I started a Neural Network Tutorial series to show how I made the neural network used in this video! Here is a link to part 1: ruclips.net/video/JXGogdI7RIE/видео.html
SUPER COOL!! massive fan of neural networks and the math behind them. what would’ve been great is if you introduced the cars to a new map to see if they were generalised and not just remembering the course!!
Absolutely fascinating video! But I’m left wondering what happens if you take the winning car and place it on a different track. Or the same track but in the opposite direction. Basically does it know how to drive, or does it know how to effectively not hit stuff on that one track?
I wan't sure if people would know it but I put it in there for me. His storytelling skills are insane. I watched so many of his videos without even playing the game
How do you use backpropagation for a system with no obvious direction of improvement? My current understanding of backpropagation only allows it to be applied to situations where you already know the best answer for each of the questions in a training set, so you know which direction the output needs to move in order to improve the network's performance - e.g. the network said 2, the correct answer was 3, so you need to increase the network's output, so you backpropagate through the system to determine how all the weights and biases need to change in order to make the output closer to three. But with driving, it is not always obvious what affect a change to the weights will have on the performance of the system, since you don't necessarily know beforehand whether driving straight or a small adjustment to the left is optimal in a given situation (or if you do there's probably not much point in training a neural network - just hard-code it). How do you resolve this issue?
You're video is awsome I'm gonna subscribe and I'm really interested to see theses IA pushed to the limits like a race track without walls or barrier+ full speed and turn Also love the mathematical explanation
Yeah in this video I did make the ray casts go through the checkpoint otherwise they would try to avoid the checkpoint thinking it’s a wall (unless you adjust the inputs so it know the difference between a wall ray and a checkpoint ray)
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
I started a Neural Network Tutorial series to show how I made the neural network used in this video! Here is a link to part 1: ruclips.net/video/JXGogdI7RIE/видео.html
Super under rated channel !! keep up the good work and don't forget - consistency is key !
Thank you! Will do! Still got to work on that consistency thing
850th sub over here 🙋🏻♂️ so many hundreds of thousands to come
I thought this was a channel with a couple hundred thousand subs for a second. Surprised you don't have more.
SUPER COOL!! massive fan of neural networks and the math behind them. what would’ve been great is if you introduced the cars to a new map to see if they were generalised and not just remembering the course!!
ayo i was 69th view too nice👌
Yeah that would be awesome see if they really learn from theirs mistake
this and your other Neural network video is sick! trying to do something like this myself, so looking forward to the Unity implementation tutorial.
This is what I have been looking for!! Thank you so much
Absolutely fascinating video! But I’m left wondering what happens if you take the winning car and place it on a different track. Or the same track but in the opposite direction.
Basically does it know how to drive, or does it know how to effectively not hit stuff on that one track?
Agree with this sentiment. Feels like this might have been more of a situation where they are overtrained on the data.
Yeh, I agree. As long as the network was big enough, it'd probably be somewhat easy to train it on new tracks for it to become more robust
5:14 love the Wirtual reference
I wan't sure if people would know it but I put it in there for me. His storytelling skills are insane. I watched so many of his videos without even playing the game
@@JohnnyCodes ah I see. Your attention to storytelling shows
Great video, more like this and this channel will grow big. Keep it up
Would've loved to see a code review, if you can make a second part, it would be awesome.
It has been a while but I just started a new series for the neural network tutorial! ruclips.net/video/JXGogdI7RIE/видео.html
@@JohnnyCodes Oh my goodness! Thank you so much!
Using a genetic algorithm instead of backpropagation in a neural network seemed unusual, but it made for a great video and results in the end!
unusual because it is not main stream ? 🤣
No, because it doesn't work for bigger networks.
@@VladimirPanteleev but do you really need monolithic large networks ?
@@dm3on Yes, for many classes of problems.
How do you use backpropagation for a system with no obvious direction of improvement?
My current understanding of backpropagation only allows it to be applied to situations where you already know the best answer for each of the questions in a training set, so you know which direction the output needs to move in order to improve the network's performance - e.g. the network said 2, the correct answer was 3, so you need to increase the network's output, so you backpropagate through the system to determine how all the weights and biases need to change in order to make the output closer to three.
But with driving, it is not always obvious what affect a change to the weights will have on the performance of the system, since you don't necessarily know beforehand whether driving straight or a small adjustment to the left is optimal in a given situation (or if you do there's probably not much point in training a neural network - just hard-code it). How do you resolve this issue?
This was very informative and an awesome video, thank you!
Glad you enjoyed it!
Cool video
I wonder how the fully trained ones would perform if you changes the map.
you are awesome, continue like this
Include collision and damage, we are set for a race!
Wow, they figured out the "line".
You're video is awsome I'm gonna subscribe and I'm really interested to see theses IA pushed to the limits like a race track without walls or barrier+ full speed and turn
Also love the mathematical explanation
Can you make them solid objects
How do you use SplineMesh to create a track?
very nice!
Is it possible to share the project?
Awesome video. Is there any way you would post up the project?
did you make the ray cast pass through the checkpoint and only hit the walls or does it hit the check point?
Yeah in this video I did make the ray casts go through the checkpoint otherwise they would try to avoid the checkpoint thinking it’s a wall (unless you adjust the inputs so it know the difference between a wall ray and a checkpoint ray)
@@JohnnyCodes okay thanks !
957th subscriber
I get that I’m probably an idiot but how the fuck do you reward a computer