Slashdot Mirror
Our Brains Don't Work Like Computers
Roland Piquepaille writes "We're using computers for so long now that I guess that many of you think that our brains are working like clusters of computers. Like them, we can do several things 'simultaneously' with our 'processors.' But each of these processors, in our brain or in a cluster of computers, is supposed to act sequentially. Not so fast! According to a new study from Cornell University, this is not true, and our mental processing is continuous. By tracking mouse movements of students working with their computers, the researchers found that our learning process was similar to other biological organisms: we're not learning through a series of 0's and 1's. Instead, our brain is cascading through shades of grey."
And it is for this reason that I loathe comparisons of computing power to brain power. "By 2015, we'll have computers as smart as humans." What kind of bullshit comparison is that? They're two completely different processes.
Fuck off.
How we know is more important than what we know.
...with floating point arithmetic. A "double" can represent a number between 0 and 1 with 15 decimals of precision, way more precise than any biological phenomenon. Computers can think like us, it's just a matter of writing the right floating-point code.
It's Roland Piquepaille, what did you expect, he's a fucktard and the only reason he's on Slashdot so much is that he has a business relationship with them.
How we know is more important than what we know.
I presume the info was a byproduct of a useful study (Cog-Neuro-Psy possibly?). I really hate it when the media picks out the And finally bit of science news stories (a la bread-landing-on-the-buttered-side, etc).
I don't know the key to success, but the key to failure is trying to please everybody. Bill Cosby (1937 - )
The article's summation is far more accurate than Slashdot. In TFA, a researcher says our minds don't work like digital computers.
The Slashdot headline says our minds don't work like computers, end of sentence.
Had TFSH (The Fine Slashdot Headline) been accurate, this would've been a mind-blowing result and in need of some extraordinarily strong evidence to support such an extraordinary claim. The question of whether the human mind--sentience, consciousness, and all that goes with it--is a computable process is one of the most wide-open questions in AI research right now. It's so wide-open that nobody wants to approach it directly; it's seen as too difficult a problem.
But no, that's not what these guys discovered at all. They just discovered the brain doesn't discretize data. Significant result. Impressive. I'd like to see significant evidence. But it's very, very wrong to summarize it as "our brains don't work like computers". That's not what they proved at all.
Just once, I'd like to see a Slashdot editor read an article critically, along with the submitter's blurb, before posting it.
Thank god we have someone like Roland Piquepaille to point out these amazing facts to us!
Yes, that was sarcasam!
"reality has a well-known liberal bias" - Steven Colbert
...the researchers found that our learning process was similar to other biological organisms....
That makes perfect sense, seeing as our brains evolved from other biological organisms.
Check out evolutionary psychology for some information. You'll view the world differently afterwards.
Evolutionary psychology (or EP) proposes that human and primate cognition and behavior could be better understood by examining them in light of human and primate evolutionary history... The idea that organisms are machines that are designed to function in particular environments was argued by William Paley (who, in turn, drew upon the work of many others).
Just because brains aren't binary or synchronously clocked doesn't mean much. One can create analog computers to represent shades of gray or create clockless computers that don't operate in lock-step synchronization. Furthermore, any digital, synchronous computer and simulate both shades of gray (with floating point numbers) and continuous processes (with sufficiently small time slices). Moreover, given the messiness of neuro-electrochemical systems, one can argue that it doesn't take a very precise float or a particularly dense time slicing to simulate neurons.
Some people ascribe the seeming magic of consciousness to some ineffable property of the brain, e.g., quantum mechanical effect. While other insist that its just what happens when you connect enough simple elements in a self-adaptive network.
The question is, are there neural input-output functions that are fundamentally not computable? If not, then a digital computer will, someday, reach human brain power (assuming Moore's law continues).
Two wrongs don't make a right, but three lefts do.
I don't think the chunk of meat in my head works using digital logic; but I'd like to think my Mind does a reasonable job of it.
Natural numbers (1,2,3...), true/false, up/down...
It's not unnatural to divide everything in half, heck our bodys are mostly symmetrical; the distiction comes in where the dividing line is.
We can weight our decisions in endless ways, if someone makes a statement, our belief of that statement depends on how many times we have heard it, our trust in the stater, if it meshes with known facts in the current context.
What I wonder is how far can a human mind be pushed in terms of concepts it can grasp and control it has, can a human visualise a 5 dimensional virtual object? control emotional responses, without supressing them? hold multiple contridictary world models? accelerate long-term memory access?
Even if you think of an electronic computer, it's just hordes of electrons rushing down pathways, only reliable because the voltage levels are continually refreshed at each step, a few electrons might wander off the path, but they are replaced at the next junction. Quantum Mob Rule.
We have no clue how the brain actually works. Sure, we know how individual neurons work, but no one can explain how a bunch of neurons creates a mind.
We look around our world and notice that computers are superficially similar to brains (e.g. they can both do math), so we hypothesize that they work similarly.
However, there's very little hard evidence supporting this hypothesis in the first place, so there's no "news" in this story.
Bottom line: The brain is not just a super-powerful computer.
The most rabid believers in American Exceptionalism are the exact same people whose policies are destroying it.
Dear Slashdot Editors,
Could we pretty, pretty please have a Roland Piquepaille section, so we can opt-out? I've been good all year, and it's almost my birthday, and I won't ask for anything for Christmas.
-Peter
There's a saying by neurophysiologists: "If the brain were simple enough to be understood, it would be too simple to understand itself"
"Gentlemen, you can't fight in here! This is the War Room!" -- Dr. Strangelove
Finally, a few good comments.
The point under discussion in this article is summed in this quote:
"More recently, however, a growing number of studies, such as ours, support dynamical-systems approaches to the mind. In this model, perception and cognition are mathematically described as a continuous trajectory through a high-dimensional mental space; the neural activation patterns flow back and forth to produce nonlinear, self-organized, emergent properties -- like a biological organism."
The goal is to forcefully point out (using an experiment) that the one way we think about mental processing, the digital computational model, is not very useful even at the trivial level of mental signal processing.
It's interesting how all the sarcastic comments about the "biological organism" reference completely miss the point. The point is that the signal is being processed in a way that could be modeled by the way a biological organism moves through space. It sniffs here, then there, then jumps to the solution. The signal processing itself exhibits emergent properties.
The reference to the dynamical system (http://en.wikipedia.org/wiki/Dynamical_system) is key. (I think people frequently fail to gloss the additional "al" and think this refers to some sort of generic "dynamic system"). Dynamical systems, although deterministic, are a foundational tool for developing chaos theory.
For me the interesting idea is that the default state of thought is in-betweeness. We stay jittering back and forth in an unresolved state until, suddenly, we aren't.
You know, we're all nerds, and we're all arrogant.
But what cracks me up is that the most arrogant assholes are the ones with the least skill or achievement. When you see someone harping the most about how he's uber-L33T because he knows what an IP address is, and how everyone else is an idiot... chances are it's someone who actually knows the _least_ about those. Chances are it's not a programmer who actually writes socket code, it's not a hardware engineer who's designed a network card, etc. No siree, it's a script-reader from the hell-desk that does the "I'm so l33t and everyone else is an idiot" fuss.
So you want to call people idiots if they don't know some computer trivia you know (off a list of canned answers)? Well, then being an EE and having some 20+ years of programming experience, I'll call _you_ an idiot, because you're below _my_ skill level.
Sure, you know what an IP or port number is or how to find it out in Windows. (Or can find it out on your list of canned answers.) But can you actually _use_ a socket on that port? Can you for example write a game server that listens on that port? If I gave you an old network card, can you find the right Linux kernel driver and change it to make it work with that card? Or what?
Or, ok, you do know what an IP address is. Congrats. Do you also know what a B-Tree is, how it works, and how to implement one in your code? Do you also know the difference between, say, MergeSort and QuickSort, and the influence of external (e.g., DB file on a disk) vs internal (in RAM) sorting on their performance? Can you implement either purely as, say, a state-machine driven by exceptions to signal state changes, just to prove that you actually understand the algorithm, as opposed to copying someone else's code off the net? Do you know the difference between bitmap indexes and b-tree indexes in Oracle, and can discuss when you might need one instead of the other?
Hey, it's computer stuff too. Very basic stuff too, nothing esoteric. We established already that computer stuff matters, and you're an idiot if there's something you don't know about them.
A polar bear is a cartesian bear after a coordinate transform.