ML - Worksheet: Alien Language
ML - Worksheet: Alien Language
This project worksheet is licensed under a Creative Commons Attribution Non-Commercial Share-Alike License
http://creativecommons.org/licenses/by-nc-sa/4.0/
1. Go to https://scratch.machinelearningforkids.co.uk/
3. Using the new Speech to Text blocks, create the following scripts.
Next, you’ll train a machine learning model for yourself to see how it was
done.
For the next part of the project, you’ll use your voice to control an alien
character that doesn’t understand English! You’ll invent two new words,
that wouldn’t be found in an English dictionary, to control your character
and train a machine learning model to recognize them.
6. Go to https://machinelearningforkids.co.uk/
11. Name your project “Alien Language” and set it to learn how to
recognise “sounds”.
Click the “Create” button
14. Click on the Add example button in the background noise bucket
Recording background noise will help your machine learning model to tell
the difference between the sounds you will train it to recognize, and the
background noise where you are.
17. Repeat that until you have at least 8 examples of background noise
18. Click the Add new label button in the top right, and create a new
training bucket called “left”
21. Click the Add new label button in the top right, and create a new
training bucket called “right”
27. If you’re not happy with how the model is working, go back to the
Train page and add more examples to all three training buckets.
28. When you’re happy with your machine learning model, click on the
Make button
29. Click on the Scratch 3 button and then click Open in Scratch 3
30. Click on the Project templates button at the top of the screen and
open the “Alien Language” project template
Unlike the pre-trained model you used before, which has been trained to
recognize tens of thousands of words, you’ve only trained it to recognize
two different words. But the principle is the same.
You’ve also seen the importance of training the machine learning model
to work with a certain background noise.
Can you think of an example of a system like this you’ve seen used
before? For example, some cars use speech recognition systems that have
been trained to recognize the different commands you can give to the in-
car computer. What other examples have you used?
Now that you’ve finished, why not give one of these ideas a try?
Try adding two more training buckets for “up” and “down” so
you can control the alien to move in all four directions.