Excellent!!! More please.😊

Kind sir. Thank you for this video. If you have the time you should totally make that follow up video. The announcement of the Nobel prize is going to rake in views!

here after hopfield won the nobel prize in physics

Anyone here after the 2024 Physics Nobel Prize 🥇 👍🏻

I asked Gemini for some follow-up materials and here is the reply: The KZbin video you linked discusses a simplified model of human memory using neural networks. The creator promised a follow-up explaining how modifications to this simplified model could create a more accurate representation of human memory. Given your interest in accessible materials building upon the initial model, finding precisely matching content might be challenging. The topic sits at the intersection of neuroscience and artificial intelligence, and bridging the gap between a simplified explanation and a highly accurate, biologically plausible model requires significant in-depth knowledge. Therefore, I cannot directly provide a single video or article that perfectly matches the creator's unfulfilled promise. However, I can suggest avenues for further research that build upon the concepts presented: To understand the advancements beyond the simplified model, you should explore resources on these topics: Neural Network Architectures for Memory: Search for papers and articles on recurrent neural networks (RNNs), especially Long Short-Term Memory (LSTM) networks and Gated Recurrent Units (GRUs). These architectures are better suited for handling sequential data and temporal dependencies, crucial aspects of memory. Look for keywords like "RNN memory," "LSTM memory," or "neural network memory models." Many research papers are available on arXiv (arxiv.org). Be prepared for more mathematically intensive material than the initial video. Associative Memory Models: Explore research on Hopfield networks and other associative memory models. These models focus on the ability to retrieve a complete memory from partial cues, a key feature of human memory. Biological Plausibility in Neural Networks: Search for literature that bridges the gap between artificial neural networks and biological neural networks. Keywords such as "biologically plausible neural networks," "spiking neural networks," and "neuromorphic computing" will lead you to relevant research. This area is highly active and complex. Cognitive Neuroscience of Memory: To understand the biological basis of memory, explore introductory texts and review articles on cognitive neuroscience. Look for information on different memory systems (e.g., working memory, long-term memory, episodic memory, semantic memory) and their neural substrates. Finding Accessible Materials: While the more advanced research will be mathematically demanding, try to find review articles or introductory chapters in textbooks on neural networks and cognitive neuroscience. These often provide a higher-level overview before diving into the mathematical details. KZbin itself might offer some channels focusing on neuroscience or AI that could have related content, though it's less likely to find a direct continuation of the specific video's simplified model. Remember that bridging the gap between a simplified illustrative model and a biologically accurate model of human memory is a significant scientific challenge. The resources suggested above will offer a pathway to understanding the complexities involved.

I don't get what's so special about it. In my understanding, it's basically just a variation of a 2D "multilayer" perceptron and not an isolated concept. In this case, the inputs (=outputs of t-1) are fully connected to the outputs of timestamp t. Also, it's an autoencoder, because the inputs double as labels (targets) for the outputs. I put "multilayer" in quotes because it has no hidden layers (or bottleneck layer). The asynchronous update also seems antiquated in the era of modern GPUs, i.e. parallel computing. These are some core aspects where biological brains differ. Parallel computing isn't an issue with biological neurons. I also believe that the human brain is more similar to a multilayer autoencoder. With this analogy, the thalamus and hippocampus are examples of relay stations similar to autoencoder bottlenecks.

here because the nobel prize

Who came from the Nobel Prize announcement?

Who all are here after the Nobel Prize in Physics 2024 announcement?

The thing I know from my own memory is it is hazy. I can remember the important points or highlights, but if you ask me to paint sequence by sequence with absolute details of everything it is not possible. So our human brains must be compressing information. Only registering features and discarding the rest. On the otherhand the data contained in the storage media is absolute, quantized to practical limitations but very specific. So there seems to be a trade off.

i lost u there(idek where)

Wow, it’s really interesting thing

😵‍💫

Only one video 💔 Add some more.

Very well explained 👏

Time is Ripe! Id love to see a follow up video on the improved version you mentioned.

This is a brilliant video

Is it fair to say that the memory is (implicitly) stored in the connection weights?

antimemory is crazy if you assume our brains work like that

amazing xd

Its in the cloud

This is an absolutely brilliant content, thanks mate!

Are those 'human mistakes' mentioned the hallucinations in ai?

After all, thank you, a rich video. But I wished to show us how do you coded the output.

So if we die, all the electron in our neural network is going to disappear which mean there is no heaven?

No name author + some = work of art

I needed extremely long amounts of time to solve simple math problems in school, while being the only one being able to solve the most difficult and hard problems. How does time required of sorting and arranging a large amount of complex memory fragments play into that?

The video is NEXT LEVEL!

This video was just amazing.

People is not machine. That memory states are consistent or inconsistent is a wrong narrative. Even if some people claim that our memory is the movement of electrons as Faraday’s Constant, this claim is a hypothesis.

amazing

Estoquista é logística e Itaperuna só tá crescendo de uma forma ou de outra ele vai ter que aprender informática.

Nram

@layerwiselectures we need you back

Excellent video! I have a question: In a Network with different patterns stored, how I can evoke the memory of these patterns?

You definitely should keep making the vids if you are still interested, this is insanely amazing! You must be a big name there!

Guy, it's 1 year already. Cmon comeback bro

I wonder if this is going to help us understand how the human brain works and fix the memory loss issues in our biological bodies (without needing a neuralink)

Is this not just the same as stable diffusion or am I missing something?

13:08. this is how we connect dots as humans ,the new fact is merged with old fact we already know and a new memory is formed.

next pls

Beautifully explained

Gimme your patreon… i need more of this

holy molly you are doing great.

this dude dropped one incredibly amazing video and left

I definitely like your first video. The way you present the material is like a professor giving the class an overview of the subject. It's different from a college lecture, though, because it's stretched out about twice as long through the inclusion of explanations of very basic underlying concepts. I'm not sure this will work for you on KZbin; educational videos really belong in an online course.

guys, just wanted to remind you that this video is easily in the top 1% videos ever produced by a human

Does anyone have any idea or even guesses on which program was used to create the visuals for this video. The way the equations looked to be almost written in with pen. I’d be very interested in learning how that was done!

Why only one video??? Why not more???

I think we kind of store it somehow with help of the eye procesor that read and writes it and even creates its own like hallucination at wil and dreams at will... but the western society is further they can send memory and change what you have made in your mind or changed it... but when they attack you with it during sleep and awake they leave behind huge headache, and they do this to me everyday, thats why I dont sleep more than 5 hours subsequently as the headache of that 5 hour is enough so I just wake up take sigaret with cold water and then sleep again.. the reason is if your eye is closed the method they use traps your eye beams and then it creates more headache then when having eyes open as the beams they use dont reflect back to your eyes and brains pumps up even more and you get their memories/videos even more, that also you can change it and if it is too good they just wake me up or try to make it ugly dream I dont like and i wake up, I win sometimes from them but sometimes they make everything black with a beam to my head so wake up turn around in bed and then go to sleep again... little pricks using me as test human let alone giving me headache reading my mind or implanting it, again these ghosts are little hiding pricks that have full go of probably secret services in nato and other northern countries... as sunni muslims are too good to do this to people and africans/latin are way behind on this technology