share that learners will practice matching fraction visuals to numerical fractions while playing a fun game of Fraction Bingo. 
Slide ImageSlide Image

Slide 10, 11

Introduce learners to the rules of the game, listed on slide 10

 

NOTE: As a less logistics-intensive alternative to Fraction Bingo, teachers may check out these alternative activities.

 

 

Icon
How can I adjust the activity for learners who require more support?
Illustrate fractions & equivalent fractions using manipulatives.
Icon
How can I adjust the activity for my more advanced learners?
Get learners to come up with different scenarios representing different fractions.

Main Activity

Slide Image

Slide 12

 Introduce learners to the task. They must make a machine smart by training it to recognize fractions. 
Slide Image

Slide 13

Share that one way to help a machine recognize fractions is through machine learning.

 

Introduce the concept of machine learning.

  • Machine learning is a computer program that can help computers learn and make data-based decisions. 
  • It's like teaching a computer how to think, learn, and improve independently.
Slide Image

Slide 14

Ask them to state instances where they have witnessed machine learning outcomes. Apart from those stated on the slide, other examples of machine learning include:

  • Instagram/Tiktok filters
  • Music recommendations on Spotify or YouTube
Slide Image

Slide 15

Ask learners if they think machines recognize different types of fruits.

  • If learners respond with a “YES,” clarify that untrained machines cannot recognize fruits, but we can train them to do so.
 

Explain how machine learning works by using a worked example. At the start, a machine cannot recognize fruits or anything else, but we can train it by showing examples of what fruits look like.

  • At first, you might show the robot a picture of an apple and tell it that this is an apple. Then you provide images of different kinds of apples from different angles for the machine to understand and identify patterns. 
  • Then, you might show the robot a picture of a banana and tell it that this is a banana. Following a similar process for training the machine on many images of bananas would help the machine identify patterns.
  • The machine would look at the pictures and learn to recognize their differences.
Slide Image

Slide 16

Now, ask learners how they would train their machine to recognize fractions. 

  • Expect responses like providing the machine with fraction images to help it learn fractions.
  • To further probe learners, ask: what features or characteristics of fraction shapes are important for the machine to recognize?
  • How would you collect and label data to train the machine to recognize fraction shapes?
Slide Image

Slide 17

 Ask learners to brainstorm and finalize the fractions they want to train their machine to recognize. They will fill out the AI training prep worksheet with the details of their project.
Slide ImageSlide ImageSlide ImageSlide ImageSlide ImageSlide ImageSlide Image

Slide 18, 19, 20, 21, 22, 23, 24

Introduction to Machine Learning for Kids 

  • Get students to navigate on their devices to Machine Learning for Kids. 
  • Guide your students through creating a project and training the model.
 

If it’s your first time using Machine Learning for Kids, please refer to the tech tutorial here.

 

You can also refer to this video tutorial.
Slide ImageSlide ImageSlide ImageSlide ImageSlide Image

Slide 25, 26, 27, 28, 29

 Share tips on how to create an efficient fraction-recognizing machine.
Slide Image

Slide 30

 Play the timer on the screen for learners to be mindful of completing the task on time. 

 

 

Icon
What if students face problems navigating the platform?
Slides 19-28 in the teaching deck are to provide learners with guided practice on navigating the platform. If learners need more support in navigating the platform, you can direct them to this video.
Icon
How can I adjust the activity for learners who require more support?
Scaffold their learning by breaking down the concept of machine learning into discrete steps. Teachers can use relevant examples (e.g., video recommendations on YouTube) to illustrate how machine learning works.
Icon
How can I adjust the activity for my more advanced learners?
Ask students to ideate how YouTube makes video recommendations for individual users.

Closing Activity

Slide ImageSlide ImageSlide ImageSlide Image

Slide 31, 32, 33, 34

 Introduce learners to the method of testing their AI for accuracy.
Slide Image

Slide 35

Invite learners to test their model and complete page 2 on this worksheet in a group. Sample answers for two of the questions are given below:

 

Q. How would you improve the accuracy of your model even further? 

  • Add more training data (more images)
  • Use a more comprehensive data set (e.g., clearer images, blurry images, bright or dim lighting).
  • Supervised learning - use models which allow for feedback to be given to improve the accuracy of the model
 

Q. How would you use machine learning to improve your classroom or school? Give an example.

  • Detect if students are wearing a mask or not to prevent the spread of disease.
  • Face recognition - attendance taking
  • Convert speech to text for note-taking
Slide ImageSlide Image

Slide 36, 37

 Consolidate learning and invite students to share reflection questions.

 

 

Icon
What advice can I give to teams whose models are inaccurate? How can they improve their models?
If learners' models are inaccurate, here is some advice you can give them to improve their models: 1. Increase the amount of training data: If the model is not accurate, it may not have enough examples to learn from. Encourage learners to add more fractions images to the training dataset to help the model learn more effectively. 2. Improve the data quality: Sometimes, poor-quality data can lead to inaccurate models. Encourage learners to ensure that the images they use are clear and high-quality.
Icon
What is supervised learning?
Supervised learning is a type of machine learning where the algorithm is trained on a labeled dataset, meaning that the input data is paired with a corresponding output label or target variable. The goal of supervised learning is to learn a mapping function from the input to the output based on the labeled examples provided during training. In supervised learning, the labeled dataset is split into two parts: training and testing sets. The training set is used to train the algorithm to make predictions, while the testing set is used to evaluate the model’s performance on unseen data.
Icon
Can you take photos that the AI model classifies incorrectly? Why do you think the model classified these incorrectly?
Yes, it's entirely possible to take photos that the AI model classifies incorrectly! ML models are trained on lots of pictures so they can learn to recognize different objects like people, animals, and things. However, sometimes these models can make mistakes and classify pictures incorrectly. This can happen for several reasons, such as if the picture is too blurry or too dark, or if the model hasn't seen enough different kinds of pictures to know what to do. Sometimes, the model might also be biased and make mistakes more often for certain types of pictures, like pictures of people with darker skin.

Assessment

 

Understanding Fractions

Expand
 

Emerging

The learner can identify numerator and denominator but struggles to explain their meaning.
 

Developing

The learner can correctly identify and explain the meaning of the numerator and denominator.
 

Proficient

The learner can accurately identify and explain the meaning of the numerator and denominator and can provide examples.
 

Identifying Fractions in Shapes

Expand
 

Emerging

The learner can identify the correct fraction of the shape but struggles to shade it accurately.
 

Developing

The learner can correctly identify and shade in the fraction of the shape.
 

Proficient

The learner can accurately identify and shade in the fraction of the shape and can explain how they arrived at their answer.
 

Creating an AI Project

Expand
 

Emerging

More support from the teacher is required to create and train the learner’s AI machine.
 

Developing

Some support from the teacher is required to create and train the learner’s AI machine.
 

Proficient

No support from the teacher is required to create and train the learner’s AI machine. Learners can work independently.
 

Testing and evaluating the AI Model

Expand
 

Emerging

The learner can evaluate the AI model but struggles to identify areas for improvement.
 

Developing

The learner can evaluate the AI model and identify some areas for improvement.
 

Proficient

The learner can accurately evaluate the AI model, identify multiple areas for improvement, and propose ways to improve the model.

 

×