1 of 3

Artificial Intelligence for GA�Unit 2 - Understanding Language: Week #

Daily Assignments:

​

  1. Module 2.1 - Understanding Language
    1. Slides
    2. Module Activities to Complete
      1. Case Studies (Slide 12)
        1. Bullying a girl named Alexa- Parents of children called Alexa challenge Amazon
        2. Risk of some features - IS ALEXA SAFE FOR KIDS?
        3. Rude kids?/ privacy concerns? - Are smart speakers really safe for children?
        4. Conversation complexity/Jeopardising teaching professions? - Is Amazon Alexa harmful to children’s educational development?
        5. Wrong/incomplete answers, discourage critical thinking - Parents say using Alexa to entertain kids comes with problems
  2. Module 2.2 - How Computers Understand Language

​

​

2 of 3

Artificial Intelligence for GA�Unit 2 - Understanding Language: Week #

Daily Assignments:

​

  1. Module 2.3 - How Intelligent Assistants Understand & Answer
    1. Slides
    2. Module Activities to Complete
      1. What can Google Search Do?
      2. Dream Bot - Ticket Out the Door (Slide 3)
  2. Module 2.4 - Word Embeddings

​

​

​

3 of 3

Artificial Intelligence for GA�Unit 2 - Understanding Language: Week #

Daily Assignments:

  1. Module 2.7 - Chatbots
    1. Teachers Slides
    2. Students Slides
    3. Module Activities to Complete
      1. Creating Your Own Keyword Matching Chatbot
      2. Creating Your Own Template-Filling Chatbot
      3. Creating Your Own Intent Recognition Chatbot
      4. Chatbot with BERT Activity

Implement a simple chatbot that represents a character in a story and can answer questions about itself and the story. This activity uses the BERT language model, a transformer neural network that can understand text and answer questions.

  • Chatbot with BERT Activity Guide
  • Software: ML4K Scratch 3
    • Note: This is a separate instance from the traditional Scratch hosted by MIT
  • Optional: Pre-made demo Chatbot with BERT.sb3
      • . Intelligent Assistant with Keywords (Cognimates)

Create a simulated intelligent assistant in Scratch using Machine Learning for Kids (ML4K) that responds to voice input, modeled after Siri or Alexa. This is a simulation that outputs predetermined responses to selected keywords in the input; real assistants do much more.

​

3. Weekly Feedback

​