Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

I animated how a neural network learns, from one neuron to backpropagation, simple math, no calculus required. Best 8 minutes you can spend if it has only ever been a black box to you.

tetsuo

237,248 subscribers

48,685 Aufrufe • vor 2 Monaten •via X (Twitter)

Bildung Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

As someone who ships LLM systems in production, this backpropagation video is the closest thing to a "why GPT actually learns" explainer I've ever seen released for free. Everyone thinks training a neural network is calculus. It's not. It's 175 billion weights getting nudged proportionally to how much each one caused the mistake. PyTorch does this backwards pass on every single training step. 20 minutes. Bookmark & watch today. The algorithm is older than the iPhone.

As someone who ships LLM systems in production, this backpropagation video is the closest thing to a "why GPT actually learns" explainer I've ever seen released for free. Everyone thinks training a neural network is calculus. It's not. It's 175 billion weights getting nudged proportionally to how much each one caused the mistake. PyTorch does this backwards pass on every single training step. 20 minutes. Bookmark & watch today. The algorithm is older than the iPhone.

Eron

20,078 Aufrufe • vor 18 Tagen

Hey there, if you had a little money, how would you spend it? I bet I can guess how you wouldn’t spend it…

Hey there, if you had a little money, how would you spend it? I bet I can guess how you wouldn’t spend it…

PoliticsGirl

12,157 Aufrufe • vor 2 Monaten

Spiking Neural Network from scratch achieves 8% accuracy. no backpropagation or SGD I created a genetic hyper parameter optimizer and it now, on average, can get 8% accuracy which is ~3% above chance Link to source code with a detailed video and markdown explanations in comment it also usually starts lower than 5% and slowly improves then eventually can start dipping below 5%, all of which leads me to believe there is glimmers of learning taking place. sometims it is stabel around 7-8-9% for a long time there is no backpropagation or SGD. it learns via STDP (spike timing dependent plasticity) and a reward mechanism each example is presented n many times (500 in this case) which produces spike train which leads to eligibility list, at the end of a turn based on if the answer was correct or not we adjust the weights using the reward as a multiplier spike timing keeps track of the sequence of neuron firings and which ones were more likely to have lead to the correct answer let me know what you think

Spiking Neural Network from scratch achieves 8% accuracy. no backpropagation or SGD I created a genetic hyper parameter optimizer and it now, on average, can get 8% accuracy which is ~3% above chance Link to source code with a detailed video and markdown explanations in comment it also usually starts lower than 5% and slowly improves then eventually can start dipping below 5%, all of which leads me to believe there is glimmers of learning taking place. sometims it is stabel around 7-8-9% for a long time there is no backpropagation or SGD. it learns via STDP (spike timing dependent plasticity) and a reward mechanism each example is presented n many times (500 in this case) which produces spike train which leads to eligibility list, at the end of a turn based on if the answer was correct or not we adjust the weights using the reward as a multiplier spike timing keeps track of the sequence of neuron firings and which ones were more likely to have lead to the correct answer let me know what you think

echo.hive

225,484 Aufrufe • vor 10 Monaten

"I want to be one of the best athlestes ever" - Arman Tsarukyan sets his sights on becoming one of the best athletes ever 🫡👏 "My sport is MMA. Wrestling and grappling are just side quests. There’s no pressure, you just go there, you win or you lose, nobody cares. I’m doing this for fun. People think I’m not hungry - no guys, I’m here for a long time. I don’t just want to make money and be a UFC champion, I want to be one of the best athletes ever. To be the best in MMA, to be the best in wrestling, to show everybody you can grapple and wrestle. And if the UFC lets me box, I’ll box full time. I can box, I can wrestle, I can grapple, and I can do my sport, MMA." (Via The Schmo )

"I want to be one of the best athlestes ever" - Arman Tsarukyan sets his sights on becoming one of the best athletes ever 🫡👏 "My sport is MMA. Wrestling and grappling are just side quests. There’s no pressure, you just go there, you win or you lose, nobody cares. I’m doing this for fun. People think I’m not hungry - no guys, I’m here for a long time. I don’t just want to make money and be a UFC champion, I want to be one of the best athletes ever. To be the best in MMA, to be the best in wrestling, to show everybody you can grapple and wrestle. And if the UFC lets me box, I’ll box full time. I can box, I can wrestle, I can grapple, and I can do my sport, MMA." (Via The Schmo )

Red Corner MMA

92,679 Aufrufe • vor 5 Monaten

i used to spend 30 minutes hunting for a twitter profile mockup tool just to see if my avatar and header worked together before going live. last night i asked computer to build me one from scratch. took less than 10 minutes. it works better than anything I ever found. now you can use it/remix it by downloading the zip file.

Phi Hoang

41,706 Aufrufe • vor 6 Monaten

this is one of my best app ideas i’ve ever posted. i don’t know why it has only 10k views. but i just found a format you can use to promote it with $0. it has 3.1M views.

this is one of my best app ideas i’ve ever posted. i don’t know why it has only 10k views. but i just found a format you can use to promote it with $0. it has 3.1M views.

Simone Canc

20,619 Aufrufe • vor 4 Monaten

Behavior expert Chase Hughes explains how you can easily "split" a personality in someone through simple conversation. He walks through the exact steps required to do this, and how the military has been using it for years...

Behavior expert Chase Hughes explains how you can easily "split" a personality in someone through simple conversation. He walks through the exact steps required to do this, and how the military has been using it for years...

Danny Jones

19,449 Aufrufe • vor 1 Jahr

Happy anniversary Bae!! I've been too busy to make a new animation, so here's my favourite one I've ever done of you. If it wasn't for you, I don't think I ever would have stuck with animation this long. Thank you for 5 amazing years!! Original artist: @/bleufld Animated by me a while ago (with no help from AI) #Baelzbrush

Happy anniversary Bae!! I've been too busy to make a new animation, so here's my favourite one I've ever done of you. If it wasn't for you, I don't think I ever would have stuck with animation this long. Thank you for 5 amazing years!! Original artist: @/bleufld Animated by me a while ago (with no help from AI) #Baelzbrush

Jusagi 🎲⏳

16,597 Aufrufe • vor 4 Tagen

I built a biologically inspired spiking neural network from scratch and it learned with %5 accuracy to do addition :) There is no backpropagation, no artificial loss functions - just spikes, synapses, and dopamine-like reward signals. it uses STDP -> "Spike-Timing-Dependent Plasticity" with modulated rewards This is super fun and I will try to get it to learn with better accuracy. I also need to better understand how all the moving parts fit together Link to source code in comment which has a detailed readme and html with animations explaining how it all works

I built a biologically inspired spiking neural network from scratch and it learned with %5 accuracy to do addition :) There is no backpropagation, no artificial loss functions - just spikes, synapses, and dopamine-like reward signals. it uses STDP -> "Spike-Timing-Dependent Plasticity" with modulated rewards This is super fun and I will try to get it to learn with better accuracy. I also need to better understand how all the moving parts fit together Link to source code in comment which has a detailed readme and html with animations explaining how it all works

echo.hive

447,674 Aufrufe • vor 10 Monaten

8½ (1963, Fellini) “I don’t like the idea of ‘understanding’ a film. Either a film has something to say to you or it hasn’t. If you are moved by it, you don’t need it explained to you. If not, no explanation can make you moved by it.” ― Federico Fellini

8½ (1963, Fellini) “I don’t like the idea of ‘understanding’ a film. Either a film has something to say to you or it hasn’t. If you are moved by it, you don’t need it explained to you. If not, no explanation can make you moved by it.” ― Federico Fellini

DepressedBergman

399,680 Aufrufe • vor 2 Jahren

Only if you've been in the deepest valley can you ever know how magnificent it is to be on the highest mountain.

Only if you've been in the deepest valley can you ever know how magnificent it is to be on the highest mountain.

Richard Nixon Foundation

47,847 Aufrufe • vor 1 Monat

Only if you've been in the deepest valley can you ever know how magnificent it is to be on the highest mountain.

Only if you've been in the deepest valley can you ever know how magnificent it is to be on the highest mountain.

Richard Nixon Foundation

126,522 Aufrufe • vor 2 Jahren

I can relate to woody in this one. If money has ever been tight you understand! Where everybody knows your name!

I can relate to woody in this one. If money has ever been tight you understand! Where everybody knows your name!

Johnny Midnight ⚡️

41,752 Aufrufe • vor 2 Monaten

As someone who's spent 3 years fine-tuning ML models, this lecture on neural networks from a Stanford math grad is the closest thing to a no-bullshit "day one of ML" briefing I've ever seen released publicly for free. Everyone thinks neural nets are magic. They aren't. They're 13,000 dials - a matrix multiplication, a sigmoid squish, done. Once you see the math, "AI" stops feeling mysterious. 18 minutes. Bookmark & watch today.

As someone who's spent 3 years fine-tuning ML models, this lecture on neural networks from a Stanford math grad is the closest thing to a no-bullshit "day one of ML" briefing I've ever seen released publicly for free. Everyone thinks neural nets are magic. They aren't. They're 13,000 dials - a matrix multiplication, a sigmoid squish, done. Once you see the math, "AI" stops feeling mysterious. 18 minutes. Bookmark & watch today.

fintex

330,207 Aufrufe • vor 1 Monat

If you spend 1200 G-Coin and level the Black Market pass to 30, you will not be able to purchase a prog skin in the shop. If you spend 3800 G-Coin and level the pass you will still not be able to purchase a single prog skin. I think the greed ceiling has finally been reached.

If you spend 1200 G-Coin and level the Black Market pass to 30, you will not be able to purchase a prog skin in the shop. If you spend 3800 G-Coin and level the pass you will still not be able to purchase a single prog skin. I think the greed ceiling has finally been reached.

SteveTheHappyWhale 🐳

32,208 Aufrufe • vor 2 Jahren

A cement company just created one of the best commercials I've ever seen. Proof that "boring" work can have a beautiful impact. No better way to spend the next 2 minutes.

A cement company just created one of the best commercials I've ever seen. Proof that "boring" work can have a beautiful impact. No better way to spend the next 2 minutes.

Alex Lieberman

146,408 Aufrufe • vor 2 Jahren

📁 Terence Tao, Fields Medalist mathematician, explains that neural networks do not understand concepts, but they uncover hidden patterns. In knot theory, a classical neural network discovered correlations between mathematical invariants no one had anticipated. It started as a black box, then became a real clue to deep structures humans had missed.

📁 Terence Tao, Fields Medalist mathematician, explains that neural networks do not understand concepts, but they uncover hidden patterns. In knot theory, a classical neural network discovered correlations between mathematical invariants no one had anticipated. It started as a black box, then became a real clue to deep structures humans had missed.

Jon Hernandez

204,985 Aufrufe • vor 7 Monaten