Brain= ???? Bytes

For the discussion of the sciences. Physics problems, chemistry equations, biology weirdness, it all goes here.

Moderators: gmalivuk, Moderators General, Prelates

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Tue Jan 12, 2010 7:52 pm UTC

Did they actually say that was sufficient to be a complete person, or just that that's the amount of pattern storage that, when combined with the particular processor they were using, could allow for a person to be recreated?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Nath
Posts: 3148
Joined: Sat Sep 08, 2007 8:14 pm UTC

Re: Brain= ???? Bytes

Postby Nath » Tue Jan 12, 2010 8:40 pm UTC

gmalivuk wrote:Did they actually say that was sufficient to be a complete person, or just that that's the amount of pattern storage that, when combined with the particular processor they were using, could allow for a person to be recreated?

Depending on what you mean by 'sufficient to be a complete person', these could be equivalent. After all, information by itself cannot be a complete person. It needs to be decoded somehow and converted into a program. Effectively, the amount of information in your brain is the amount of information it takes to reconstruct you. This measure depends on the reconstruction program. In the extreme case, if you have a program that can simulate everybody in the world, the additional information is merely a few bits to tell the program who it's supposed to simulate. Or if you have a program that does nothing other than simulate you, the additional information in your brain is zero.

Probably a more intuitive measure of information in the brain is: what's the length of the shortest possible program that can simulate you (to the required level of accuracy)? But this is just another way of asking the same question. The length of the program depends on what programming language you use. If your programming language has a 'simulate gmalivuk' function, you're back to the situation described above.

As far as I can tell, there's no way around this. The amount of information in something always depends on what you already know.

psyck0
Posts: 1651
Joined: Wed Mar 19, 2008 5:58 pm UTC
Contact:

Re: Brain= ???? Bytes

Postby psyck0 » Tue Jan 12, 2010 8:40 pm UTC

poxic wrote:Just to clarify: I don't think there are "superhuman abilities" hiding underneath ordinary consciousness, just that we do notice and record more than we're usually aware of. In the article I linked to, the author went from stick-figure drawings of a cat to detailed sketches with facial expressions (also of a cat) during a TMS session. Not because he suddenly developed artistic ability, but because he became better able to accurately recall the details of cats he'd seen.

Think of how you can be unable to recall someone's face, no matter how hard you try, until you remember that you saw them at a party last week. Suddenly you see a scene from your friend's house, where the TV was, where your drink was sitting, and the person you were trying to remember who happened to be standing by the kitchen. All that information was stored but most of it wasn't ... indexed, if you will. (You couldn't have answered the question "where did you put your drink down when so-and-so was over by the kitchen?" until you remembered that scene.)


OK, yes, but there's a reason for that. You don't recall it because memory works as a network- you need to get to memory items from links from related memory items. Some items only have one link. In that sense, I'm not sure that anything is ever "erased" but the odds of accessing it are very low (and this is good because they are irrelevant clutter). Also, memory items decay and change if not used- there may not be an erasing procedure, but they do disappear. I'm certain that over half of the "vivid details" you recall in a scene are interpolated and even made-up. There was a famous study done on flash-bulb memories, memories people have of where they were when they heard a famous event- 9/11, pearl harbour, etc. These are no more accurate than any other memory, yet people are absolutely certain in them. There's a pretty good chance that your memory of what you were doing right before 9/11 is flat-out wrong. That shows how reliable memory is.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Tue Jan 12, 2010 8:54 pm UTC

psyck0 wrote:There's a pretty good chance that your memory of what you were doing right before 9/11 is flat-out wrong.

No, I'm quite certain I was asleep. But also I think 9/11 is too recent for this to actually be true. I may not remember all the details surrounding it, but I remain sure that I found out when I got an email from my mom telling me a plane had crashed into WTC. I can verify this by looking through old emails. And I suppose that's where it can become difficult to judge things like this: I know I *am* correct about some of those details, and can prove it, but it would be quite hard to determine whether that's because my "original" memory of the event is accurate, or instead because in this technological age it's easy enough to just go back and check if I want to.

A better example for me would be the Challenger explosion in early 1986. I do recall it happening, and watching news about it is my earliest datable* memory, but I am at the same time quite content to admit that much of the specific visual elements of that memory have been interpolated or added later or whatever.

* Approximately datable, anyway. I suppose I could be remembering news coverage the next day, though. Then again, my memory is of being at daycare, which with that particular babysitter meant that watching TV would have only been around lunchtime, which is when it happened, according to Wikipedia.

(In about 8 months, there will be people who are truthfully able to answer the COPPA question to join the forums, who were as old when 9/11 happened as I was when the Challenger blew up...)
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

psyck0
Posts: 1651
Joined: Wed Mar 19, 2008 5:58 pm UTC
Contact:

Re: Brain= ???? Bytes

Postby psyck0 » Tue Jan 12, 2010 11:44 pm UTC

There is a fantastic study about flashbulb memories that you should read: Weaver, C.A., III. (1993). Do you need a “flash” to form a flashbulb memory? Journal of Experimental Psychology: General, 122, 39–46.

He was having his psych 100 class do a memory test experiment thing by chance on the day the gulf war broke out, so he converted it to a study of the memories of students hearing about it. Very interesting results. Accuracy of recall 6 months later was no better than accuracy of recall of the control (they were instructed to remember the details around the next time they met their closest friend or roommate or something after that class), which was pretty poor.

So, I think that plenty of time has passed since 9/11.

Your memory really is a lot worse than you think it is, at least for events. That's why witness testimony is so unreliable (and on a scientific basis shouldn't be considered admissible in a trial).

User avatar
Rockberry
Posts: 109
Joined: Tue Jan 12, 2010 9:24 am UTC
Location: UK

Re: Brain= ???? Bytes

Postby Rockberry » Wed Jan 13, 2010 12:32 am UTC

gmalivuk wrote:Did they actually say that was sufficient to be a complete person, or just that that's the amount of pattern storage that, when combined with the particular processor they were using, could allow for a person to be recreated?


I think they meant the amount of information stored in the brain if I recall. And yes, I gather it was sufficient to create/copy a complete personality.

User avatar
Cobramaster
Posts: 201
Joined: Fri Nov 20, 2009 2:31 pm UTC
Location: Georgia Southern University Department of Chemistry, and Department of Biology.

Re: Brain= ???? Bytes

Postby Cobramaster » Wed Jan 13, 2010 12:47 am UTC

Ahh yes but that is merely the software on a persons brain memory is separate still.
SlyReaper wrote:Did you never notice the etymological link between "tyrannosaur" and "tyrant"? 1% of the dinosaurs had 99% of the prey. Occupy Pangaea.

Technical Ben
Posts: 2986
Joined: Tue May 27, 2008 10:42 pm UTC

Re: Brain= ???? Bytes

Postby Technical Ben » Wed Jan 13, 2010 12:51 am UTC

On the 200mb capacity thing... I would expect a high error margin on that one. Without a quote or example of how it was calculated I don't trust it. Why? Because a lot of the TV science is over simplified. And when it comes to simulating the human brain, a lot of short cuts are taken (reasonably so, as we cannot simulate one in real time currently).
So this means, little short cuts as "the human brain contains x amount of brain cells so must only be x size in bits". Or "the human brain recalls x memories so must have x bytes of storage". They seem to forget that each braincell has a synapse, each synapse is weighted etc.
If you can show me the calculations "X amount of brain cells, by Y amount of synapses, by the number of possible patterns within the brain..." to make sure you have not missed something, then I'd be happy to agree with it.
Also, if we calculate the power of a grid of 12 by 12 cells, can we extrapolate it? It follows exponential growth right?
It's all physics and stamp collecting.
It's not a particle or a wave. It's just an exchange.

RabidAltruism
Posts: 135
Joined: Fri Aug 29, 2008 9:25 pm UTC

Re: Brain= ???? Bytes

Postby RabidAltruism » Wed Jan 13, 2010 1:11 am UTC

Nath wrote:Probably a more intuitive measure of information in the brain is: what's the length of the shortest possible program that can simulate you (to the required level of accuracy)? But this is just another way of asking the same question. The length of the program depends on what programming language you use. If your programming language has a 'simulate gmalivuk' function, you're back to the situation described above.


Perhaps what we really want to find is Big Theta(Human Being)! :D

edit: I suppose an important question, in that case, would be: what could we mean by input size? Genome characteristics, environment? Implicit in looking for a Big O/Theta/Omega approach to this question seems to be a belief that some kind of input should be provided on which to base the simulation. I'm not sure whether that's a virtue or a vice..

I think we'll find ourselves running into pretty serious problems with individual differences here, also; simulating a newborn would no doubt require a far shorter program than simulating, say, Feynman (who would also be much more complicated than simulating me, but I think I'd beat the baby out).
...she reminds you of the invisible form of the soul; she gives life to her own discoveries; she awakens the mind and purifies the intellect; she brings light to our intrinsic ideas; she abolishes oblivion and ignorance which are ours by birth...

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Wed Jan 13, 2010 5:54 am UTC

psyck0 wrote:Your memory really is a lot worse than you think it is, at least for events.

Oh I know. (Wait, is that possible? Knowing that memory is worse than I think it is...?)

But part of the point I made with the 9/11 and Challenger things was that, especially in modern times, there is so much *other* information around that can be used to corroborate and flesh out memories, that even if the "original" has degraded, it's entirely possible for a pretty accurate "copy" or something to have taken its place. Not accurate in a lot of the details we think remain consistent over time, like where people were standing or what exact words someone said or what a person looked like. But accurate for things like I listed with those two memories.

I'm pretty sure I still have the email saved somewhere from 9/11 (though not on this computer so I can't check immediately), and I know it was a Tuesday so I know what time I would have woken up and started watching CNN on my computer. And it's quite probable that I was in fact sitting in my babysitter's kitchen watching TV when the Challenger exploded, because I know that it happened around lunchtime.

I think in general it's important to look more closely at precisely what kinds of details we are good and bad at remembering when it comes to flashbulb memories. Because I know that these two as well as a lot of other memories I have really amount to nothing more than a single mental image combined with a couple facts about day of the week or whatever, but from there I can often pretty accurately work out a lot of the other details deductively, even if I'm not really "remembering" them, whatever that means technically.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Cobramaster
Posts: 201
Joined: Fri Nov 20, 2009 2:31 pm UTC
Location: Georgia Southern University Department of Chemistry, and Department of Biology.

Re: Brain= ???? Bytes

Postby Cobramaster » Wed Jan 13, 2010 2:46 pm UTC

My 9/11 memory is relatively complete as far as what happened events wise and what I saw, the audio is gone and replaced by a rough transcript since my brain is awesome like that, what I remember was sitting in Shop class working on a draft for a project when an announcement was made over the intercom, from there I went to English where we started to watch the news but we were prevented from going outside since we were on the world's largest military base and were considered a primo target.
SlyReaper wrote:Did you never notice the etymological link between "tyrannosaur" and "tyrant"? 1% of the dinosaurs had 99% of the prey. Occupy Pangaea.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Wed Jan 13, 2010 4:33 pm UTC

Cobramaster wrote:My 9/11 memory is relatively complete as far as what happened events wise and what I saw, the audio is gone and replaced by a rough transcript since my brain is awesome like that, what I remember was sitting in Shop class working on a draft for a project when an announcement was made over the intercom, from there I went to English where we started to watch the news but we were prevented from going outside since we were on the world's largest military base and were considered a primo target.

Yeah, I think that sort of general memory (where you're not claiming to remember what people said or what they were wearing or where they were standing or whatever) is rather more robust than the little details. Especially, as I said, when objective information exists (like what day and time something happened, and what class you would have been in at that time) which you can continue to use to reinforce what you remember.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Telchar
That's Admiral 'The Hulk' Ackbar, to you sir
Posts: 1937
Joined: Sat Apr 05, 2008 9:06 pm UTC
Location: Cynicistia

Re: Brain= ???? Bytes

Postby Telchar » Wed Jan 13, 2010 7:24 pm UTC

The problem that many models have run into is that in the hippocampus, there are layers of neurons that will keep firing after they have been stimulated and will occasionally fire in reverse order. They also have a rather unique structure which is radically different than most neural striata.

The other problem one would run into in terms of weighting is the placement of the synapse on the neuron. There are several types of neurons that will synapse directly onto an axon, or onto the cell body. There are still others that have directly electrical synapses which bypass neurotransmitters all together. I don't know much about the AI field, but it'd be interesting to know how they account for these variations.
Zamfir wrote:Yeah, that's a good point. Everyone is all about presumption of innocence in rape threads. But when Mexican drug lords build APCs to carry their henchmen around, we immediately jump to criminal conclusions without hard evidence.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Wed Jan 13, 2010 7:53 pm UTC

Someone wrote in to ask essentially this question on the latest SGU podcast. Basic reply was that we really have no idea, because the brain works so vastly differently from a digital computer, being more analog, for one thing. In addition to that, there's the massively parallel stuff that goes on in the brain.

It's easy to ask how much information a digital rendition of something has, in a very technical sense of information, because you can just look at the filesize. But what about a painting? How much information does the Mona Lisa have? In a strictly digital sense, it would seem to be for all practical purposes infinite, since you can discover strictly more detail in higher and higher resolution digital images of it. But which details are "important"? Different answers to that question, I think, are somewhat analogous to the different answers you can give to how much information the brain holds. Do you need to describe every atom? How about every molecule? Can you get by describing how blocks of cells are connected, or do you need to model each cell separately along with all of its connections to other cells? How many of the differences between possible networks of connections can be described as "information" the brain holds?

I suspect a lot of these questions will run into the problem of vagueness pretty quickly. To the point where we'll probably never get better than a few orders of magnitude in our estimates of how much information a human brain holds.

(And I also suspect that attempts to "upload" onto silicon hardware will run into some of the same issues as digitizing things like paintings: perfect transfer is probably impossible, so how much loss and compression is acceptable?)
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

psyck0
Posts: 1651
Joined: Wed Mar 19, 2008 5:58 pm UTC
Contact:

Re: Brain= ???? Bytes

Postby psyck0 » Wed Jan 13, 2010 9:04 pm UTC

Even very basic memories can be wrong. Weaver says he remembers hearing about Pearl Harbour on the radio as an announcement that interrupted a baseball broadcast he was listening to. He believed that for decades. Then he realised that Pearl Harbour was in December and no baseball was being played at that time.

I'm not saying YOUR memory is wrong, merely giving more examples.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Wed Jan 13, 2010 9:13 pm UTC

psyck0 wrote:He believed that for decades.

Was this also independently verified, or does he merely "remember" having believed that for decades?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Nath
Posts: 3148
Joined: Sat Sep 08, 2007 8:14 pm UTC

Re: Brain= ???? Bytes

Postby Nath » Wed Jan 13, 2010 10:35 pm UTC

Telchar wrote:The other problem one would run into in terms of weighting is the placement of the synapse on the neuron. There are several types of neurons that will synapse directly onto an axon, or onto the cell body. There are still others that have directly electrical synapses which bypass neurotransmitters all together. I don't know much about the AI field, but it'd be interesting to know how they account for these variations.

Standard neural network models in AI do not account for this sort of thing. Ever since AI and cognitive science went their separate ways, the biologically-inspired models used in AI have been drifting away from the real thing. It turns out we can build better problem-solving systems if we don't get too caught up in making them work the same way we do.

psyck0
Posts: 1651
Joined: Wed Mar 19, 2008 5:58 pm UTC
Contact:

Re: Brain= ???? Bytes

Postby psyck0 » Thu Jan 14, 2010 1:13 am UTC

gmalivuk wrote:
psyck0 wrote:He believed that for decades.

Was this also independently verified, or does he merely "remember" having believed that for decades?

I assume that was a joke (ha, ha, rather clever).

User avatar
Josephine
Posts: 2142
Joined: Wed Apr 08, 2009 5:53 am UTC

Re: Brain= ???? Bytes

Postby Josephine » Thu Jan 14, 2010 1:16 am UTC

psyck0 wrote:
gmalivuk wrote:
psyck0 wrote:He believed that for decades.

Was this also independently verified, or does he merely "remember" having believed that for decades?

I assume that was a joke (ha, ha, rather clever).

probably not. valid question.
Belial wrote:Listen, what I'm saying is that he committed a felony with a zoo animal.

psyck0
Posts: 1651
Joined: Wed Mar 19, 2008 5:58 pm UTC
Contact:

Re: Brain= ???? Bytes

Postby psyck0 » Thu Jan 14, 2010 2:45 am UTC

Of course it's not independently verified. He could be lying for all I know. It's a personal anecdote that I find humorous, given his position in research into flashbulb memories.

User avatar
joshz
Posts: 1466
Joined: Tue Nov 11, 2008 2:51 am UTC
Location: Pittsburgh, PA

Re: Brain= ???? Bytes

Postby joshz » Thu Jan 14, 2010 2:48 am UTC

By "independently verified," he meant that someone else recorded in the past that he believed that, not just that he thinks he believed that.
You, sir, name? wrote:If you have over 26 levels of nesting, you've got bigger problems ... than variable naming.
suffer-cait wrote:it might also be interesting to note here that i don't like 5 fingers. they feel too bulky.

User avatar
RockoTDF
Posts: 582
Joined: Thu Oct 18, 2007 6:08 am UTC
Location: Tucson, AZ, US
Contact:

Re: Brain= ???? Bytes

Postby RockoTDF » Thu Jan 14, 2010 5:03 am UTC

psyck0 wrote:I have some background in cognitive psychology. This is what I have learned in my schoolings about memory:

1) Each memory is represented by an activation pattern of every neuron in the brain (or a subset) represented in that activation. It is complex to describe which neurons are involved since some regions are clearly not involved with memory at all, and other regions are more involved with particular types of memories, so I'll just go with every neuron in the brain for simplicity of discussion.


It's a subset. There isn't a normal state where every neuron is talking to every other neuron. The overall brain state (caused by fatigue, being drunk, etc) does matter, however. This is called state dependent memory.

3) Memory is HOLOGRAPHIC. This is the kicker, and I don't fully understand it. The basic idea is that each memory is added to each other memory by adding the activation patterns using matrix algebra, with the net result of a single activation pattern that represents all of our memories. Because of this, our memory capacity is infinite. Additionally, because memory is holographic, memories are stored throughout the brain, not in one location. Just like a real hologram, removing a chunk of the brain only makes the memories more fuzzy, it doesn't remove any of them (taking out a chunk of a hologram makes the image fuzzy, but doesn't remove any part of the image).


Removing pieces of the brain will remove memories. Whether or not they are removed or inaccessible is another story (or one that is entirely unnecessary if you think of brain states/memories as dynamic).

gmalivuk wrote:
poxic wrote:There's some evidence that we do record everything but don't usually recall most of it. I just spent a couple of minutes searching for a reference to a documentary I once saw, but came up empty. It interviewed a young autistic woman who, when her mother gave her the past year's calendar, filled in every single day with what had happened, who had visited, the weather...

How is that evidence that we record everything and not, say, that this particular autistic woman recorded everything?


We don't record everything. Our memory is controlled by attention, which is very different in an autistic person. That and we would need to very what these autistic people say to make sure it is true.

Nintendon't wrote:From my understanding (at least what I learned in first year psychology) is that the brain is largely reconstructive, rather than an exact storage place for memories. Especially when considering false memories, the forgetting of memories, and such.


Yep. The brain rebuilds memories from a variety of things, and this can account for all that stuff you listed. It also explains how one can implant false memories in another by suggestion.

Telchar wrote:The problem that many models have run into is that in the hippocampus, there are layers of neurons that will keep firing after they have been stimulated and will occasionally fire in reverse order. They also have a rather unique structure which is radically different than most neural striata.

The other problem one would run into in terms of weighting is the placement of the synapse on the neuron. There are several types of neurons that will synapse directly onto an axon, or onto the cell body. There are still others that have directly electrical synapses which bypass neurotransmitters all together. I don't know much about the AI field, but it'd be interesting to know how they account for these variations.


The hippocampus is likely consolidating memories, ie refiring recently fired neurons so that their connections strengthen. The differences in structure have been modeled successfully.

The AI field generally doesn't take too much biology into account. However, in these discussions computational neuroscience is often forgotten. Comp neuro has managed to model the hippocampus (in fact, I'm working on updating an established model at the moment) as well as other areas while taken biology into account. Sometimes it is not too important to worry about every last ion channel or conductance, which is often the case when attempting to model larger brain systems.
Just because it is not physics doesn't mean it is not science.
http://www.iomalfunction.blogspot.com <---- A collection of humorous one liners and science jokes.

User avatar
gmalivuk
GNU Terry Pratchett
Posts: 26767
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

Re: Brain= ???? Bytes

Postby gmalivuk » Thu Jan 14, 2010 12:51 pm UTC

psyck0 wrote:Of course it's not independently verified. He could be lying for all I know.

My point wasn't that he might have been lying, but rather that his memory of a memory is probably orders of magnitude less reliable than his memory of an incredibly important flashbulb event. So unless he wrote down the belief about baseball decades ago, I'm less inclined to believe that he actually had that memory for decades.
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

User avatar
Xanthir
My HERO!!!
Posts: 5413
Joined: Tue Feb 20, 2007 12:49 am UTC
Location: The Googleplex
Contact:

Re: Brain= ???? Bytes

Postby Xanthir » Thu Jan 14, 2010 2:06 pm UTC

gmalivuk wrote:
psyck0 wrote:Of course it's not independently verified. He could be lying for all I know.

My point wasn't that he might have been lying, but rather that his memory of a memory is probably orders of magnitude less reliable than his memory of an incredibly important flashbulb event. So unless he wrote down the belief about baseball decades ago, I'm less inclined to believe that he actually had that memory for decades.

To make gmal's post clearer, he thinks it's possible the dude may have accidentally fabricated the memory of having that memory for decades. He may have actually had that memory for only a relatively short time, having fabricated it recently, and then to maintain some internal consistency he also fabricated the memory of it being an old memory.
(defun fibs (n &optional (a 1) (b 1)) (take n (unfold '+ a b)))

User avatar
Telchar
That's Admiral 'The Hulk' Ackbar, to you sir
Posts: 1937
Joined: Sat Apr 05, 2008 9:06 pm UTC
Location: Cynicistia

Re: Brain= ???? Bytes

Postby Telchar » Thu Jan 14, 2010 2:51 pm UTC

RockoTDF wrote:Removing pieces of the brain will remove memories. Whether or not they are removed or inaccessible is another story (or one that is entirely unnecessary if you think of brain states/memories as dynamic).


Unless you're talking about procedural memory. There was an early psychologist that did cortical cuts in mice to prove it. Can't remember the name.

The hippocampus is likely consolidating memories, ie refiring recently fired neurons so that their connections strengthen. The differences in structure have been modeled successfully.


But how? How can you model mossy neurons when their lack of nodes and myelinization is so unique but not due to a lack of evolution as in other animals, but because it is required for function?

The AI field generally doesn't take too much biology into account. However, in these discussions computational neuroscience is often forgotten. Comp neuro has managed to model the hippocampus (in fact, I'm working on updating an established model at the moment) as well as other areas while taken biology into account. Sometimes it is not too important to worry about every last ion channel or conductance, which is often the case when attempting to model larger brain systems.


At least in the case of the hippocampus, it's more than taking ion channels into account. It's completely unique function and structural organization.
Zamfir wrote:Yeah, that's a good point. Everyone is all about presumption of innocence in rape threads. But when Mexican drug lords build APCs to carry their henchmen around, we immediately jump to criminal conclusions without hard evidence.

User avatar
RockoTDF
Posts: 582
Joined: Thu Oct 18, 2007 6:08 am UTC
Location: Tucson, AZ, US
Contact:

Re: Brain= ???? Bytes

Postby RockoTDF » Thu Jan 14, 2010 5:22 pm UTC

Telchar wrote:
RockoTDF wrote:Removing pieces of the brain will remove memories. Whether or not they are removed or inaccessible is another story (or one that is entirely unnecessary if you think of brain states/memories as dynamic).


Unless you're talking about procedural memory. There was an early psychologist that did cortical cuts in mice to prove it. Can't remember the name.


I'm pretty sure you mean he did cortical ablations (ie removing parts of the brain) in mice, and that they could still find their way around a maze. This is one of those cases where there are far more studies of humans with brain damage that lose episodic memory that are more relevant to the current conversation. Procedural memory requires so many parts of the brain and body to work in sync that one cannot expect to merely ablate one region and lose a skill altogether (doing it as well as before is another story, though).


The hippocampus is likely consolidating memories, ie refiring recently fired neurons so that their connections strengthen. The differences in structure have been modeled successfully.


But how? How can you model mossy neurons when their lack of nodes and myelinization is so unique but not due to a lack of evolution as in other animals, but because it is required for function?


One does not always need to take mylination into account. One can construct easily make a model with the systems-level differences in hippocampal organization that can explain results of past experiments (or brain damaged patients) as well as make predictions for future experiments or what will happen when subregion X is damaged. The idea that one must model every last detail (which is computationally intractable) to have a good model is probably one of the biggest misconceptions about computational neuroscience.


The AI field generally doesn't take too much biology into account. However, in these discussions computational neuroscience is often forgotten. Comp neuro has managed to model the hippocampus (in fact, I'm working on updating an established model at the moment) as well as other areas while taken biology into account. Sometimes it is not too important to worry about every last ion channel or conductance, which is often the case when attempting to model larger brain systems.


At least in the case of the hippocampus, it's more than taking ion channels into account. It's completely unique function and structural organization.


Loads of articles here:
http://psych.colorado.edu/~oreilly/pubs-online.html

and here:
http://www.science.mcmaster.ca/Psycholo ... tions.html

Some of them are about cellular level stuff/neurogenesis, others are about systems level modeling.
Just because it is not physics doesn't mean it is not science.
http://www.iomalfunction.blogspot.com <---- A collection of humorous one liners and science jokes.

Outchanter
Posts: 669
Joined: Mon Dec 17, 2007 8:40 am UTC

Re: Brain= ???? Bytes

Postby Outchanter » Fri Jan 15, 2010 2:22 am UTC

Nath wrote:What this means is that the actual memories are stored in the weights of the neurons. (And the topology, but that tends to be relatively static).

How static is the topology of human brains though? I had the impression that connections appeared and disappeared all the time.

And I don't know how often they're used, but there are ANN algorithms that change the topologies of their networks. I recently read a neat article on AI for Go (the boardgame) which uses SANE, an algorithm for coevolving network topologies together with neurons, to train computer players.

User avatar
Telchar
That's Admiral 'The Hulk' Ackbar, to you sir
Posts: 1937
Joined: Sat Apr 05, 2008 9:06 pm UTC
Location: Cynicistia

Re: Brain= ???? Bytes

Postby Telchar » Fri Jan 15, 2010 9:02 am UTC

RockoTDF wrote:I'm pretty sure you mean he did cortical ablations (ie removing parts of the brain) in mice, and that they could still find their way around a maze. This is one of those cases where there are far more studies of humans with brain damage that lose episodic memory that are more relevant to the current conversation. Procedural memory requires so many parts of the brain and body to work in sync that one cannot expect to merely ablate one region and lose a skill altogether (doing it as well as before is another story, though).


The comparison to complex systems is odd. I can not remove your liver and expect your body to survive. I can not remove your fuel pump and expect your car to run. There are certainly similarities, I could remove a spark plug, but I think that particular comparison is rather misleading.


One does not always need to take mylination into account. One can construct easily make a model with the systems-level differences in hippocampal organization that can explain results of past experiments (or brain damaged patients) as well as make predictions for future experiments or what will happen when subregion X is damaged. The idea that one must model every last detail (which is computationally intractable) to have a good model is probably one of the biggest misconceptions about computational neuroscience.


What metric do you use to measure success? When you say you can successfully model results of similar brain damage or normal brain function, are you talking about firing patterns? General activity levels? Can you induce amnesia? Can you damage the POA and have sexual side effects? I'm just curious how detailed the metric for success is.



Loads of articles here:
http://psych.colorado.edu/~oreilly/pubs-online.html

and here:
http://www.science.mcmaster.ca/Psycholo ... tions.html

Some of them are about cellular level stuff/neurogenesis, others are about systems level modeling.


I'll try and read up in the meantime. Thanks!
Zamfir wrote:Yeah, that's a good point. Everyone is all about presumption of innocence in rape threads. But when Mexican drug lords build APCs to carry their henchmen around, we immediately jump to criminal conclusions without hard evidence.

User avatar
RockoTDF
Posts: 582
Joined: Thu Oct 18, 2007 6:08 am UTC
Location: Tucson, AZ, US
Contact:

Re: Brain= ???? Bytes

Postby RockoTDF » Sat Jan 23, 2010 8:34 am UTC

Telchar wrote:
One does not always need to take mylination into account. One can construct easily make a model with the systems-level differences in hippocampal organization that can explain results of past experiments (or brain damaged patients) as well as make predictions for future experiments or what will happen when subregion X is damaged. The idea that one must model every last detail (which is computationally intractable) to have a good model is probably one of the biggest misconceptions about computational neuroscience.


What metric do you use to measure success? When you say you can successfully model results of similar brain damage or normal brain function, are you talking about firing patterns? General activity levels? Can you induce amnesia? Can you damage the POA and have sexual side effects? I'm just curious how detailed the metric for success is.


Success here is defined as being able to replicate human behavioral results - ie list learning, task performance, etc. You have a good model if it succeeds when humans do, fails when humans do (such as brain damage) and can make predictions for future studies.
Just because it is not physics doesn't mean it is not science.
http://www.iomalfunction.blogspot.com <---- A collection of humorous one liners and science jokes.

Wummi
Posts: 14
Joined: Thu Jul 02, 2009 8:38 pm UTC

Re: Brain= ???? Bytes

Postby Wummi » Thu Feb 11, 2010 3:36 pm UTC

even on a super simple model, comparison gets quite messy

tanking on a 100 cell / bit model,

i challenge you, to memorize this:

Spoiler:
1111111111100000000110010010011000000001100000000110000000011010000101100111100110000000011111111111


and to memorize this:

Spoiler:
Image


both contain 100 bits of information.

in fact, both spoilers contain the _same_ 100 bits of information

one quite easy to reproduce from memory accuratly, the other, not so much.

User avatar
Meteorswarm
Posts: 979
Joined: Sun Dec 27, 2009 12:28 am UTC
Location: Ithaca, NY

Re: Brain= ???? Bytes

Postby Meteorswarm » Fri Feb 12, 2010 2:16 am UTC

Wummi wrote:even on a super simple model, comparison gets quite messy

tanking on a 100 cell / bit model,

i challenge you, to memorize this:

Spoiler:
1111111111100000000110010010011000000001100000000110000000011010000101100111100110000000011111111111


and to memorize this:

Spoiler:
Image


both contain 100 bits of information.

in fact, both spoilers contain the _same_ 100 bits of information

one quite easy to reproduce from memory accurately, the other, not so much.


Sure, but the picture triggers the brain's "It's a face" module, which is extremely specialized, so it's hardly a fair comparison. Regardless, simple models of the brain fail all over. Why can we memorize stuff better in 7-digit or fewer chunks, even if it's the same information? Countless examples of non-linear responses abound.
The same as the old Meteorswarm, now with fewer posts!

User avatar
Tass
Posts: 1909
Joined: Tue Nov 11, 2008 2:21 pm UTC
Location: Niels Bohr Institute, Copenhagen.

Re: Brain= ???? Bytes

Postby Tass » Fri Feb 12, 2010 5:39 pm UTC

If I didn't notice that it made such an easy picture then I would convert the first sequence into: 11.8.2.2.1.2.1.2.2.8.2.8.2.8.2.1.1.4.1.1.2.2.4.2.2.8.11 and then use a mnemonic to convert it into pictures and situations. Quite doable (I was actually about to take up your challenge, but then I clicked the second spoiler), but in the process I would ironically make it quite a bit more than 100 bits in order be able to remember it.

User avatar
Josephine
Posts: 2142
Joined: Wed Apr 08, 2009 5:53 am UTC

Re: Brain= ???? Bytes

Postby Josephine » Sat Feb 13, 2010 3:23 pm UTC

So, what does that mean? Do ANNs create unnecessarily large 'files' to store small amounts of data?
Belial wrote:Listen, what I'm saying is that he committed a felony with a zoo animal.

MarvinM
Posts: 95
Joined: Tue Jul 29, 2008 11:45 pm UTC

Re: Brain= ???? Bytes

Postby MarvinM » Sun Feb 14, 2010 4:07 pm UTC

There are two results from psychology that I remember, probably from over 15 years ago, based on memorisation experiments. I can't cite them. One experiment resulted in an estimation of the capacity of long term memory at 1010bits. The other I'm fairly sure from a seperate study was an estimation of the rate data enters long term memory at about 2 bits per second roughly constantly, suggesting we store around 109 to 1010 bits over a typical human lifetime.

Once we get over how insulting these numbers seem, accepting 1011 neurons in a human brain the estimates seem quite reasonable.

How much data we'd need to simulate a human brain at the neural level is a somewhat different question and could easily be 7 orders higher.

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

Re: Brain= ???? Bytes

Postby negatron » Wed Feb 17, 2010 2:31 pm UTC

http://www.fhi.ox.ac.uk/Reports/2008-3.pdf

This should be reference material for anyone who wants reasoned estimates with explanation. It's a 130 page report on brain emulation.

The most plausible estimates seem to be 5 x 10^13(50 Terabytes) to 10^16 (10 Petabytes).

It also goes into processing requirements for the brain and even virtual environments. It also makes cost estimates for this computational capacity, which I find quite poor, but I digress.
Image
I shouldn't say anything bad about calculus, but I will - Gilbert Strang


Return to “Science”

Who is online

Users browsing this forum: No registered users and 16 guests