Learning Pathway for AI Coding

[info-creaticode]

 

Introduction

 

The CreatiCode Playground provides a complete set of tools for creating AI-based projects, including games and apps. It offers a safe and engaging environment where K-12 students can learn and experiment with advanced AI technologies, such as ChatGPT and face tracking.

For beginners, there are step-by-step tutorials and projects that introduce the basics of AI coding. These resources are designed to help students build a solid foundation, making it easier for them to create enjoyable and practical AI projects.

Teachers can also use these tutorials to design introductory courses, enabling students to gain hands-on experience with AI coding and build confidence in their skills.

The first part of the tutorials focuses on AI tools available on the CreatiCode playground for generating images and coding assistants. The second part will focus on generative AI blocks (e.g. ChatGPT). The third part covers predictive AI blocks (e.g. classification models).

 
 
 
 
 

 
 
 
 

Part 1 - AI Tools in CreatiCode Playground

 

 
 

Topic 1 - Search or Generate Backdrop Images Using AI

 

• Tutorial Link: Search or Generate Backdrop Images Using AI

• Overview: Unlike the very limited backdrop library of MIT Scratch, CreatiCode offers a vast library of AI-generated backdrop images that you can browse or search. They are all carefully examined to ensure they are appropriate for K-12 school settings. If you can’t find any backdrop you like, you can also ask AI to generate a new backdrop for you with a detailed description.

 
 

Topic 2 - Search or Generate Sprite Images Using AI

 

• Tutorial Link: Search or Generate Sprite Images Using AI

• Overview: You can also choose from a rich collection of sprite images that are generated by AI and safe for use in schools. Each sprite image features one character or object over a transparent background. You can also generate new sprite images with your own description.

 

 
 

Editing Images

 
Instead of generating new images, you can also edit an image: select an image, then input a request on how to change that image.

• Tutorial Link: AI - Editing Images (Generate Variation)

• Example:

Request: put on a green t-shirt
Reference and Result Images:

 
 

Project Idea - Journey of a Waterdrop (Difficulty: 1)

 

• Tutorial Link: Journey of a Waterdrop (Difficulty: 1)
• Overview: Create a simple illustration of the lifecycle of a waterdrop using a few AI-generated images.

 
 

Topic 3 - Learn new things with CreatiCode XO’s help

 

• Tutorial Link: CreatiCode XO - Explain To Me
   

• Overview: CreatiCode XO is an AI assistant that you can chat with when you are in the CreatiCode playground. You can use XO to help you learn new concepts or blocks, with simple questions like “explain xyz to me” or “how to do xyz”.

   

 
 

Topic 4 - Debugging with CreatiCode XO’s help

 

• Tutorial Links:

  • How to debug
  • CreatiCode XO - Debugging
     

• Overview: Debugging is one of the most challenging task for most students, and very often our teachers do not have the time or expertise to help each student with debugging. Every student needs to learn the basics of debugging. XO is not designed to debug on behalf of the students. Instead, XO can assist students in the debugging process in diagnosing issues and brainstorming solutions.

   

 
 

Topic 5 - Design New Projects with CreatiCode XO’s help

 

• Tutorial Link: CreatiCode XO - New Project Design
   

• Overview: XO can help us get started with new projects. It can suggest project ideas when we “blank out” and need some inspirations. If you already know the project you would like to build, XO can suggest a plan on how to do that.

   

 
 

Topic 6 - Use XO to generate quizzes

 

• Tutorial Link: CreatiCode XO - Give me a quiz
   

• Overview: XO can be used to generate quiz questions in several formats, like short answers, multiple-choice or code completion. Students can use it for self-evaluation, and teachers can also use it to design assessment problems.

   

 
 

Topic 7 - Use XO to generate feedback

 

• Tutorial Link: CreatiCode XO - Automatic Feedback
   

• Overview: XO can be used to generate feedback on our project automatically. This allows students to get realtime feedback as they work on a project, and they don’t have to wait for teachers.

   

 
 
 
 
 

 
 
 
 

Part 2 - Programming with Generative AI

 

 
 

Topic 8 - Introduction to ChatGPT blocks

 

• Tutorial Links:

  • The Chat Window Widget
  • OpenAI ChatGPT
     

• Overview: Learn the basics of working with ChatGPT and also how to use the chat window widget to build a simple chat application. All ChatGPT requests are moderated to ensure they are appropriate for K-12 schools.

   

 
 

Project Idea - A Chat With Einstein (Difficulty: 3)

 

• Tutorial Link: A Chat With Einstein
• Overview: Create a simple chat app where the user can chat with ChatGPT, which pretends to be Einstein.

 
 

Project Idea - An Improved Chat App (Difficulty: 3)

 

• Tutorial Link: An Improved Chat App
• Overview: Improve the previous chat app with streaming response and make sure ChatGPT stays in the role assigned to it.

 
 

Topic 8 - Chat-based Apps

 

• Tutorial Link: Yes Minister!
• Overview: The improved chat above can be modified into many different types of chat-based apps, and the only thing that needs to be changed is the prompt for ChatGPT. This is very different from traditional app development, where each new app requires new logic and code. In this example, we can change it to a game where the player is the Prime Minister of UK, and the ChatGPT AI will ask the player to make some important decisions and rate the decisions at the end.

 
 

Project Idea - MBTI Personality Test (Difficulty: 3)

 

• Tutorial Link: MBTI Personality Test
• Overview: This another example of building a new app by changing the ChatGPT prompt. In this app, ChatGPT will ask the user a few questions, then determine the MBTI personality of the user.

 
 

Topic 9 - Simple Widgets: Label, Button, Textarea

 

• Tutorial Links:

  • Label
  • Button
  • Textarea

• Overview: learn to use these 3 most basic types of widgets, which are useful for building most apps.

 
 

Project Idea - An AI-based Story Writer (Difficulty: 3)

 

• Tutorial Link: An AI-based Story Writer
• Overview: ChatGPT has many other use cases beyond chatting. In this tutorial, you learn to use it to write a story based on user input.

 
 

Project Idea - A Quiz Writer (Difficulty: 3)

 

• Tutorial Link: A Quiz Writer
• Overview: In this project, we ask ChatGPT to generate a multi-choice quiz question based on user request, then evaluate the user’s answer.

 
 

Project Idea - Cloze Game (Difficulty: 4)

 

• Tutorial Link: Cloze Game (Difficulty: 4)
• Overview: In this project, we ask ChatGPT to create a cloze game, which require some more advanced prompting techniques.

 
 

Topic 10 - Voice-based Apps

 

• Tutorial Links:
  • AI for Converting Speech to Text
  • AI for Converting Text to Speech
  • Talk to an AI Sprite (Difficulty 2)
  • A Voice Translator (Difficulty: 2)
• Overview: It is easier for many users to talk than typing on a keyboard, so we can use speech recognition to convert what a user says to text and send that to ChatGPT. When we get ChatGPT’s response, we can use text-to-speech to convert it back to audio.

 
 

Topic 11 - Vision-based Apps

 

• Tutorial Link: ChatGPT AI - An assistant that can see and talk (Difficulty: 2)
  
• Overview: Large language models like ChatGPT are generally “multi-modal,” which means they can not only chat with us in words but also understand images. This is very useful since, in many cases, it is much easier to describe what we need with a picture than words. In this tutorial, you will build an AI assistant that can see and talk. To use it, the user just needs to take a picture using the camera and then ask the AI assistant any questions about that picture.

 
 

Topic 12 - AI-driven Games

 

• Tutorial Link: Who’s the Spy?
• Overview: ChatGPT can also be used as the “reasoning engine” to create interesting games that weren’t possible before. In this tutorial, you will build a game where the player needs to interogate a few characters to find out who is the spy, and all responses are generated using ChatGPT.

 
 

Topic 13 - Using ChatGPT as text processing tool

 
 

• Tutorial Link: Product Review Summary
• Overview: ChatGPT is very good at processing large amounts of text. In this tutorial, you will learn to use it to process many product reviews and summarize them.

 
 

Topic 14 - Teach ChatGPT to use tools

 

• Tutorial Link: Enhance ChatGPT to Do Math Calculations (Difficulty: 3)
• Overview: ChatGPT has limitations in many tasks, such as math calculation. In this tutorial, you will teach ChatGPT for “tool use”: you will provide a calculator tool for ChatGPT, and you will instruct ChatGPT to use this tool whenever it needs to do calculations.

 
 

Topic 15 - Using 2 Chatbots At the Same Time

 

• Tutorial Link: AI - 2 Chatbots Debating Each Other (Difficulty: 4)
• Overview: Sometimes you need to use more than one chatbots in your project, so that each chatbot will take on a different role or work on a different task. In this tutorial, you will learn to make 2 chatbots debate each other on any topic the user specifies.

 
 

Topic 16 - Enhance ChatGPT with Semantic Search

 

• Tutorial Links:
  • ChatGPT AI: QA Bot Using Semantic Search (RAG) (Difficulty: 4)
     
• Overview: ChatGPT has limited knowledge, so when we create a QA chatbot, we need to feed it additional knowledge based on the user’s question. This can be achieved using “semantic search”, which allows us to search for text based on meaning rather than matching words.

 
 

Topic 17 - Use ChatGPT to create knowledge base from web

 

• Tutorial Link: Prepare Knowledge Data Using Web Content (Difficulty: 4)
• Overview: In this tutorial, you will learn to prepare the knowledge data for a new chatbot, which will answer questions about any organization based on its web site. The basic idea is to download data from its website and generate questions/answers using ChatGPT based on that data.

 
 
 
 
 

 
 
 
 

Part 3 - Programming with Predictive AI

 

Topic 18 - AI Motion Sensor

 

• Tutorial Links:

  • Video Sensing
  • Bouncing Ball with Motion Sensor (Difficulty: 3)

• Overview: The video sensing blocks allow us to detect the amount of object movement in the camera video. This information can be used to create simple games where the player’s body controls sprites in the game.

 
 

Topic 19 - Hand Tracking

 

• Tutorial Link: Counting Fingers with Hand Detection (Difficulty 3)

• Overview: AI models allow us to track human hands in a camera video in realtime and detect the keypoints of the hand/fingers. In this tutorial, you will build a small project that counts how many fingers are stretching out.

 
 

 
 

Topic 20 - Body Pose Tracking

 

• Tutorial Links:

  • Body pose detection
  • Fitness Game Using Body Pose Detection (Difficulty: 3)

• Overview: Motion sensing only gives us very limited information on the amount of motion in the camera video. To track the exact pose of the player in the camera video in realtime, we can use the body pose detection block. That allows us to build games where the body pose of the player is used to control sprites in the game.

 
 

Topic 21 - Beginner Level Image Analysis

 

• Tutorial Link: Recognize 0’s and 1’s (Difficulty: 4)

• Overview: Image analysis requires walking through all the pixels (components) in an image and analyze them. In this tutorial, you will build a smart program that can recognize whether a hand-written number is 0 or 1, which shows the basic methods of image analysis.

 
 

Topic 22 - Data Classification

 

• Tutorial Links:

  • The KNN (K-Nearest-Neighbor) Classifier
  • KNN Classifier for Diabetes (Difficulty: 4)

• Overview: AI classifiers allow us to classify data into different categories. We need to first train such a classifier model using training data, then we can use it to classify new data. In this tutorial, we will build a classifier that predicts whether a person has diabetes, using a KNN classifier model.

 
 

Topic 23 - Neural Network Models

 

• Tutorial Link: Use Neural Network Model for Training and Prediction (Difficulty: 5)

• Overview: A Neural Network model is made of many “layers” of “neurons”. Each neuron is a small calculator that takes some input values and calculates an output value based on the input values. We can train these neurons based on some training data, and then use the model to make predictions on new data. In this tutorial, we will train a simple neural network model to do a simple calculation to illustrate the basic workflow.