Modern computing vs Biology

The magic smoke.

Moderators: phlip, Moderators General, Prelates

cspirou
Posts: 147
Joined: Wed Jun 11, 2008 4:09 pm UTC

Modern computing vs Biology

Postby cspirou » Tue Aug 10, 2010 6:22 pm UTC

There's a BBC article about how the brain works more like the internet then a traditional computer.

http://www.bbc.co.uk/news/science-environment-10925841

I'm not much of a computer scientist but I am very interested in the relation between computer science and nature. Is it more efficient to design computers similar to how things are done in nature instead of just having a CPU in charge of everything? There is somewhat of a trend right now with most of the graphics being processed in a GPU instead of a CPU. Does this mean that the computer of the future should have several chips with each one working on a specialized routine? Or possibly a CPU with several cores and each core specialized towards a different task?

User avatar
Obby
Posts: 785
Joined: Thu Mar 06, 2008 11:37 pm UTC
Location: Philadelphia

Re: Modern computing vs Biology

Postby Obby » Tue Aug 10, 2010 6:33 pm UTC

That kind of makes sense. This, at least, provides a bit of an explanation as to why the human brain can so readily adapt to certain kinds of damage (for instance, a blind person being able to hear more effectively).
The story so far:
In the beginning the Universe was created.
This has made a lot of people very angry and has been widely regarded as a bad move.

ImagingGeek
Posts: 380
Joined: Fri Aug 06, 2010 6:41 pm UTC
Location: Canada

Re: Modern computing vs Biology

Postby ImagingGeek » Tue Aug 10, 2010 6:40 pm UTC

If you have the time, this is a great video that goes into some aspects of how our minds work, in the context of morality:
'Morality: From the Heavens or From Nature?' by Dr. Andy Thomson

Word of warning, that vids a hour long. And once you watch one of the AAI presentations, you'll want to watch them all (they're all good, BTW).

The fascinating thing is that our minds do not consist of very powerful or fast "computers". Instead they consist of thousands (if not millions) of "cores" connected in parallel, each "core" running at a few hertz. Both humbling (average chip these days has 4-6 cores) and unbelievable (our brain works at a few Hz, verses hundreds of MHz to low GHz for computers) :!:

Even more amazing is that different "cores" can "argue" with each other - when your faced with a difficult decision and feel "torn", that is an actual reflection of what is going on in your mind - one part of your brain fighting with another part of your brain. The middle section of the above video covers that - in the context of decisions which involve balancing moral, numerical and self-preservation issues - all of which are computer by different "cores" of our brains.

Bryan
I have a new blog, on making beer.

Not that anyone reads it...

User avatar
psykx
Posts: 408
Joined: Sat Feb 23, 2008 11:24 pm UTC
Location: England
Contact:

Re: Modern computing vs Biology

Postby psykx » Tue Aug 10, 2010 7:28 pm UTC

I've heard it said a human brain woks at 40 Hrz, although I can't remember where from at the moment.
Berengal wrote:Only if they're killer robots. Legos are happy robots. Besides, even if they were killer robots it wouldn't stop me. You can't stop science and all that.

ImagingGeek
Posts: 380
Joined: Fri Aug 06, 2010 6:41 pm UTC
Location: Canada

Re: Modern computing vs Biology

Postby ImagingGeek » Tue Aug 10, 2010 10:09 pm UTC

psykx wrote:I've heard it said a human brain woks at 40 Hrz, although I can't remember where from at the moment.


Not exactly. The frequency of brain waves, along with the pattern of pulses, is what carries the information. So its a little different from a computer with a set clockspeed. Brain waves like what you measure with an EEG represents an averaging of waves generated by individual neurons. These average waves (which represent a high-order degree of coordination between groups of neurons) do not represent the activity of individual neurons, or even the "cores" of our brains. There are 5 types of these waves, raging in frequency of ~1Hz upto 100Hz.

Individual neurons are much less coordinated - you can think of the brain waves above as sort of clocks that the neurons use as a reference. The actual potentials that travel from one neuron to another (and also inside of an isolated "core" of the brain) are very short - the pulses themselves average 5ms in length. The actual frequency of the waves vary - that is how information is carried, but normal frequencies are in the 50-80hz range. But some neurons are slow, emitting a pulse every couple of minutes. Others get beyond 100Hz.

Bryan
I have a new blog, on making beer.

Not that anyone reads it...

Hello1024
Posts: 9
Joined: Tue Aug 10, 2010 5:02 pm UTC

Re: Modern computing vs Biology

Postby Hello1024 » Fri Aug 13, 2010 3:36 pm UTC

To summarize:

Brain: more neurones - hence more parallel
CPU: More frequency - hence more calculation throughput per processing unit per second.

Brain: Unpredictable - errors\failures are common and tolerated
CPU: Deterministic - Errors are catastrophic (ie. usually cause complete failure of a task)

Brain: self configuring, using not entirely understood mechanisms
CPU: Preconfigured to do a particular task.

Both CPU's and brains appear limited in performance mainly by energy use/dissipation at the moment.

markop2003
Posts: 60
Joined: Sun Jun 06, 2010 4:21 pm UTC

Re: Modern computing vs Biology

Postby markop2003 » Sat Aug 14, 2010 3:01 pm UTC

There was a TED speech on this a while back which said that the spec sheet of the human brain matches up pretty well with the spec sheet of the internet.

We already have lots of specialized chips in PCs, not just GPUs but also memory controllers, disk controllers, RAID chips ect, you can even get these specialized CPUs that fit in a CPU socket but are optimized towards a certain task. Specialized chips are always going to be better than general chips for a given task however general chips are more flexible and so may give better overall performance as you don't have one half of the system idle whilst the other half is struggling.
Though i think things are actually going to head more general now with vertex and shader processors in GPUs being replaced with general stream processors several years ago and it's set to go more in this direction with AMD's plans for Fusion. Also it seems things are going to follow the multicore trend not only because of the heat issues but also because the vast majority of the cost of semi-conductors is in the design and it works out much cheaper to just add a second chip than designing a better one, we're already seeing this in high end GPUs and AMD's Bulldozer seems to be heading in this direction too. Eventually we'll end up with something like what Zii is trying to accomplish (though they currently seem to be failing commercially) where we'll have one general processor consisting of a bunch of general processing modules with the more powerful chips simply having more processing modules.

User avatar
Benzene
Posts: 18
Joined: Tue Jun 10, 2008 4:15 am UTC
Location: Your local gas station.

Re: Modern computing vs Biology

Postby Benzene » Sun Aug 15, 2010 6:15 am UTC

Too many misconceptions are drawn when people compare the brain to computers. On a very high level, they share some similarities, like electrical circuits to plumbing, but if you drill down even slightly, things can be taken the wrong way.

I laugh a little when I hear people try to state specific "frame rates" of eyes, as though the cells all sampled an image at the exact same time, then did nothing until however long later. While they do have a bandwidth (due to the time constant for the pigments to be converted between forms, nerves to repolarize, etc), they are the only continuous-time form of video I can think of, and thus have a near-infinite "frame rate"

Memory is another curious aspect of the mind; some piece on Radiolab discussed how when memories are recalled, they are destroyed (but remade elsewhere). When people with traumatic memories are administered with a drug that blocks memory formation and led to recall those memories, they are eroded (yes, Eternal Sunshine of the Spotless Mind has a large bit of truth behind it.)
--llo? What beep; I didn't hear anything? Is it even recording? (inaudible) Hey, how do I get this thing to--- *BEEP*

User avatar
PhoenixEnigma
Posts: 2303
Joined: Fri Sep 18, 2009 3:11 am UTC
Location: Sasquatchawan, Canada
Contact:

Re: Modern computing vs Biology

Postby PhoenixEnigma » Mon Aug 16, 2010 6:37 pm UTC

Benzene wrote:Memory is another curious aspect of the mind; some piece on Radiolab discussed how when memories are recalled, they are destroyed (but remade elsewhere). When people with traumatic memories are administered with a drug that blocks memory formation and led to recall those memories, they are eroded (yes, Eternal Sunshine of the Spotless Mind has a large bit of truth behind it.)

Interestingly, I do believe that's similar to how (most common forms of) RAM works, as well - reading is actually destructive (at a hardware level), and the page must be written back after every read.
"Optimism, pessimism, fuck that; we're going to make it happen. As God is my bloody witness, I'm hell-bent on making it work." -Elon Musk
Shivahn wrote:I am a motherfucking sorceror.

User avatar
negatron
Posts: 294
Joined: Thu Apr 24, 2008 10:20 pm UTC

Re: Modern computing vs Biology

Postby negatron » Wed Aug 18, 2010 5:34 pm UTC

Hello1024 wrote:CPU: Preconfigured to do a particular task.

Was yours preconfigured to browse xkcd? CPUs are "preconfigured" for universal computation. Even DSPs which are designed for a particular purpose are logically universal.

Computers do not need to be, and probably should never be, a dynamic network like a brain. Such an approach would prevent reliable processing, and although not critical for high performance computing tasks, it is critical for application logic.

Don't confuse software for the hardware that runs it. If you have a simulation of a brain in software it is entirely inconsequential how static the supporting hardware may be.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang


Return to “Hardware”

Who is online

Users browsing this forum: No registered users and 7 guests