I definitely think that this transfer of "how stuff that feels super complex on first glance directly translates to code" is SUPER valuable for me at least. Thanks!
Positive and negative in electrochemical charges(signal) is relative to the charge value of the opposite charge +1.45 v is "negative" compared to +1.5 v. Also the weight of a prediction is compared to the threshold that a receptor(type of neuron) fires a signal at, instead of pulling your hand fast when it's just a breeze stimulating a receptor rather than fire. I ❤ you man, best introduction to Nn in history.
I concur that - I was a little disappointed when after 15 minutes of a great explanation of concept the video just concluded that "some magic happens here". The followup would be much appreciated. 👍
You're right, it would be great to cover the training part as well. Ideally, I'll have some time and retake this video - I feel the second part is underwhelming and doesn't really lead anywhere. The problem is that training is such a juicy topic, and complicates the first part. Welp! I'll try to figure it out. In any case, one of my goals was to make people curious, which I guess succeeded! :)
Bro it is very interesting you showing cool ways of using dart. Thank you. I need more of your videos, your explanation is great. I see you as my teacher.
I agree with many of the other comments. By far the best explanation of how a neural network operates and such a great example to show it in code, but by the end I was left dangling because I just wanted to play, try out some data and some different activation functions, but can't work out what NetworkEvaluator and train() are doing so cannot proceed. Any updates on the next video and when its coming, there are a few eagerly waiting ... please don't stop now!!!
Amazing video! I was wondering how to code a neural network, but there was nothing good online. I followed along with your code in java and it works well. It would be awesome if you could make a video training the neural network.
Thank you very much Filip for this awesome video ! It's so great to have an explanation of how something so "magic" works under the hood through simple cases ! I learned a lot ! Do you plan on making another video about the training part ? I'd love to see how to implement the training algorithm to transform a training set into a set of well defined parameters for my network 😅
Mannnnnnn! what kind of research are you doing before creating videos. I was working on something like this but using javascript and i got distracted mimicing the neuron details/composition rather than keeping it simple as this. I love this content even if it might be out-dated for the current world, it still gives you the direction in which you need to think and put your efforts in. Love the way you explained it👍👍👌👌. This is more than giving access to Chat-GPT 3 to me! 😅
Great video Filip! In college I worked on a Tetris game with genetic programming to create the best AI for the game! I couldn't ever get it working properly but this video explains some things I had issues with! Maybe one day I'll go back and finish it haha 😉
Funny enough, yesterday before heading to bed I just felt like checking if you've released a new video; found nothing but now that this is here, I'm pretty sure it was still being uploaded at the time i checked😅
I'm actually using a genetic algorithm in this case, so that's quite similar. But backpropagation would make sense in the particular case of XOR or AND learning. And it would not be that much harder. Problem is: not sure when I have time to record the video that explains this, as it's a lot harder to explain in simple OOP terms (imho).
In a biological system, the equivalent of the negative weight sum of the connection between neurons is the reaction of neurotransmitters from the presynaptic ending with receptors at the postsynaptic ending, which reduces the probability of impulse initialization. And positive - increase. The most abundant inhibitory neurotransmitter in the brain is γ-Aminobutyric acid (GABA). Am i right?
Hi) I really like the way you make the video, very much like you are familiar with the theme:-) Is it possible to ask you to help become a Flutter developer? i Senior Python backend developer, and would like to learn a new direction faster) If you agree, I would be very happy, it would be enough to have some tasks that would help me to dial the right skills
I definitely think that this transfer of "how stuff that feels super complex on first glance directly translates to code" is SUPER valuable for me at least. Thanks!
Please make more videos on ML 🙏
Positive and negative in electrochemical charges(signal) is relative to the charge value of the opposite charge +1.45 v is "negative" compared to +1.5 v.
Also the weight of a prediction is compared to the threshold that a receptor(type of neuron) fires a signal at, instead of pulling your hand fast when it's just a breeze stimulating a receptor rather than fire.
I ❤ you man, best introduction to Nn in history.
What I enjoy the most is you showcasing how good and flexible the Dart language is!
Great video, thank you! Will you do a further video about training a neural network in detail?
I hope to. And thanks!
We hope you do it too. Thanks in advance 👍👍
I concur that - I was a little disappointed when after 15 minutes of a great explanation of concept the video just concluded that "some magic happens here". The followup would be much appreciated. 👍
this knowledge changed everything for me, this is pure fire 🔥 thanks Filip!
It would be interesting to see a video about how hard it actually is to train your network. Like the same basic code, but for training.
You're right, it would be great to cover the training part as well. Ideally, I'll have some time and retake this video - I feel the second part is underwhelming and doesn't really lead anywhere. The problem is that training is such a juicy topic, and complicates the first part. Welp! I'll try to figure it out. In any case, one of my goals was to make people curious, which I guess succeeded! :)
@@filiphracek you did succeed Filip! I'm super curious about it and would love to see a new video covering the topic. Thank you!
I'd love to see a follow up video on training too! Although I know the theory of NNs, but seeing the actual code and flow is extremely helpful
Bro it is very interesting you showing cool ways of using dart. Thank you. I need more of your videos, your explanation is great. I see you as my teacher.
I enjoy all Your work Filip
I agree with many of the other comments. By far the best explanation of how a neural network operates and such a great example to show it in code, but by the end I was left dangling because I just wanted to play, try out some data and some different activation functions, but can't work out what NetworkEvaluator and train() are doing so cannot proceed. Any updates on the next video and when its coming, there are a few eagerly waiting ... please don't stop now!!!
PERFECT, I was just looking for who Tweeted about this, ty
Very good video! I would like to see more of this!
The kind of content I need! Thanks
Thanks for your great videos Filip and i wish you a happy new year 2023! 🍀
Amazing video! I was wondering how to code a neural network, but there was nothing good online. I followed along with your code in java and it works well. It would be awesome if you could make a video training the neural network.
Thank you very much Filip for this awesome video ! It's so great to have an explanation of how something so "magic" works under the hood through simple cases ! I learned a lot !
Do you plan on making another video about the training part ? I'd love to see how to implement the training algorithm to transform a training set into a set of well defined parameters for my network 😅
interesting thing with dart
Thanks for this explanations Philip !
So useful, waiting for a whole "Game AI with dart" playlist.
Hi Filip, I never imagined Neural network can be this easy, can you please (with some extra sweet) make a video on the concept how to train part.
awesome video mate
Awesome video
thank you for this great video
Mannnnnnn! what kind of research are you doing before creating videos. I was working on something like this but using javascript and i got distracted mimicing the neuron details/composition rather than keeping it simple as this. I love this content even if it might be out-dated for the current world, it still gives you the direction in which you need to think and put your efforts in. Love the way you explained it👍👍👌👌. This is more than giving access to Chat-GPT 3 to me! 😅
Thank you so much!
NIce !!!! I would like to see more videos about it.
Thanks for the great explanation!
Great video Filip! In college I worked on a Tetris game with genetic programming to create the best AI for the game! I couldn't ever get it working properly but this video explains some things I had issues with! Maybe one day I'll go back and finish it haha 😉
Can you make more videos about this topic
great video.. thanks
Can you also show how you train the network?
Sure, I'll do it either in the next video, or I'll retake this one, make it snappier, and include training. Thanks for the suggestion!
You are great
can you come up with the following video for training the neural network in the same flow?? please.. i liked the simple thought process alot!
What a valueable video thanks 🙏
Very very nice Fillip! The best explanation about all this A.I. stuff. ))) Will you use this network in your game?
Yes I will! It's going to be super helpful for low-level API, such as obstacle allowance and aiming.
I hate you, for not uploading this 2 months ago when I needed. (jk amazing work).
i need more of this video because i want dart to be the main code language for ai not only python which is made by ai
Funny enough, yesterday before heading to bed I just felt like checking if you've released a new video; found nothing but now that this is here, I'm pretty sure it was still being uploaded at the time i checked😅
Wow, thank you so much. Can you please share the training code? :)
Thanks for this, but the hardest thing is probably the functions that trains it, probably you can't do it just by simulated annealing?
I'm actually using a genetic algorithm in this case, so that's quite similar. But backpropagation would make sense in the particular case of XOR or AND learning. And it would not be that much harder. Problem is: not sure when I have time to record the video that explains this, as it's a lot harder to explain in simple OOP terms (imho).
great. continue more about game dev and flutter.
i can use full stack ml with flutter without python and django ?
Does anyone know a(n actively maintained) dart library for all that?... training, gradfient descent, back propagation...?
can you share the source code please
Can darts be used for even more ai? like deepfake / speech to text, text to speech, pure dart face recognition without additional libraries?
In a biological system, the equivalent of the negative weight sum of the connection between neurons is the reaction of neurotransmitters from the presynaptic ending with receptors at the postsynaptic ending, which reduces the probability of impulse initialization. And positive - increase. The most abundant inhibitory neurotransmitter in the brain is γ-Aminobutyric acid (GABA). Am i right?
I have no idea! Now I kind of want to spend the holidays reading up on biological neurons, to be honest. :)
Hi)
I really like the way you make the video, very much like you are familiar with the theme:-)
Is it possible to ask you to help become a Flutter developer?
i Senior Python backend developer, and would like to learn a new direction faster)
If you agree, I would be very happy, it would be enough to have some tasks that would help me to dial the right skills