?? Work at Pickup Points � Flexible schedule

But what is a neural network? | Deep learning chapter 1

3Blue1Brown 18 views

Download

Download Video

Comments

Thats the best explanation I have seen
Comment from : @5max62

Nice 👍
Comment from : @SurajprasadVerma-y6v

On 15:02 should be b0 to bk not bn
Comment from : @HCS-e6n

15:16 - code snippet of how feed forward might look like in Python
Comment from : @disouzam_bh

just amazing
Comment from : @deepakdas4513

Salute to you for your simplicity, deep explanations, highly suitable and visualizing animations, lessons structure, proper wording and great teaching language I can't explain the addition you does in the way of deep learning Very much confident and clear to the knowledge in Deep learning and Maths associated with that
Comment from : @akashvarsani

best video ever
Comment from : @Manojkumarashok

FYI, our instructor at IISER-Mohali officially recommends watching your videos as part of the coursework on Machine Learning
Comment from : @sharatsa6008

God created universe and everything, blesses people in worldbrOur vision is a miracle of God, not an evolution processbrMilitary classified of neural networks are hard, soft work is like Las Vegas
Comment from : @張洪鈞

3Blue1Brown’s explanation of neural networks isn’t just teaching—it’s pure intuition brought to life, making even the hardest concepts feel natural Huge respect for that! 🙌
Comment from : @MuhammedShaheb

Hi, I know near zero about ML, algorithms etc brOnly thing that worries me, is that IF the humans that makes algorithms and provide datasets, have mental heath issues, we are going to produce a crazy machine that learns to be even more crazy
Comment from : @lucianogiudice8569

Such superbly done video
Comment from : @sheetaleinstein

Que bien explicación ❤
Comment from : @Olehop123

"Якщо мигдаль став молоком " чудовий вислів
Comment from : @valentynkostiuk

I love when you expect a neural network to learn the way humans do, but it ends up learning in ways we would never think about
Comment from : @jesusgracia4113

great, simple clear many thanks from Chile
Comment from : @claudioacuna7775

10:04 這邊不太懂，如果圖片中暗像素代表0(第一層的圖片像素Activation範圍是0-1)，即使給予負權重，0乘上任何數依然是0，為何可以得到最大的加權總合呢？brbr然後我也很好奇層數跟每層需要有多少個神經元是如何決定的，希望之後章節會介紹到br(影片為了示範需要設定為兩層跟每層16個)
Comment from : @after-8-years

Omagad 😅, actually i didn't understand but thank you for adding my voice in the end (russian dub) ❤
Comment from : @microerne

At 14:48, shouldn’t the bias matrix be from b0 to bk instead of b0 to bn? Just confused a little
Comment from : @meghanabellamkonda1443

i love how my real life eye is 3 brown 1 blue, just reverse of you ahahahhaa
Comment from : @akisa452

Thank you
Comment from : @tylercorderman5386

Saw this video on your Hindi channel, and now you've added here too Great ❤️❤️
Comment from : @Jazz-j1m

great thank you so much
Comment from : @franzjoseftheissen5450

The fact that I could watch this amazing video in Hindi 😭😭 Thanks
Comment from : @SANTOSHKUMAR-zo4ef

You probably won't see this, but if you do, can you please make a series on LSTM models, how BPTT works and how Adam optimizers help? I feel like these are skimmed over in most courses, but in reality, it is probably the most realistic representation of a human brain there is in my opinion, and having a deep understanding of these could be an integral part of understanding just how an AI actually thinks
Comment from : @AbhradeepDe

Can you make a video on activation functions
Comment from : @DanielAnani-s5y

Thank you
Comment from : @Aourioon

Why didnt I find this chanel 7 years ago
Comment from : @wingsoffreedomm

Голос уе ищный
Comment from : @prochti-puranulinga5765

1233 - It should be 16 * 16 right?
Comment from : @dataviz1135

Hail @3blue1brown 🙇
Comment from : @DheerajNarlajarla

thanks bro you give me whole explaination on neural network
Comment from : @NAKULBHAGWANDAFALE

WOW WOW WOW'
Comment from : @tiktok12345

You will never be able to teach that AI how cute pi creatures are
Comment from : @DoFliesCallUsWalks

Three words , " You're the best" !!!!:)
Comment from : @Edward_Kenway_

I think im too stupid to understand this 😔 Do i need any prior knowledge of this to be able to understand it?
Comment from : @RobinBli

This reminds me what i did in my 1st neural netowrk assignement 25 years ago The time I spent on training is 10x the porgramming itself, still accuracy is not very good This is true 101 of AI
Comment from : @alexsiuwh

Whats wrong with your voice, Lisha
Comment from : @ManaMatsuda-w6b

I guess at 14:38, the range of the 'b' matrix should be [b0, b1, b2, b3, , bk] instead of [b0, b1, b2, b3, , bn] as there are going to be 'k' biases in particular isn't it so??
Comment from : @smitchandi498

Спасибо за русскую звуковую дорожку
Comment from : @saniks3174

I love your visual methods of explaining the concepts If I want to understand any mathematical concept in depth, your channel is the one I turn to Thank you so much and I hope you continue making more videos like this
Comment from : @shreyassrinivasan5891

If only he knew…
Comment from : @SirDamatoIII

Thank you for all your effort! The animations are fantastic brAs a viewer, you may need to pause and rewind several times to fully appreciate these beautiful visuals that accurately represent the processes in the ANN
Comment from : @edtout

Comment from : @RichRufus-ti7ts

Not clear at all
Comment from : @cryptoinside8814

Image on eyes- matching it with pre images memory or information- resultthis is deep process of mindit can be easily obtain if you go threw diractionseven a whole scene can be grasp by computerbrMind also uses 4 sides of a scene or a figurbrI can tell you whole if neededpeople like you can open new doors with it what information i haves
Comment from : @mrperfect1067

Really good Thank you
Comment from : @NicholasCookhk852

2025 learners gather here
Comment from : @vinhquangnguyen5307

Excellent video It explains the concept of neural network so clearly
Comment from : @kenwang8595

Very good content
Comment from : @HenikajaAndriamahayIRIMANANA

It is incredible how well you explain it, very impressive for a topic like this
Comment from : @jovanplavsic5235

12:00
Comment from : @adam-v7p6k

Now I understood my maths lessons
Comment from : @moaazkamal78

I was a software developer from 1968 to 2014 and worked from just above the hardware level to the topmost application layer during my career, but neural nets are "magic" to me I hope I can grasp all this "magic" with your videos!
Comment from : @mdturnerinoz

hello, i know that it was a long time ago that you made this video, but i dont really understand the neurons in the hidden layers Are they also just numbers, or did someone code them?
Comment from : @Myself

13:37 There's an error on the SRT subtitle file, i hope it gets fixed later
Comment from : @izzuddin_

This video makes things easy to understand thanks
Comment from : @mk1212-t7n

I hope someday educational institutions play your videos as a standard Absolute legend with all the visuals and engaging content
Comment from : @shirsaksahoo

should‘nt all the weights unrelated to the to be recognised pattern, here the segment, be negative instead of just the ones around the edge ? Otherwise a pattern covering the entire image would be identified as a segment
Comment from : @yvesbernas1772

Have huge respect for these videos' creators
Comment from : @gdthegreat

Becoming a good trader takes time and patience When i first got into trading i was liquidated twice, and lost my entire mortgage deposit I could have given up, but decided to learn how to trade and put it into practice 4 years later and i am glad i made that decision
Comment from : @RobertLEndicott

I want to make a donation but I just lost my job :(
Comment from : @philtrem

Woah there is so much to learn when u actually pay attention to what ur watching lol
Comment from : @jbox77

Zaberdast
Comment from : @SuryaBano-c9w

12:46
Comment from : @manishsuthar6778

any chance you can expand this with the chain of through reasoning model and reinforcement learning?
Comment from : @nicksingh991

I have exam tomorrow
Comment from : @depresty

5:27 so, because the total no to be recognized are 10, ie 09, thats why there are 10 neurons in the last layerbrso if there is something that requires distinction between 4 objects, there would be 4 nuerons in the last layer? Am i right?
Comment from : @mukulwadhokar1192

I have completed the book now what should I do to keep learning Machine Learning I would be thankful for any gudance
Comment from : @rkgarg2953

Still don't understand the weights How do you choose what numbers to use for the weights
Comment from : @fahim3159

Love the energy in this one, 3Blue1Brown! brIf you’re ever brainstorming ways to improve storytelling or post-production, I’d be happy to chat!
Comment from : @pratapanurag757

BRO PUT HIS SOUL TO MAKE PRESENTATION❣
Comment from : @piyushsavaladekar3551

This is amazing, thank you for sharing
Comment from : @danieleflorenzano1100

Thankyou, I finally understood the basic principle of a neural network which 4 years of college failed to teach me
Comment from : @atharvakarawade9054

Amazing ❤❤
Comment from : @classicbaby-v3s

Sometimes I forget that it is such a great time to be alive, to be able to learn and understand complex mathematical analyses like these from educators like yourself, which has far reaching applications in the modern world
Comment from : @adarshagrl

How do you make this visualisations?
Comment from : @kleffy

@3blue1brownbr14:39 brfor bias matrix - number of rows should be k rite?? coz one bias for every neuron in the layer?brcorrect me if I am wrong
Comment from : @pavansainarasimha6741