-
Launch
Machine learning model that can sort images
-
Natural Language Processing
Machine learning model that can recognise natural language commands
-
Recommendation Systems
Machine learning model that can recommend the reading age of a book based on data about the book
-
Decisions and Ethics
Presentation or report summarising key points
-
Machine Learning Algorithms
In this session we will be looking at some of the algorithms that make machine learning possible.
-
(Optional) Python and Orange
Data visualiastions using Orange Python code for importing data and running machine learning algorithms (decision trees and kNN)
[LAB] Building a model in Scratch3
ML2Scratch connects Machine Learning(TensorFlow.js) with Scratch.
If you take a few images with a webcam, label them, and learn them, you can classify similar new images based on the learning results. The captured images are not sent to the server, and all learning and classification are performed in the browser. (However, a network connection is required to load the application at startup and to download the learning model.)
Demo Movie
Learn/Classify Scratch stage images
Learn/Classify Scratch webcam images
Other samples
- Rock/Scissors/Paper Demo YouTube | .mov file
- Control a toy robot, MiP, by hand gestures YouTube | .mov file
You’ve started to train a computer to recognise pictures. But instead of trying to write rules to be able to do this (programming), you are doing it by collecting examples. These examples are being used to train a machine learning “model”.
This is called “supervised learning” because you are supervising the computer’s training. The computer will learn from patterns in the example photos you’ve chosen, such as the shapes and the use of colour. These patterns will then be used by the model to recognise new images.
Your teacher will use the examples you have collected to train the model to recognise cars and cups.
How accurately did your model sort the images?
If it made any mistakes have a look at the images it got wrong – is there anything similar about them? Do they have the same background or have they got the same colour on them.
Add some more images to your test data and rerun you model – have you improved the accuracy?
Examples of use
- Try to avoid obstacles with machine learning # ML2Scratch # ev3(Google Translated)
- Control Wagara-saurus(Japanese style dinosaur) using ML2Scratch
- Control an electric fan with illustration
- Smart Trash Box(Japanese)
- Making a coin sorting AI robot with Scratch and micro:bit
- Go forward with jasmine bottle, go backward with canned coffee (movie)
- ML2Scratch bookshelf arrangement check (movie)
- ML2Scratch detects parking space fullness (movie)