Machine Learning - AI4kids

Computers learn from data. Machine learning is the core technology of artificial intelligence. Through practical exercises, children can understand that machines learn through "data". There are three methods of machine learning: supervised learning, unsupervised learning, and reinforcement learning. The basic architecture of the neural network and the choice of training data affect the results of machine learning. Computers can learn from data.

Table of Contents

• Machine Learning Examples
  • Personal experience:
  • Industry Application:
• Key Concepts
• Indicators of students’ learning ability at different stages
  • Elementary school first and second grade:
  • Elementary school third to sixth grade:
  • Junior high school 1st to 3rd grade:
  • High school 1st to 3rd year:

Machine Learning Examples

Personal experience:

● Mobile phone photo albums recognize the images of contacts ● Youtube learns users’ preferences by clicking on videos

Industry Application:

● Training speech recognition systems ● Training translation systems: Google can translate between over 100 different languages ​​● Image search: training classifiers to recognize images of people, animals, vehicles, etc.

Key Concepts

● What is machine learning ● Methods of machine learning ● Types of machine learning algorithms ● Basics of neural networks

● Types of neural network architectures ● Training data and how it affects learning efficiency ● Limitations of machine learning

Indicators of students’ learning ability at different stages

Elementary school first and second grade:

● In an unplugged activity, find correlations between data. ● Use a classifier to recognize images. Use Google Autodraw or Cognimates Train Doodle to study how machines are trained and how they know what they are drawing

Elementary school third to sixth grade:

● Describe and compare three different machine learning approaches: supervised learning, unsupervised learning, and reinforcement learning ● Modify an open machine learning project ● Describe how algorithms and machine learning can exhibit bias

Junior high school 1st to 3rd grade:

● Identify bias in the training dataset and augment training data to address bias ● Manually simulate a simple neural network in training

High school 1st to 3rd year:

● Use TensorFlow to train a three-layer neural network ● Experiment and understand simple machine learning algorithms

For more information please Refer to the book published by AI4kids <Learning AI is Really Easy Series - Machine Learning>

Want to learn more? Click the button for more information!