Brain: complex quantum device
Computer: electrons flying around.
Why do AI researchers even hope that they can emulate the brain or create some intelligence and shit? It's clearly retardedly hard. Why are we being warned against rogue AIs and shit?
Basically AI nowadays is just a massive curve fitting operation be it RNNs or LSTMs or whatever you're using. Is it a dead fucking field?
Pic very unrelated.
Are AI researchers really trying to emulate a brain in software. Maybe they might be trying to understand some heuristics used by humans so they can use them in programming a theorem prover or something but is anyone actually trying to build a copy of a brain using electronics?
>>7660063
Fuck your shit OP.
I want to know if that troll science comic would work at a small amount.
The brain is no more a quantum device than computer circuitry. Both exist in a world subject to quantum laws. And neither exploit those laws in any particular way, but rather seem to do their best to cancel out all quantum jitter from their mechanisms.
The brain is several magnitudes more complicated than the wildest fever dreams of a CPU architecture designer, but the difference is not some hand-wavy thing like "quantum mechanics." Instead it is raw fucking complexity.
The days of "we'll just throw in a bunch of transistors and a pinch of programmer common sense and maybe AI will happen" are long, long gone.
For a while, expert systems had a lot of steam but they are out of fashion.
But now, the word of the day is (like everywhere else): BIG DATA. You can actually do pretty cool stuff with massive amounts of data so it did blow a bit of life into the field.
>>7660105
no you fucking moron
>>7660063
Nice bait m8. Do yo seriously think people take penrose's ideas seriously?
>>7660105
Without sufficient light, a plant won't generate oxygen even if it's still alive.
>>7660063
>Is it a dead fucking field?
It's the state of the art in many areas of computer science so it's hardly dead.
It's obvious you're butt blasted for some reason, maybe realizing that your single digit IQ brain can be outperformed by computers is a bit painful.
>>7660063
If by "quantum device" you mean "individual receptors interact with their ligands and allosteric modulators probabilistically" then yes
>>7660108
I wouldn't be so dismissive of the brain potentially having unknown processes. We still don't know why humans perceive time the way they do.
>>7660189
Even though the brain has untold many emergence properties we do not understand it's hard to believe it contains such dramatically novel physical functions. It's not like the studied physical parts are particularly small, so where's the quantum enhancing process hiding?
>>7660142
So you need to add a flash light to the equation is what you're telling me.
>>7660200
Are magnetic fields enough to massively interfere with any quantum system?
>>7660221
yes
>http://www.express.co.uk/news/uk/611992/Scientists-experiment-magnets-immigrants-God-magnetic-waves
finally proof that liberalism is a mental illness.
>>7660189
>I wouldn't be so dismissive of the brain potentially having unknown processes.
Why not? The low-level function of neurons is actually pretty well studied. We know how transistors work, but not how to arrange them into CPUs.
>>7660220
Flash lights don't emit much in the way of UV radiation that plants need.
>>7660302
>plants need
>UV radiation
Aw, you've never seen a chlorophyll absorption spectrum, have you?
>>7660171
Lol fail fuck
OP here. By dead field I don't mean its out of use, I mean like there's really nothing else to see until our hardware gets a big boost some time. Honestly I don't see many developments in AI anymore, that's why. All that's happening is just applications. Maybe its getting more specialised like physics, we can no longer have einsteins popping up with random epic shit and making everyone crap their pants anymore. I dunno.
>>7660063
>brain
>quantum device
source needed
>>7660063
An AI doesn't even have to be that smart to start fucking everything up. That's the scary part.
And it is probably ded if google made their shit open source.
>>7661995
Penrose, I don't have to explain shit
>>7661985
You just see the applications? Then you aren't looking hard enough. Neural turing machines was a recent development to give an example
>>7662138
It was a shit made like last year or something.
>>7660063
If you had enough plants and some kind of water purification system, and you didn't dive to deep, why would your pic not work?
>>7662596
Sunlight?
>>7662596
the excess of carbon dyoxide would make you suffocate
>>7660063
Obvious troll, but I'll try to answer anyway.
>Brain: complex quantum device
Citation Needed
>Why do AI researchers even hope that they can emulate the brain
Why would they try to emulate the brain? The human brain is extremely inefficient compared to the current ANN architectures which achieve the same sub-tasks.
>Basically AI nowadays is just a massive curve fitting operation be it RNNs or LSTMs
So?
Empirically, AI is now advancing at an increasing pace: Every few months, an R&D groups in Facebook, Google or Microsoft comes up with new algorithms that achieves results which were considered impossible only two years past.
>>7660063
>Curve fitting
>LSTM
Op clearly has no fucking clue what he is talking about. Also, LSTM IS a subclass of RNN
>>7662596
Too slow, and too trivial. Even plants that have a decent net oxygen release (which is not all plants), you'd need too many and they wouldn't keep up with you. There's also the matter of sunlight, nutrients in soil, bacteria to perform nitrogen fixation, etc.
>>7663391
You're a genius m8
The very act of correcting weights regardless of using gradient descent or some other method is, fucking, curve fitting. Do you understand curve fitting as drawing a line of best fit? If so, please return to your high school physics class.
>>7663365
Obvious troll, but I'll try to answer you anyway
>brain in inefficient
Citation needed
>ANN can do same shit as brain, better
Nothing has passed the turing test yet. There's something they can't do properly.
>>7660063
can we see more of this
>>7660189
Why do computers perceive time? They can't store all the states that they have ever had, or if they do store them, its only fragments or important parts. Logic gates don't really work backwards per se: I can't know what the inputs of an OR gate were just by looking at the output and nothing else. Furthermore computers need a constant tick-tock of a clock cycling the inputs to outputs, and those back into more inputs at a very basic level. Supposing a computer was self aware, wouldn't only be "aware" of its current state at that particular clock cycle?
Maybe we're the same way, we can only be aware of what "cycle" or second it is now, we only have enough storage to have a partial memory of the past, and trying to look forward become very murky and muddled the further we look out.
I could go on, but its not very /sci/ more poorly philosophical. fft I suppose.
Literally how is brain a quantum device? It's a deep deep neural mesh.
>>7660117
Good post. Why are expert systems dead? I remember sitting wide eyed during the coursera lectures on artificial intelligence and imagining the possibilities of those systems.
Those multi agent search algorithms were out of this world, so to speak. I feel data Mining is dumbed down in comparison
>>7663932
We don't know.
Right now it appears to be able to be described well enough using more "classical" interactions, but a quantum aspect could be involved.
Why you ask?
Because we don't know shit about the universe either. Quanum anything somehow is still in its infancy, for all we can guess, the brain evolved to best utilize future events influencing the outcome of past ones. Somehow. Machines that worked in such a way would have an obvious cumulative advantage.
>>7661985
If you aren't seeing developments, you aren't looking.
>>7663939
>utilise future events influencing the outcome of the past ones
What?
>turn on calculator
>prints 0
>press 2
>prints 2.0
>press + 3
>prints 5.0
>prints + 3 again
>prints 8.0
HOLY SHIT IT IS LEARNING!
>>7660063
>Brain: complex quantum device
Stopped reading there. Seriously this hypothesis has absolutely no support, at least not when I last read about it.
