Predicting numbers with Machine Learning for Kids

Hello, today we are going to teach machine learning for kids how to predict numbers, okay? 00:00:01

So the first thing we are going to use is the normal Scratch webpage. 00:00:09

We go to it, we google it, we can click on it, we don't need to log in, we can click on create. 00:00:15

And we are going to use a basket simulator, okay? 00:00:20

You already have the sb3 file, okay? 00:00:24

You can download it, and the moment you have downloaded it, you can load it from your computer. 00:00:28

Load it from your computer, and the file is called Basket, the SB3 file, okay? 00:00:33

So, this is a simulator. 00:00:38

Every time you press on the green flag, the ball moves to a random position. 00:00:40

Now it's minus 64, minus 55, and here you can decide the power and the direction of your shot. 00:00:44

So, let's consider that 529, I find. 00:00:52

If you click on the ball and you press S, you shot the ball. And it's a score. Fine, perfect. So you are going to, if you click on R, if you press on R, I mean, you will type these values. You will register these values. So I'm going to register them using the block. You can register them on a sheet of paper, but I think it's easier if you use the block. You will see why later. 00:00:56

x, y, power and direction. These are the four values you need to register. Direction. So x minus 64. You can see it here. Y minus 55. Every value must be separated by commas. Power 5, 29. 00:01:23

So every time you score, you must register the X position, the Y position, the initial ones, the power and direction. So if you click on the red button, again on the green flag, the ball changes its original position. 00:01:50

now is minus 215 and minus 40 so if I click on the ball very important click 00:02:10

on the ball and then press S now is now we are failing so we need to go back to 00:02:16

this position we press on air very very very important air don't press the red 00:02:23

button air we go back to the our X and Y initial position I think I need a more 00:02:28

powerful one let's try with seven I press on S no don't remember to click on 00:02:33

Now I press on S, and almost, R, don't press, and now instead of 7, I'm going to use 8, I 8, I click on it, I press S, and now I have a score. 00:02:38

this is a good position again, a good shot 00:02:52

I press on air, again I'm going to copy 00:02:56

I go to the block, minus 215 00:02:59

comma minus 40 00:03:05

comma 8 00:03:08

comma 29 00:03:11

to change the position, remember just to change the position, red button, green flag 00:03:14

And now to shot, if I press S, I click on it, let's try a little bit less powerful, maybe 4S, no, remember, don't forget to click on the ball, now S, again shot, so R, I press on R and I have minus 32 is the X initial position, minus 2 is the Y initial position, 4,29. 00:03:19

So you will repeat 10 times until you have 10 sets of different values that allow you to score. And the moment you have them, you will have to save it. For example, data.csv. Okay, this is the way you will save the information. 00:03:49

I already had one, so I'm going to upload it. 00:04:12

So, imagine you have ten different values, okay? 00:04:17

So, we have finished with Scratch. 00:04:20

Now, we go back to Machine Learning for Kids. 00:04:23

The exercise we did, the part we did with Scratch was just to find a group of values 00:04:28

that are the ones that allow us to score. 00:04:35

Now, we are going to use a new project. 00:04:39

We click on Añadir un nuevo proyecto. 00:04:40

Nombre del proyecto, basket, name of the project, basket, project type, predictive numbers, storage, your web browser. Remember that if you use the web browser, you are not allowed to open exactly this one in a different computer, okay? Just on this one. 00:04:42

So, create, crear. Okay, I already have my project. If I click on it, I need to train it, as always, and I need to add the values I already have. I can type them one by one, but it's easier if you type them this way. 00:05:01

If you use the upload CSV file. If you click on it, I will upload this one. This one is larger than the one I created before. So I have it. And I need to, I have all the values, but I need something else. I need to select the values I'm going to predict. 00:05:20

I want to predict power, and I want to predict two directions. 00:05:39

So, as you already know, I have two values that are going to be predicted, power and direction. 00:05:47

Okay, so the moment I have two of them, I can go back to the project, and I can train the model. 00:05:52

Entry a new model, or train the model. 00:06:00

I click on it, and the system will need some time to train a new one. 00:06:03

It's been really, really fast. 00:06:08

So, for instance, imagine I'm again in Scratch, and I click on the red button, green flag, and I'm here in this position. Minus 2, minus 66. I'm going to try to, I'm going to ask the Machine Learning for Kids what are the power and direction correct values to score. 00:06:09

minus 2, minus 86 are my values, minus 2, minus 86, minus 86, probar, try, and these are the solutions. 00:06:32

I will need to round, for example, I will use 5 and 27, so I go back here, 5 and 27, 27, 5 and 27, 27, now, I click on the ball and I press S, I fail because my model is not perfect, but I can correct it, I can, for instance, if the direction is a bit, 00:06:48

For instance, instead of 27, I can use 24. Let's try again. I click on the ball, and I press S. And this one is correct, okay? So I have, and now I have a new value that is correct. Minus 2, minus x is 6, 5, and 24. 00:07:24

The machine made a mistake, it happens from time to time, remember we have a short number of pieces of information, but it doesn't matter. 00:07:43

So, what else? I have my model and the exercise, your exercise will be about the creating part. You will create, click on create and you will go to scratch, open in scratch and here again you will open the file already, I have already given to you, archivo file, load from your computer and here you will 00:07:56

use the basket.sb3 file and here you will complete it, but the instructions to 00:08:26

complete it, you will read it in the instructions of your task, ok? 00:08:33