But what is a neural network? | Deep learning chapter 1
Video Overview & Insights
What are the neurons, why are there layers, and what is the math underlying it?
The handwritten-digit example made neural networks feel much less mysterious. Seeing how pixels, weights, biases, and layers work together was far clearer than just hearing the usual “it learns like a brain” explanation.
Help fund future projects: https://www.patreon.com/3blue1brown
Written/interactive form of this series: https://www.3blue1brown.com/topics/neural-networks
Your explanation is underrated I learned a lot from it.
Additional funding for this project was provided by Amplify Partners
For those who want to learn more, I highly recommend the book by Michael Nielsen that introduces neural networks and deep learning: https://goo.gl/Zmczdy
I'm just seeing this to learn AI. Thanks 3b1b!
There are two neat things about this book. First, it's available for free, so consider joining me in making a donation to Nielsen if you get something out of it. And second, it's centered around walking through some code and data, which you can download yourself, and which covers the same example that I introduced in this video. Yay for active learning!
https://github.com/mnielsen/neural-networks-and-deep-learning
This tutorial is underrated I enjoyed every second of it.
I also highly recommend Chris Olah's blog: http://colah.github.io/
For more videos, Welch Labs also has some great series on machine learning:
The editing is worth sharing It helped me understand better.
https://youtu.be/i8D90DkCLhI
https://youtu.be/bxe2T-V8XRs
Bro are people really this smart
For those of you looking to go *even* deeper, check out the text "Deep Learning" by Goodfellow, Bengio, and Courville.
Also, the publication Distill is just utterly beautiful: https://distill.pub/
I feel like I struggled most of my life with how to manifest. Sometimes it felt so out of reach that I decided to give that archetype quiz by Vera Hart a try, and it really did wonders. It matched my personal archetype exactly, and that’s how I finally figured out how to do it properly. I highly recommend anyone who struggled like me to try it as well! It only takes a couple of minutes anyway ❤️
Lion photo by Kevin Pluck
Звуковая дорожка на русском языке: Влад Бурмистров.
Reticulating splines.... reticulating splines....
Ok, done, the souls of the dead will now transfer information through fermion pattern quasi - particles, to become Ai generated suggestions, then categorized, and animated into sprites, for VIDEO GAME BAD GUYS
Thanks to these viewers for their contributions to translations
German: @fpgro
This is so insightful. I’m trying to learn AI and this video was recommended to me so I understand what I’m going into and the fact it was 8 years ago is mind blowing 🤯 ❤️
Hebrew: Omer Tuchfeld
Hungarian: Máté Kaszap
as a 14 year old, this is so surprisingly simple?????? the last time I tried to learn about how neural networks work I was so confused somehow. I just hope training it isn't as hard
edit:it sounds so simple?? I mean they didn't show the math but just. the way it works in general makes so much sense. thank you SO MUCH for this, it really kickstarted a project I was thinking of making.
Italian: @teobucci, Teo Bucci
-----------------
It was wonderful to me. appreciated
Timeline:
0:00 - Introduction example
P. S. ❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤
1:07 - Series preview
2:42 - What are neurons?
Pat Salis ❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤
3:35 - Introducing layers
5:31 - Why layers?
Patrick Salis ❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤
8:38 - Edge detection example
11:34 - Counting weights and biases
P. S. ❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤
12:30 - How learning relates
13:26 - Notation and linear algebra
Patrick Salis ❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤❤
15:17 - Recap
16:27 - Some final words
This was incredibly well made.
17:03 - ReLU vs Sigmoid
Correction 14:45 - The final index on the bias vector should be "k"
the shit... it has even 9 languages... haahhh!
------------------
Animations largely made using manim, a scrappy open source python library. https://github.com/3b1b/manim
crazy how this was made before Chatgpt released haha
If you want to check it out, I feel compelled to warn you that it's not the most well-documented tool, and has many other quirks you might expect in a library someone wrote with only their own use in mind.
Music by Vincent Rubinetti.
This channel is worth sharing It deserves more views.
Download the music on Bandcamp:
https://vincerubinetti.bandcamp.com/album/the-music-of-3blue1brown
Me on neural networks: living under a rock
Stream the music on Spotify:
https://open.spotify.com/album/1dVyjwS8FBqXhRunaG5W5u
i dont think anyone has every explainted it to me so clearly and beautifully. thank you for your efforts for explaining complex concepts in such ansimple way. following and going to keep learning.
If you want to contribute translated subtitles or to help review those that have already been made by others and need approval, you can click the gear icon in the video and go to subtitles/cc, then "add subtitles/cc". I really appreciate those who do this, as it helps make the lessons accessible to more people.
------------------
這對我的深度學習很有幫助。
3blue1brown is a channel about animating math, in all senses of the word animate. And you know the drill with YouTube, if you want to stay posted on new videos, subscribe, and click the bell to receive notifications (if you're into that).
If you are new to this channel and want to see more, a good place to start is this playlist: http://3b1b.co/recommended
awesome
Various social media stuffs:
Website: https://www.3blue1brown.com
❤️❤️
Twitter: https://twitter.com/3Blue1Brown
Patreon: https://patreon.com/3blue1brown
Ugh i love this sm
Facebook: https://www.facebook.com/3blue1brown
Reddit: https://www.reddit.com/r/3Blue1Brown
sooo many people talking about that vera hart archetype quiz its crazy, lol. Tried it myself and the results are quite promising.
More User Perspectives
I feel like I struggled most of my life with how to manifest. Sometimes it felt so out of reach that I decided to give that archetype quiz by Vera Hart a try, and it really did wonders. It matched my personal archetype exactly, and that’s how I finally figured out how to do it properly. I highly recommend anyone who struggled like me to try it as well! It only takes a couple of minutes anyway ❤️
@bachelor-point-4why green and red :(((, colorblind ppl struggling fr
@AkiAkiVThis channel is worth sharing I enjoyed every second of it.
@NapalRoy-sm3qxI LOVE YOU
@halizneThis is the kind of explanation that makes learning genuinely enjoyable. Great work! 🙌
@aman_rawat85Brilliant explanation—made a complex topic feel surprisingly easy to understand. 👏
@sahil.bk47Awesome video, looking forward to the next one!
@isaiassoliz-y9mКонец эры точных вычилений
@user-mn7pob0bwThis video is top notch It helped me understand better.
@NapalRoy-sm3qxYour videos are always so helpful, thank you!
@AnjjiMaggamThis video is well made It deserves more views.
@AmireHussen-m3oWait, was THIS advanced 8 years ago?!!
@cronoxonixThank you for the great content!
@TitanKaisenLoved this video, keep up the great work!
@imran0otik276I just thought of something interesting related to the animation that you showed of the 784 dimensional input space flowing through the network to the 10 dimensional output space.
A neural network like this could be thought of as a compression algorithm that compresses the higher dimensional input down to a lower dimensional output, and training the network is just fine tuning the compression algorithm.
Great job, this video was very useful!
@ChandramohanChandramohan-m3h9ucool
@thecutepro2447GenZ going nuts at 15:21
@nungbear1398Great job, this video was very useful!
@SofeyaMoina-m2uno one can explain such a complex topic like you did, also I'm watching this after 8 years since it was posted. The complexity of the topic depends on how it was explained...and you did a really great job on making it the simplest one .
@TOP_G-xoxmuy bueno. Detalle: es un castellano muy de Miami Cartoon Network, prefiero los subtítulos. Igual entiendo que es difícil traducir a los distintos castellanos
@martinditullio1411My goat
@animatrix7721Given how important LLMs / "AI" (machine learning) are these days for many jobs, it would be dope if you had a series that goes further into the research of LLMs. Pick a topic, dive into it, and as always, give understanding to hold up the overarching guidelines rather than it just being a guideline we trust on blind faith. An example: Why is there a U-shaped curve where LLMs tend to focus mostly on the beginning and end of its context? Asking Claude, it identified softmax dispersion, something called RoPE, something called attention sinks, and training data bias (most RLHF etc. was done on smaller contexts, so simply by how it was trained, large contexts by their length can take an LLM into out-of-distribution space, causing it to malfunction. And that's just one practical nugget of information with four technical explanations found in the research to explain it. Could be an entire series of deep diving into deep learning.
@AG-ld6rv