I have to question the intelligence of the person who made the graphic of a brain with gears which are arranged in way that would prevent them from turning.

@MAGA_Extremist

Жыл бұрын

😅

@OferSadan85

Жыл бұрын

I know a few of those people whose brain works exactly like that 😂

Excellent presentation 8⁠-⁠)

Great Presentation !

What a cool presentation! Brilliant..... Brilliant!!

Thanks!

THANK YOU...!!!

for those familiar with AI in general, start viewing at 34:27 to get into the real topic here.

@joyboy-zx

Жыл бұрын

Hero without a cape

@stefan_popp

Жыл бұрын

Should have looked for this comment 32min ago...

I fear that field of research is going the wrong direction by trying to replicate yet another way to achieve a Von Neumann architecture: the brain doesn´t do logic gates or colocation of computing and memory! The connections *are* what the brain uses to compute *and* memorize. The memories and processing units *are* the connections. I´ll find this research field interesting when it will focus on circuits and materials that can dynamically reconfigure their connections, like the brain does. That´s the key ability to be searching for, here.

@lesliespeaker668

Жыл бұрын

I think that was just a proof of concept for the actual materials and design. Training them as logic gates is probably done because it's easy to measure. They could have trained them to do anything if they wanted, but keep in mind that with a single electronic neuron, there's not much you can do. The real magic of versatility happens when you have bigger matrixes of these things. Currently they are at the stage of figuring out how to best build these. And this is a huge step ahead should they succeed because the training happens in hardware, in the atom layer, in the material itself, without any computation at all. Current deep learning hardware is static, it doesn't change its own structure, because it's made of classic chip dies. As I understood it the goal here is to make deep-learning-in-material faster and more energy efficient than the current method which is deep-learning-in-software (on hardware which materials are not altered in the process).

@noorulali1184

Жыл бұрын

Good insight. We need to look at brains as a set of neurons and connections and explore how intelligence emerges from it.

In biochemistry, we wonder how linear behavior can arise from reaction kinetics that behave nonlinearly. Apparently the reverse problem is true in material science? Seems iffy.

Depending on the application, a slightly lower accuracy (96% compared to 99%) is totally acceptable if the computational cost is much lower.

1.) thanks a lot for this excellent presentation 2.) From my point of view there is another step missing: Adapting the network by evaluating the results of action, i.e. an evaluating instance => stored experience; => leading to "self optimization" ;-)

@kano6325

Жыл бұрын

Theoretically 😂🌎🧲👌

PILOT MAGICIAN

Im very surprised how little people showed up to this lecture, I dont know anyone who isnt at least curious about AI lol

You know the human civilization is in trouble when leading scientists defer to Wikipedia for basic definitions.

5:35 "For me the most amazing aspect is ... power consumption 20 Watts". I see you understand brain very well, that's certainly the most important thing about brain when you talk about information processing. Sharing this information with us makes it all clear and certainly almost nothing more left to do. Thanks for your lecture sir. 😀

😵‍💫

Is this man even human??

Needs better comparative metaphors to address recent cognitive neuroscience paradigm to shift to global cortex dynamic interactions ( and embodied mind and embodied cognition) away from 192 specialized loci networking ( thus departure for neural network modeling )

@Daniel-ft4of

Жыл бұрын

Have you read about "The non Local Universe"? I ask because the NIH says thats going to be the model of science moving forward. Funny how they haven't changed school curriculums yet.

Considering how complex and wonderful our brains it’s amazing how most of the population have yet to use even a tiny percentage of their brains potential

@kooshikoo6442

Жыл бұрын

That's an old myth that's been debunked.

Amazing ! First here 🥰

The brain is to produce more nueons and more connections physically, it is not about movement of flow of electrons and static infrastructure. The Infrastructure keep changing and learn, and the states are infinite. so the path should not limit the state for future. expect every nano second, there is repair, there is cleansing, there is construction, there is connection and disconnection is going on in a physical format. So not just power consumpsion also there is force to create, destroy, repair, connect with fundamental unit such as DNA and nerve gene unit.

Hmmm. I think neural biologists like Andrew Huberman can easily single out neural groups responsible for certain emotions and the causes and effects with mechanisms. Complex for sure and incpomple.

@Squirrel-zq6oe

Жыл бұрын

Right!? I've heard him do something of the sort quite a few times

@kooshikoo6442

Жыл бұрын

What makes you believe this? And what makes you think Andrew Huberman is so brilliant? His podcast? You do realize that he's shilling all sorts of products?

@scott-hr3hd

Жыл бұрын

@@kooshikoo6442 he has done it. He has a doctorate in neural biology and entomology and he teaches at Stanford University of medicine.

It's nanotech, you like it?

@kano6325

Жыл бұрын

Absolutely

Can a machine know it's being stared at like a human?

The replacement of human emotions and free will.

@Neptoid

Жыл бұрын

Yeah we often handwave away all the problems of humanity like that. What if emotions fill our networks with goal-orientedness? To understand how stuff like this evolved you need stuff like game theory, the gene's-eye view, but the adaptation of emotions probably is something computationally rational to have and to tug on the rest of the social network. Emotions have a purpose, all animal have it mate, it will play a role in life's landscape in the future and it even might turn out to be the ideal so we can interact with them. I'm sorry for getting biopunk here, but the closest extensions of ourselves, we might want to feel for ourselves and not them them feel on their own. I don't know what the landscape of life will be in the future, but we have only scratched the surface of what life can be

@47f0

Жыл бұрын

I see you have met my ex-wife.

@stefanolacchin4963

Жыл бұрын

Actually evidence is piling up that there is no such thing as free will.

@albert6157

Жыл бұрын

Free will is an illusion

@davidmiller9485

Жыл бұрын

@@albert6157 funny, You are the biggest illusion i've seen responding to this video so far.

I would be interested in a more technical lecture on this topic, maybe by bit better lecturer / grasp of english and without more than half the time spent connecting with novices or laymen.

Word to the unwise; Boo!

How about using mRNA? Oh hang on, I think someone already thought of that....

I really enjoy his class until he could not stop swallowing his saliva and I could hear it so badly on my Headphones for a better concentration but the fact that he refuses to drink water so he can stop the annoying sound of swallowing saliva. !!😫😖😫😠🤯

34:57 or 39:33 - Begins here, really. What a pity. Do you really think your audience 22:20 doesn't know what a neural net is, or how many neurons there are in the human brain? ---This is one of those terrible wastes of everyone's time, reviewing for at least 35 minutes information that everyone remotely acquainted with the field already knows. Gee. Lee Sodol lost to AlphaGo in 2016. Who knew?

what a mixed up lecture. sorry

@kano6325

Жыл бұрын

Relatively 😂🧲🌎👌

Idealizing materials has really been a great testable repeatable process of idealism for me. Like most my mind was cornered in school and heavily influenced by the industrial revolutionary mindset. we teach ourselves evolutionary mythology and materialism. Nothing wrong with this except how dogmatic it become. How hard headed even now it is being as it tries to change itself to adopt idealistic evidence. But this revival and rise of idealism brought about by computation with much better explanations of life and our universe are all very contradictory to Darwin's context and our models. Hijacking philosophy lol Idealism better explains idealism just materialism better explain material. take your cognitive bias medicines.

@ExtantFrodo2

Жыл бұрын

*"evolutionary mythology"* Evolution works whether you believe it or not. In point of fact, you can't even stop it.

@MAGA_Extremist

Жыл бұрын

@@ExtantFrodo2 right!?

@spaztor7723

Жыл бұрын

ew its a person that cant see the forest threw the trees

@dadsonworldwide3238

Жыл бұрын

@@spaztor7723 exactly they can't enjoy the beauty of the forest and it's engineering

@ExtantFrodo2

Жыл бұрын

@@dadsonworldwide3238 A forest is no less beautiful for having formed on it's own. A very wise man once said that nothing in biology makes sense outside the understanding of how life evolves. Our understanding does not diminish the wonder, it enhances it!

this, uh, was, umm, uh, really, uhhh, hard to watch.. couldn't finish it (he should have practiced; almost more filler-words than not... sigh..).