worst part about shorts is that if you miss something then you have to re-watch the whole short.

This taught me more than i learned about neural network in my previous life as a computer scientist

This was the most impressive explanation of neural networks I've seen.

Some of the activation functions are : 1) Sigmoid function 2) ReLu 3) Leaky ReLu

Wonderful short depiction however all of the magic happens in the back propagation/error correction phase .. it occurs to me that while the architecture is clear, any reasonably sufficient representation of a NN must spend at least some time discussing the training.

I don't subscribe to many but very grateful I did subscribe to Assembly AI ! Such a great breakdown

It is actually hard to explain such complex concepts in 60 seconds. Whether the viewer will understand it or not, depends from the knowledge he comes with.

A simplified explanation but extremely good for beginners

Hey I just wanted to say a person literally built a neural network in Minecraft you should watch the video shorts don't support links but the video name is "I built a working neural network in Minecraft using redstone"

I taught this content for over 30 years. This short video does not clarify anything important about neural networks, nor does it do anything but confuse the viewer.

The best explanation of Neural Network. I have ever seen 😮 I am suprised ✨💯

Great vid! As always, the toughest talks are the shortest ones!

Well explained 👍🏼

Very good explanation for just 60 seconds! Another way to think about it is "it is just a mathematical function" and that is honestly scary to me. Are our brains just a gigantic mathematical function? A neural network is just a giant mathematical function that takes multiple inputs and outputs multiple values. Training a neural network is giving the giant function some inputs, comparing the output with the desired output and updating the constants in the function accordingly.

Oh hell yeah, love the music too

Great video

awesome explanation and background music 🤘 🤘

This is really impressive

Finally someone did it right. THANK YOU!

Nice work girlll

Thankyou! I do have a question though, how are the weights and biases chosen by the training algorithm?

Say you are seeing a ball. The visual input is one. The sound is an other. The shape is another. Touch is an other. Bias closer to circles. Rolling. Sound rolling. Etc.

Very nice

That was awesome. Instant sub.

thankyou ☺️ please explain more concepts in long form videos

Mindblown ❤

the last half video is quite good.

Amazingggggg.....thanku.....finally....understood....!!!👍🏻

Amazing explaation

saved my life !

Amazing explanation

Thank you

remember kids: always use reLU activation

U r greaT Mam Love From india keep sharing♥️✨

Stellar work! If you seek more, a book with like-minded content is what I'd propose. "From Bytes to Consciousness: A Comprehensive Guide to Artificial Intelligence" by Stuart Mills

Weights are optimised with a descent gradient, and the prediction is given by logistic regression, but how do you optimise the biases in a neural net?

Adding weights to inputs creates exponentially accentuates biases at the end results. So there is no escaping biases. Only a different type of bias

So a neural network has a designed bias? Designed inputs, as well as designed outputs. What’s not controlled about it? The middle layer but still starts off (controlled/biased) How’s it thinking? I’m dumb.

Mam you just looking like wow ❤❤❤

Merci 🙏

The use of x, y variable names is a bit misleading since x is often refered to as the input, while y, as output. Though if you are knowlegable about anns, it might not be a problem.

Watching this 15 minutes before exam 😭

How does one interpret the output?

WOW 😳 amazing 😍 explain

Is this a pathway to quantum neural computers? As in, one which is robust to quantum uncertainty(thus doesn't need many forward passes, or uses multiple averaged outputs at once to speed it up), handles superposition within the hidden layers, and benefits from the exponential parallelization.

Love from India 🇮🇳❤️

My sis explained this in 60 seconds while my teacher couldn't in a semester

Can you clarify how do you handle errors at the end to mitigate them on the next iteration?

Anyone link me something on how to do back propogation / good techniques

00:00

But why are hidden layers needed and how many of them?

0:00 for non-short form

To call this an "explanation" is like calling a firecracker a rocket that flies you to the moon.

I need more info❤😮😮

Its seems pretty easy to understand 😂😅😂😂😂

have you ever seen a dual output layer in line?

bios what for

So complex

Well explained. NN for Dummies.

What is the value of A ??

I don't know the application of this because i hve never seen a person applying this meanwhile i often see people saying that it's used un training AI

60 seconds is not enough to understand what have been said in 60 seconds 😂

im still confused [edit: i understand now}

Well yeah, but that's missing the point. It's a matrix-vector transform in latent space.

So do I integrate by parts? Or is there a trig identity??? 😂

All explanations of AÍ and neural network are very ambiguous and leave more image driven people behind. What is never answered is the nature of the neurons in a neural network. What are they in fact? Functions; concepts; math equations; transistors etc… like, in the brain is easy to picture the nature of neurons as a cell like structure exchanging neurotransmitters, but in the AÍ field it is harder to visualize what they rather than what they do. Maybe I should go ask chatgpt… 😅

The music, tone and speed of speech, and the timer at the top are not really helping :/

Get explanation!

Worst thing about this short is that you skipped a few steps...

красивая девушка в современности

😮

It's helpful but still not totally clear. It would be perfect if there were a simple example, with simple values

Or, in one second. it's complicated....

dada

xdd

I don't see how you have explained them if you don't discuss backpropagation.

She sound like Turkish

Could not understand anything. Is there any connection with real human body?

how learining the english

Challenge: now try the same in Human language. ;) its harder, right?

please dont include the time, because our brain will focus on the time and the discussion which is not good in your content.

Is this some sort of competition of explaining this concept in 60 sec?. Why do you need a timer?

I wonder what the purpose of such a video is.

Falls short You won’t get deep knowledge in 60 secs by no means sorry

The biggest issue is this is a dumb explanation

what is this obsession with background music!!!!!! we are watching a tutorial; not watching a hollywood fight scene. stop putting bgm everywhere.