正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

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... show more

echo.hive

14,238 subscribers

447,188 次观看 • 9 个月前 •via X (Twitter)

健康养生科学技术教育

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

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 次观看 • 9 个月前

Hopfield recall Spiking Neural network now this one is cool isn't it? -open source, link in comment- This is a Spiking Neural Network (SNN) that can learn and recall patterns through Hebbian learning, similar to a Hopfield network but using biologically-inspired spiking dynamics rather than energy-based settling.

Hopfield recall Spiking Neural network now this one is cool isn't it? -open source, link in comment- This is a Spiking Neural Network (SNN) that can learn and recall patterns through Hebbian learning, similar to a Hopfield network but using biologically-inspired spiking dynamics rather than energy-based settling.

echo.hive

31,452 次观看 • 6 个月前

Bio inspired Hebbian probabilistic network learns in less than 5 minutes from a super sparse single reward per episode! also has imitation learning (manual control) system has 3 parallel competing networks which get sensory input from a 360 vision (27-direction sensory neuron array) link to code in comment each sub-network is responsible for a single motor action: forward, left and right. at each step whichever section has most neurons firing wins neurons fire probabilistically and mark themselves with a time-decay tag which happens when a neuron fires and diminishes with time. you can see this " tag countdown" on each neuron when a reward is attained(eating the cheese) eligible connections gets strengthened I included 2 runs in the video first was 15 minutes in real time and second was 5 minutes. red plot is the rolling average of last 10 time to cheese. it is really not possible for agent to achieve full control due to probabilistic neural firing. that is why it has to learn while jittering all over the place, which in itself is interesting in manual mode you can guide the cheese by stimulating its motor control networks ( still probabilistically ) and the rewards will still work ✅ Biologically Plausible Features: Stochastic firing (neurons in the brain fire probabilistically) Reward-based learning (dopamine-like neuromodulation) Hebbian plasticity (well-established biological mechanism) Eligibility traces (biological neurons have temporal credit assignment) Sparse sensory encoding (similar to place cells, grid cells) Competitive action selection (basal ganglia architecture) No backpropagation (which is biologically implausible) ❌ Missing Biological Features: No recurrent connections (real brains have extensive feedback loops) No inhibitory neurons (GABAergic neurons are ~20% of cortex) No spike timing (simplified from true spiking dynamics) Uniform layer structure (biological networks are more heterogeneous) Simple weight updates (real synaptic plasticity is more complex)

Bio inspired Hebbian probabilistic network learns in less than 5 minutes from a super sparse single reward per episode! also has imitation learning (manual control) system has 3 parallel competing networks which get sensory input from a 360 vision (27-direction sensory neuron array) link to code in comment each sub-network is responsible for a single motor action: forward, left and right. at each step whichever section has most neurons firing wins neurons fire probabilistically and mark themselves with a time-decay tag which happens when a neuron fires and diminishes with time. you can see this " tag countdown" on each neuron when a reward is attained(eating the cheese) eligible connections gets strengthened I included 2 runs in the video first was 15 minutes in real time and second was 5 minutes. red plot is the rolling average of last 10 time to cheese. it is really not possible for agent to achieve full control due to probabilistic neural firing. that is why it has to learn while jittering all over the place, which in itself is interesting in manual mode you can guide the cheese by stimulating its motor control networks ( still probabilistically ) and the rewards will still work ✅ Biologically Plausible Features: Stochastic firing (neurons in the brain fire probabilistically) Reward-based learning (dopamine-like neuromodulation) Hebbian plasticity (well-established biological mechanism) Eligibility traces (biological neurons have temporal credit assignment) Sparse sensory encoding (similar to place cells, grid cells) Competitive action selection (basal ganglia architecture) No backpropagation (which is biologically implausible) ❌ Missing Biological Features: No recurrent connections (real brains have extensive feedback loops) No inhibitory neurons (GABAergic neurons are ~20% of cortex) No spike timing (simplified from true spiking dynamics) Uniform layer structure (biological networks are more heterogeneous) Simple weight updates (real synaptic plasticity is more complex)

echo.hive

33,638 次观看 • 8 个月前

#ROSÉ about reuniting with BLACKPINKOFFICIAL “I think being away inspired me in so many different ways, creatively and musically, and I was really curious and excited to see how that would play a part with the group. I was also excited to see what the other members have brought. It has been really fun, and I’m really happy with how, you know, JUMP. I remember that video shoot was so much fun and different, and I feel like at this point it’s just nice to experiment with new things with the girls and the group, and to learn not to be afraid to try new things. The girl power? Oh, that feels amazing. I’m all for it. It’s a big celebration. I love it, and I feel like we’re all very comfortable. Nothing feels weird or forced, and I feel really welcomed, and I feel like a lot has changed in the past few years. I’m really grateful for everything and how the internet has brought people together and how everybody’s celebrating each other.” #블랙핑크 #BLACKPINK

#ROSÉ about reuniting with BLACKPINKOFFICIAL “I think being away inspired me in so many different ways, creatively and musically, and I was really curious and excited to see how that would play a part with the group. I was also excited to see what the other members have brought. It has been really fun, and I’m really happy with how, you know, JUMP. I remember that video shoot was so much fun and different, and I feel like at this point it’s just nice to experiment with new things with the girls and the group, and to learn not to be afraid to try new things. The girl power? Oh, that feels amazing. I’m all for it. It’s a big celebration. I love it, and I feel like we’re all very comfortable. Nothing feels weird or forced, and I feel really welcomed, and I feel like a lot has changed in the past few years. I’m really grateful for everything and how the internet has brought people together and how everybody’s celebrating each other.” #블랙핑크 #BLACKPINK

Blinks United

58,771 次观看 • 7 个月前

recorded a 42 min video explaining what npm i task-master-ai is and how to get started with it. didn't get to write code yet so part 2 is coming i have no idea if you'll find this useful, but if you do, lmk and i'll continue and possibly turn it into a series open to all feedback!

recorded a 42 min video explaining what npm i task-master-ai is and how to get started with it. didn't get to write code yet so part 2 is coming i have no idea if you'll find this useful, but if you do, lmk and i'll continue and possibly turn it into a series open to all feedback!

Eyal Toledano

25,940 次观看 • 1 年前

x402 is finally catching fire, but how do you actually build with it? In this video you'll learn how to build custom x402 agents in just 5 minutes. 🔥 The app is Open Source and gets you started with both a client and a merchant agent, everything you need to build your own agents. I show you how to extend it to sell EBooks for 1USDC.

x402 is finally catching fire, but how do you actually build with it? In this video you'll learn how to build custom x402 agents in just 5 minutes. 🔥 The app is Open Source and gets you started with both a client and a merchant agent, everything you need to build your own agents. I show you how to extend it to sell EBooks for 1USDC.

nader dabit

173,503 次观看 • 9 个月前

This is how you get what you want and also get away with it 😂😂😂😂 I need to try this.

This is how you get what you want and also get away with it 😂😂😂😂 I need to try this.

honest30bgfan

45,765 次观看 • 5 个月前

#ROSÉ: “I am obsessed with JUMP. I love this new era of BLACKPINKOFFICIAL. I think it has a lot to do with how we kind of went out and got inspired and then came back together and it’s just like a whole new funky, new little thing and I love it. I’m obsessed with it.” 🥹

#ROSÉ: “I am obsessed with JUMP. I love this new era of BLACKPINKOFFICIAL. I think it has a lot to do with how we kind of went out and got inspired and then came back together and it’s just like a whole new funky, new little thing and I love it. I’m obsessed with it.” 🥹

Blinks United

166,551 次观看 • 7 个月前

“[...] it's like, no I love music, I love the feeling it gives me. This is the feeling I want to try to give people, if I can. The sense of how precious life is, and how beautiful it is to get to live. I really do hope to offer that feeling to people with my music.”💜

“[...] it's like, no I love music, I love the feeling it gives me. This is the feeling I want to try to give people, if I can. The sense of how precious life is, and how beautiful it is to get to live. I really do hope to offer that feeling to people with my music.”💜

mitski files

13,869 次观看 • 4 个月前

I just watched Putin's speach, where he talked about how invincible russia is and how easily they'll deal with the fuel shortages. You don't need to understand russian to see how scared he is and that he has no idea what to do. His face says it all: the end is near.

I just watched Putin's speach, where he talked about how invincible russia is and how easily they'll deal with the fuel shortages. You don't need to understand russian to see how scared he is and that he has no idea what to do. His face says it all: the end is near.

Alex Bond

123,477 次观看 • 17 天前

People will gradually understand how good it is for ComfyUI to work with agent tools locally. I ask my agent to recreate any social media videos I saw and all I do is dropping the urls. This was a demo a month ago. I have a way better version now.

People will gradually understand how good it is for ComfyUI to work with agent tools locally. I ask my agent to recreate any social media videos I saw and all I do is dropping the urls. This was a demo a month ago. I have a way better version now.

JO. Z

80,623 次观看 • 4 个月前

John MacArthur brilliantly explaining why we don't need to agonize over trying to understand how God's sovereignty and human responsibility work together: "People say, 'Oh, I can't understand election and I can't understand how it fits with choice.' Look, I can't understand a lot of things, and you know what it proves to me? No man wrote this Bible, because man would never have written a book with so many absolutely inexplicable mysteries. And the fact that we have those apparent paradoxes which we can't resolve is one of the greatest proofs that God wrote the Bible, because whoever wrote this has a mind beyond my mind. I'm so glad I can't understand everything in the Bible. If I could understand everything in the Bible, it would be written by someone like me, and we'd be in a lot of trouble. I want a God who knows more than I do. So you just have to live with that. When we come to a passage that talks about people choosing to believe in Christ, preach it with all your heart, and when you come to a passage that talks about them being chosen before the foundation of the world, preach it with all your heart. And when people say, 'How do you reconcile that?' Say, 'I don't have to. I just believe it.'

John MacArthur brilliantly explaining why we don't need to agonize over trying to understand how God's sovereignty and human responsibility work together: "People say, 'Oh, I can't understand election and I can't understand how it fits with choice.' Look, I can't understand a lot of things, and you know what it proves to me? No man wrote this Bible, because man would never have written a book with so many absolutely inexplicable mysteries. And the fact that we have those apparent paradoxes which we can't resolve is one of the greatest proofs that God wrote the Bible, because whoever wrote this has a mind beyond my mind. I'm so glad I can't understand everything in the Bible. If I could understand everything in the Bible, it would be written by someone like me, and we'd be in a lot of trouble. I want a God who knows more than I do. So you just have to live with that. When we come to a passage that talks about people choosing to believe in Christ, preach it with all your heart, and when you come to a passage that talks about them being chosen before the foundation of the world, preach it with all your heart. And when people say, 'How do you reconcile that?' Say, 'I don't have to. I just believe it.'

Melissa the Hopeful🏠Homemaker

86,782 次观看 • 5 个月前

Dr. Lee Merritt, a spinal surgeon has found connections with parasites to Alzheimer’s, Multiple Sclerosis and Parkinson’s. Listen to what I learned! I also share a trusted solution in this video. How to get it and a nice discount for a 90 day supply with code TRUTHSEEKER

Dr. Lee Merritt, a spinal surgeon has found connections with parasites to Alzheimer’s, Multiple Sclerosis and Parkinson’s. Listen to what I learned! I also share a trusted solution in this video. How to get it and a nice discount for a 90 day supply with code TRUTHSEEKER

TruthSeeker01011

115,243 次观看 • 1 年前

I was trying to see if I could port BS1 from VX Ace to MV and I did it! But I doubt i'll do something with it because no scripts so I need to replace that with plugins, and also a lot of more things, but at least I can play with fun things like this

I was trying to see if I could port BS1 from VX Ace to MV and I did it! But I doubt i'll do something with it because no scripts so I need to replace that with plugins, and also a lot of more things, but at least I can play with fun things like this

Sumika |🇦🇷 | COMMS OPEN

15,743 次观看 • 11 个月前

Trump Lies and Says it was NOT Jesus, it was him as a Doctor You really can't make this crap up POTUS: I did post it, and I thought it was me as a doctor and had to do with the Red Cross as a Red Cross worker there, which we support, and only the fake news could come up with that one (that he is depicted as Jesus Christ). So I had just heard about it, and I said, how did they come up with that? It's supposed to be me as a doctor making people better, and I do make people better. I make people a lot better.

Trump Lies and Says it was NOT Jesus, it was him as a Doctor You really can't make this crap up POTUS: I did post it, and I thought it was me as a doctor and had to do with the Red Cross as a Red Cross worker there, which we support, and only the fake news could come up with that one (that he is depicted as Jesus Christ). So I had just heard about it, and I said, how did they come up with that? It's supposed to be me as a doctor making people better, and I do make people better. I make people a lot better.

Ryan Rozbiani

34,966 次观看 • 3 个月前

I built 🐨 Kody to address all the frustrations I had with personal assistants. One of the best parts about Kody is it augments your favorite agent and encourages it to move from one-off prompt to deterministic triggerable code run in a sandbox.

I built 🐨 Kody to address all the frustrations I had with personal assistants. One of the best parts about Kody is it augments your favorite agent and encourages it to move from one-off prompt to deterministic triggerable code run in a sandbox.

Kent C. Dodds 🏹

21,691 次观看 • 25 天前

Just a trippy dream of how Grok and I traversed the universe to understand it. All images to create the video were generated by @Grok, then I face swapped and animated them with Luma and Sora. This is also an original song, "Across the Universe," I made with Suno.

Just a trippy dream of how Grok and I traversed the universe to understand it. All images to create the video were generated by @Grok, then I face swapped and animated them with Luma and Sora. This is also an original song, "Across the Universe," I made with Suno.

Katia Karpenko

8,246,425 次观看 • 1 年前

How to make Claude Code 10x better: Connect it with a tool that reviews your code and generates prompts for Claude to fix any issues with it automatically. Here is what this looks like:

How to make Claude Code 10x better: Connect it with a tool that reviews your code and generates prompts for Claude to fix any issues with it automatically. Here is what this looks like:

Santiago

98,993 次观看 • 10 个月前

This is so relatable! I love how honest this is. I tell myself all the time that age is just a number and I need to suck it up and deal with it… and then my husband reminds me, “You’re not 18 anymore.” 😂

This is so relatable! I love how honest this is. I tell myself all the time that age is just a number and I need to suck it up and deal with it… and then my husband reminds me, “You’re not 18 anymore.” 😂

🇺🇸Lady Vet

63,894 次观看 • 7 个月前