rjo_, several years ago, I saw an article somewhere in which some scientists used a high power microscope to look closely at a single brain cell. What they saw was a string of between 4 and 8 (I can't remember exactly) proteins switching on and off like transistors. Do you know what I'm talking about? I cannot remember where I saw it, but the implications were amazing.
deSilva said...
Oops, Mark, where have you got your notions from of what happens inside a transistor???
I'm just quoting what I remember from the article. They mentioned how it looked like the proteins were switching on and off. I understand how a transistor works.
Has anyone read "On Intelligence" by Jeff Hawkins (inventor of Palm Pilot)? Interesting way of thinking of how the brain might work. Makes a lot of sense intuitively. The theory is that the brain is basically a simulator which constantly adjusts the simulations it runs based on differences between the simulation and resulting inputs from your senses. There is a lot more detail in the book, but that's the gist.
It's interesting to note that Jeff Hawkins went into business making Palm Pilots for the sole purpose of making enough money so he could spend it studying the human brain.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Post Edited (Ken Peterson) : 10/1/2007 1:13:39 AM GMT
Intelligence isn't about what you know, it's about how you learn. Although intelligent people tend to know a lot because of how they learn, knowledge isn't intelligence.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Transistors are magic. There's little gremlins living inside, but they're rather lazy. You need to zap them pretty hard in their Bottoms so that they Carry the Electrons. It's rather simple, really - once you know what's going on.
Also, the Prop reminds me of the Transputer because they are both interesting, and could be used for both low level and high level processing. I know they're not really similar >_<
A bipolar transistor is basically three semiconductors in a stack, with the center piece rather thin. In the case of an NPN transistor, the top and bottom pieces are N-doped silicon, and the middle is P-doped silicon. The middle piece is the base, and the top and bottom are the collector and emitter, respectively.
If you put a positive voltage on the collector and a negative voltage on the emitter, the C-B PN junction is reverse biased because the positive lead is on the N-type semiconductor. The B-E junction is forward biased, but there's no current flow because of the reverse-biased C-B junction. This somewhat resembles two diodes in series, with opposing polarities.
Now, put a positive voltage on the base, and you get current through the B-E junction, which fills the middle P-doped silicon with charge carriers. If you get enough charge carriers, you can turn the P-doped silicon piece into a conductor. When that happens, current can now flow from the collector to the emitter!
How's that for an explanation without pictures??
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
It is quite fine, Ken, but it is more or less a metaphore.
Or in other words: It leaves as many questions as it answers..
What is X-doped silicon?
What is the effect of different doping strengths?
How thin ist "rather thin" for the base?
Why has it to be thin in the first place?
Why is this "current amplification" so precise? And yet so different for each single transistor?
What is the cause for the "forward voltage"?
Why can't you simply exchange collector and emitter?
OK, deSilva...you talk as if you know the inner workings of a transistor and the rest of us are hopelessly ignorant. Unless you're designing transistors, it's usually good enough to understand the implications of the information readily available in the data sheets. Otherwise we can debate the inner workings of the Propeller and how an "AND" instruction is actually carried out in the multitude of logic gates within.. For all practical purposes related to the Propeller, a transistor is just a switch.
It's not a matter of who has the intelligence, it's a matter of who has the time.....I, for one, barely have enough time to play with the propeller at all, and I cherish that time tremendously!
OK.....back to Propeller Supercomputing!!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
I'm not sure what's used these days, but boron and phosphorus are very nice dopants that come from adjacent columns of the periodic table from silicon. I forget which is which, but one has one less valence electrons (P dopant) than silicon and the other has an extra valence electron (N dopant). Because of the way the dopant interacts with the crystal structure of the silicon (or germanium or whatever), the extra electron is loosely bound and can move freely through the material. If there's one less electron, an adjacent atom lends its electron and that "hole" is what effectively wanders through the crystal.
The "forward voltage" drop represents the difference in the energy levels of the electrons on the two sides of the junction due to the use of the dopants. It's the contribution made to the kinetic energy of the electrons to enable them to slip through to the corresponding conduction energy level on the opposite side of the junction.
These days you can't just switch emitter and collector because the physical construction (mostly dimensions) of the base-emitter and base-collector junction is somewhat different. "Back in the old days", the two junctions were made very similarly and you could swap them. Once they started making planar transistors, that all changed. The transistor will still work, but the characteristics will be different (gain, impedance, maximum current, etc.)
Wow! 3 answers just off the top of my head. I did a science project in high school over 40 years ago on the construction of a junction transistor (in my kitchen no less!). I met a couple of folks at Bell Labs who were doing development of the LED, particularly green ones. They had made a kit for students to make solar cells (for science projects) and I wanted to see if I could make a transistor the same way ... just diffuse two successive junctions into the same area on a piece of silicon. I needed to use both a P and an N dopant and get the timing and concentration (let alone the furnace temperature) right.
Sorry ... it's down memory lane time.
Post Edited (Mike Green) : 10/19/2007 12:33:19 AM GMT
Your explantion is really a model for what is happening, works well when you are thinking about how to use a transistor but it is not really an explanation, I think this is why deSilva refers to it as a metaphor. You have to just take it as a fact and use it.
The details deSilva talks about are not rocket science but they are not easily expressed in a few words (an animation could do it pretty well though) if it was perhaps we would have been treated to them rather than a list of questions.
Anyway back to supercomputing, oh, no one was actually talking about that [noparse]:)[/noparse]
Just saw your post... I've been slightly deflected for the last couple of days.
I didn't see the article, but I would be interested if you can figure out where you saw it.
Ken,
So, what happened? He invented the Palm Pilot... made gobs of money and then figured out that he needed a security clearance?
If he really wants to study the brain, he ought to go to Moscow.
Rich
Much of what we would like to know is already known, and if someone wanted us to know it, we would know it already... and yet, still we exist.
(We didn't win the Cold War... the Soviet Union died laughing.)
Ken,
I haven't read Hawkin's book yet, but the idea of the brain constantly running simulations and comparing them to the outside world is much more real than you might think. There are some good studies in the last year or so looking at blood flow (a proxy for brain activity) in areas of the brain during certain activities. For example, areas of the brain associated with movement of a finger become active while you're thinking about doing something that might involve movement of the finger, before you actually decide to do the movement.
I happen to have an interest in acupuncture and there are points on the external ear that are related to pretty much the whole rest of the body. When you stick an acupuncture needle in the spot on the ear associated with the foot, then the area of the brain associated with the foot shows activity.
Come to think of it, I might know what you are talking about.
You might be talking about an experiment using radio-labeled calmodulin (or other ion binding proteins) ... calmodulin can be made to act pretty much like a transistor in quasi-cyclotronic fields. Abe Liboff is the American expert and created something of a stir when he published his data, resolving a long standing controversy between Blackman and another fellow, whose name escapes me. "News" of such "transistor" events is probably signaled broadly on a higher frequency signal, which is generated as a part of the "transistor event."
Lednev covered Liboffs outburst with a parametric resonance model, which seemed to quiet everything down for a while. And to seal the deal Blackman and Blanchard at the EPA did some superb "obfuscationatory" work on neurite outgrowth (which was chillingly close to stating a labratory model of the autistic brain defect.)
It certainly does appear that autism is an environmental problem... and I remain absolutely charmed that work done by our best people, paid for by tax dollars... and reported by our Federal agencies got buried. And since my son has an autism varient, I've more than interested in seeing what I could do to unclog the bottleneck.
Keeping a lid on this kind of information is possible... but takes a lot of work. We still do it because apparently our final analysis has remained essentially unchanged. The Russians thought we were all nuts. The older I get, the more respect that I have for our decision making... but the final verdict is still out there.
rjo_: can you elaborate on what you know about autism being related to the environment? I understand that an otherwise perfectly normal child can develop it after a few years. On the face of it, this might seem off topic, but we are talking about supercomputing aren't we?
BTW: I think my Propeller is showing some signs of autism. It keeps doing the same thing over and over...with meticulous precision.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Once, as a kid, a group of us had a talk about that. Came down to knowing what one knows. More importantly, what the scope of unknowns are.
When we model something, we make associations. Without these, memory is difficult to impossible. Try this with a friend: Take some word that is just a combination of common phonetics: grindlespoink, darglefoog, etc... say it to them, then continue the conversation. Have that conversation be about memory and associations. Most all people, after an hour or so, cannot remember the word, yet will recall the entire conversation, where it occured, etc...
If you repeat the process, only this time, linking that word to something, the result will be much better. One such interesting link is the conversation about memory and association itself! The average person, having gone through this cycle, will realize they need to make some effort to associate that random set of sounds, and will literally develop an improved ability to do it! That ability will be retained, if used from time to time. It's learning about learning in action.
When we model something internally, we are associating rules, behaviors to other known rules and behaviors. If you think about it, there are really very few NEW rules and behaviors. Most are common across a lot of things.
It all starts out when we are young. We mimic things so as to better grok them. What we experience literally becomes part of who we are and how we think. "Play" is a very adaptable form of learning. Most of the senses are involved and links are rich and often deep.
This is normally triggered by new things. Young people, when exposed to new things, have no model to work from, so they "play" with it, building links and associations to their existing body of abstractions and observations.
As we age, we see fewer and fewer new things, thus we "play" less, having aquired and modeled most of the things necessary for our given environment.
When the urge to "play" strikes, I tend to cultivate it and exploit it, because learning is happening, and it's the kind of learning that is extremely valuable.
Where scope is concerned, modeling helps! If we have explored and modeled something successfully, and we encounter something else, that appears to fit the models we have, we grasp, at a low level, roughly what it's scope might be. Essentially, we look for those places where the model breaks, than store that delta, not the entire process.
At the end of the day, the pool of rules is quite small compared to the very high number of abstractions and associations we maintain to identify things and their relationships.
Most of us have core paths we depend on. They might be visual, kinesttic, audio, tactile, etc... Others can be learned, but one has to go back to "play" mode for them to really take root and be useful. There is a reason people engage in the fine arts. This is to improve on their general cognitive ability. it's more than raw stimulation of the neurons, it's actually linking processes together and learning to apply them to what is experienced...
Core things, dance, acting, painting, singing, moving, drawing, writing, etc... all involve complex paths. All also share some core rules that, when associated and reinforced, open the doors for more diverse kinds of sensory input to make better sense and be more easily processed / integrated into the body of understanding we all leverage each day.
(the doctor has left the building now --sorry all, it's a topic I've always found interesting.)
So, if we were modelling this... it would look like an FET controlling something that looks like an FM signal... and on the other end you would need a rather complex filter.
I guess what I am trying to say is... if I were trying to build the computer analog of the human brain, I wouldn't worry about flops... I'd think more in terms of frequency synthesis and bandwidth... These can be thought of in terms of flops... but only if you have a serious error in your thinking.
Are, now deSilva is on safe ground again!
So I agree: Talking of computer "performance" in terms og FLOPS is like talking of a persons "peformance" in terms of his bank account. It is "per se" nonsense, but not generally sneered at
The computer is only as good as the software, no matter how many FLOPS you have.
This has been an interesting discussion about thought and the mind. I have noticed over and over again that you remember things more easily if you can recognize it as fitting a pattern or model of something you already understand. That is why it's so important to understand what you learn and not memorize it. The things I only memorized in school are of little use to me now.
I'm just curious...seems like we have a lot of doctors participating in this forum. How many here are doctors? I'm an electrical/mechanical engineer if anybody cares.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
I'm not a doctor! That was self-deprecating humor!
That childhood conversation has stuck with me the entire time! Over the years, I've explored this topic, during idle moments, just out of raw, morbid interest. Comes with wanting to know how things work --including people!
Comments
I'm just quoting what I remember from the article. They mentioned how it looked like the proteins were switching on and off. I understand how a transistor works.
Very few do..
It's interesting to note that Jeff Hawkins went into business making Palm Pilots for the sole purpose of making enough money so he could spend it studying the human brain.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Post Edited (Ken Peterson) : 10/1/2007 1:13:39 AM GMT
"Intelligence is not what you do when you know what to do. Intelligence is what you do when you don't know what to do."
He may have been quoting someone more famous, but I've not heard this elsewhere.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Intelligence isn't about what you know, it's about how you learn. Although intelligent people tend to know a lot because of how they learn, knowledge isn't intelligence.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Its complicated and I couldn't describe the function in detail if you held a gun to my head, but its just a switch.
Also, the Prop reminds me of the Transputer because they are both interesting, and could be used for both low level and high level processing. I know they're not really similar >_<
no it's not but it can be used like one, but it can also be used like a lot of things!
If you put a positive voltage on the collector and a negative voltage on the emitter, the C-B PN junction is reverse biased because the positive lead is on the N-type semiconductor. The B-E junction is forward biased, but there's no current flow because of the reverse-biased C-B junction. This somewhat resembles two diodes in series, with opposing polarities.
Now, put a positive voltage on the base, and you get current through the B-E junction, which fills the middle P-doped silicon with charge carriers. If you get enough charge carriers, you can turn the P-doped silicon piece into a conductor. When that happens, current can now flow from the collector to the emitter!
How's that for an explanation without pictures??
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Or in other words: It leaves as many questions as it answers..
What is X-doped silicon?
What is the effect of different doping strengths?
How thin ist "rather thin" for the base?
Why has it to be thin in the first place?
Why is this "current amplification" so precise? And yet so different for each single transistor?
What is the cause for the "forward voltage"?
Why can't you simply exchange collector and emitter?
It's not a matter of who has the intelligence, it's a matter of who has the time.....I, for one, barely have enough time to play with the propeller at all, and I cherish that time tremendously!
OK.....back to Propeller Supercomputing!!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
The "forward voltage" drop represents the difference in the energy levels of the electrons on the two sides of the junction due to the use of the dopants. It's the contribution made to the kinetic energy of the electrons to enable them to slip through to the corresponding conduction energy level on the opposite side of the junction.
These days you can't just switch emitter and collector because the physical construction (mostly dimensions) of the base-emitter and base-collector junction is somewhat different. "Back in the old days", the two junctions were made very similarly and you could swap them. Once they started making planar transistors, that all changed. The transistor will still work, but the characteristics will be different (gain, impedance, maximum current, etc.)
Wow! 3 answers just off the top of my head. I did a science project in high school over 40 years ago on the construction of a junction transistor (in my kitchen no less!). I met a couple of folks at Bell Labs who were doing development of the LED, particularly green ones. They had made a kit for students to make solar cells (for science projects) and I wanted to see if I could make a transistor the same way ... just diffuse two successive junctions into the same area on a piece of silicon. I needed to use both a P and an N dopant and get the timing and concentration (let alone the furnace temperature) right.
Sorry ... it's down memory lane time.
Post Edited (Mike Green) : 10/19/2007 12:33:19 AM GMT
Your explantion is really a model for what is happening, works well when you are thinking about how to use a transistor but it is not really an explanation, I think this is why deSilva refers to it as a metaphor. You have to just take it as a fact and use it.
The details deSilva talks about are not rocket science but they are not easily expressed in a few words (an animation could do it pretty well though) if it was perhaps we would have been treated to them rather than a list of questions.
Anyway back to supercomputing, oh, no one was actually talking about that [noparse]:)[/noparse]
Graham
edit : not addressing Mike
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Just saw your post... I've been slightly deflected for the last couple of days.
I didn't see the article, but I would be interested if you can figure out where you saw it.
Ken,
So, what happened? He invented the Palm Pilot... made gobs of money and then figured out that he needed a security clearance?
If he really wants to study the brain, he ought to go to Moscow.
Rich
Much of what we would like to know is already known, and if someone wanted us to know it, we would know it already... and yet, still we exist.
(We didn't win the Cold War... the Soviet Union died laughing.)
I haven't read Hawkin's book yet, but the idea of the brain constantly running simulations and comparing them to the outside world is much more real than you might think. There are some good studies in the last year or so looking at blood flow (a proxy for brain activity) in areas of the brain during certain activities. For example, areas of the brain associated with movement of a finger become active while you're thinking about doing something that might involve movement of the finger, before you actually decide to do the movement.
I happen to have an interest in acupuncture and there are points on the external ear that are related to pretty much the whole rest of the body. When you stick an acupuncture needle in the spot on the ear associated with the foot, then the area of the brain associated with the foot shows activity.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
Come to think of it, I might know what you are talking about.
You might be talking about an experiment using radio-labeled calmodulin (or other ion binding proteins) ... calmodulin can be made to act pretty much like a transistor in quasi-cyclotronic fields. Abe Liboff is the American expert and created something of a stir when he published his data, resolving a long standing controversy between Blackman and another fellow, whose name escapes me. "News" of such "transistor" events is probably signaled broadly on a higher frequency signal, which is generated as a part of the "transistor event."
Lednev covered Liboffs outburst with a parametric resonance model, which seemed to quiet everything down for a while. And to seal the deal Blackman and Blanchard at the EPA did some superb "obfuscationatory" work on neurite outgrowth (which was chillingly close to stating a labratory model of the autistic brain defect.)
It certainly does appear that autism is an environmental problem... and I remain absolutely charmed that work done by our best people, paid for by tax dollars... and reported by our Federal agencies got buried. And since my son has an autism varient, I've more than interested in seeing what I could do to unclog the bottleneck.
Keeping a lid on this kind of information is possible... but takes a lot of work. We still do it because apparently our final analysis has remained essentially unchanged. The Russians thought we were all nuts. The older I get, the more respect that I have for our decision making... but the final verdict is still out there.
BTW: I think my Propeller is showing some signs of autism. It keeps doing the same thing over and over...with meticulous precision.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
When we model something, we make associations. Without these, memory is difficult to impossible. Try this with a friend: Take some word that is just a combination of common phonetics: grindlespoink, darglefoog, etc... say it to them, then continue the conversation. Have that conversation be about memory and associations. Most all people, after an hour or so, cannot remember the word, yet will recall the entire conversation, where it occured, etc...
If you repeat the process, only this time, linking that word to something, the result will be much better. One such interesting link is the conversation about memory and association itself! The average person, having gone through this cycle, will realize they need to make some effort to associate that random set of sounds, and will literally develop an improved ability to do it! That ability will be retained, if used from time to time. It's learning about learning in action.
When we model something internally, we are associating rules, behaviors to other known rules and behaviors. If you think about it, there are really very few NEW rules and behaviors. Most are common across a lot of things.
It all starts out when we are young. We mimic things so as to better grok them. What we experience literally becomes part of who we are and how we think. "Play" is a very adaptable form of learning. Most of the senses are involved and links are rich and often deep.
This is normally triggered by new things. Young people, when exposed to new things, have no model to work from, so they "play" with it, building links and associations to their existing body of abstractions and observations.
As we age, we see fewer and fewer new things, thus we "play" less, having aquired and modeled most of the things necessary for our given environment.
When the urge to "play" strikes, I tend to cultivate it and exploit it, because learning is happening, and it's the kind of learning that is extremely valuable.
Where scope is concerned, modeling helps! If we have explored and modeled something successfully, and we encounter something else, that appears to fit the models we have, we grasp, at a low level, roughly what it's scope might be. Essentially, we look for those places where the model breaks, than store that delta, not the entire process.
At the end of the day, the pool of rules is quite small compared to the very high number of abstractions and associations we maintain to identify things and their relationships.
Most of us have core paths we depend on. They might be visual, kinesttic, audio, tactile, etc... Others can be learned, but one has to go back to "play" mode for them to really take root and be useful. There is a reason people engage in the fine arts. This is to improve on their general cognitive ability. it's more than raw stimulation of the neurons, it's actually linking processes together and learning to apply them to what is experienced...
Core things, dance, acting, painting, singing, moving, drawing, writing, etc... all involve complex paths. All also share some core rules that, when associated and reinforced, open the doors for more diverse kinds of sensory input to make better sense and be more easily processed / integrated into the body of understanding we all leverage each day.
(the doctor has left the building now --sorry all, it's a topic I've always found interesting.)
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Propeller Wiki: Share the coolness!
So I agree: Talking of computer "performance" in terms og FLOPS is like talking of a persons "peformance" in terms of his bank account. It is "per se" nonsense, but not generally sneered at
This has been an interesting discussion about thought and the mind. I have noticed over and over again that you remember things more easily if you can recognize it as fitting a pattern or model of something you already understand. That is why it's so important to understand what you learn and not memorize it. The things I only memorized in school are of little use to me now.
I'm just curious...seems like we have a lot of doctors participating in this forum. How many here are doctors? I'm an electrical/mechanical engineer if anybody cares.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
The more I know, the more I know I don't know.· Is this what they call Wisdom?
That childhood conversation has stuck with me the entire time! Over the years, I've explored this topic, during idle moments, just out of raw, morbid interest. Comes with wanting to know how things work --including people!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Propeller Wiki: Share the coolness!
Post Edited (potatohead) : 10/19/2007 3:30:43 PM GMT