Getting from cause effect to awareness

[nazz]

Thanks in no small part to the problems of the computer metaphor, people tend to think of memories as "stored" like they are in a computer. A memory for "dog" is stored in this sequence over here, a memory for "car" over here, etc. That's not the way that the brain works. Storage relies on a steady-state. A bit that is part of the representation of some variable must be either 0 or 1. If it changes, then that is no longer the same variable. In a computer, data is represented by the permanent (so long as the representation doesn't change) state of specific bits. In the brain, there is nothing akin to bits. That's because the representation is patterns of firing.

You may have heard of the Stroop test. Usually, it involves looking at words like red, blue, green, etc., with each word being a particular color:

The task is to give the color of the words not what the word says.
Let's imagine that we simply had some place in the brain where "blue" was stored. Then all we'd have to do is call up that memory. It shouldn't matter what the word spells. Only it does. That's because our information about the color depends on neural activity which corresponds to the perception of the color, not just the concept of the color or the color's spelling. In reaction time tests, people process abstract words faster if they just saw a corresponding item first such as first seeing an upward pointing arrow and then seeing the word "hope" because part of the way we represent "hope" is upward direction. By overlapping concepts like this, we basically "overlap" memories of concepts. The concept of "hope" is represented by certain patterns that also represent "up". At a very basic level, when certain features are presented together or represented in the brain by sensorimotor activation that occurs at the same time, they get represented by correlations between the firing of lots of neurons. In a very real way, we can store an unlimited number of concepts because of the way we are able to overlap the representation of information about these concepts. Mental connections are active neural connections.

To simplify, think of the way that conditioned learning works. A bell rings when the dog is presented with food. After a while, the same physiological response that the presentation of food results in occurs with the bell. There is "overlap" between how the brain represents the bell and food.

This is kind of the opposite of redundancy, which is why I'm not clear on what idav means by this.

That's probably a terrible explanation but usually it takes me several to get sufficiently non-technical and yet informative.

I don't know if it is a good explanation or not but I might not be getting it in either case. But then again, maybe I am. Because the only thing I could come up with in thinking about this was that a memory could be like a repeated neural pattern. So let's say "dog" has a specific neural signature in the brain. We could draw a diagram showing what neurons fired when we saw a dog and how they interconnected. So every time we see a dog the neural pattern is familiar to our conscious mind. It's the like the brain says "Oh yeah, I know this path. I've walked it before". When it is something new to us there is a different neural pattern that is established. But once that pattern has been established multiple times the same type of recognition occurs.

So the memory is not "stored" anywhere in anything analogous to how we think of things being stored. It's just a repetition of a familiar, recognizable pattern that has occurred before. The more that route is traversed the stronger, more durable the memory of it.

Is that anything close to what you are saying?

 

[LegionOnomaMoi]

So let's say "dog" has a specific neural signature in the brain.

It doesn't. It tends to involve particular neurons. But it is also context specific. The same neural regions that are involved when you say "that's a dog" while pointing to a dog will not be involved when you read the word "dog" in this thread. In fact, reading the word dog with quotes around it as in "dog" changes things. We tend to think of concepts as discrete entities. They aren't and they aren't represented as such in the brain. When you read the word dog, see a dog, say the word dog, there will generally be some neural activity that is similar in cases. But
1) that same activity will occur for other concepts
&
2) if we just said "that sort of activity is "dog" for you" we'd be wrong because limiting it to just that activity you'd be unable to think of the concept "dog".

We could draw a diagram showing what neurons fired when we saw a dog and how they interconnected

It's important to understand that the connections aren't so much between neurons as they are between firing patterns among neurons. neurons are constantly firing. The information is conveyed both in the rate of neural firing and in the ways that neurons fire in time with one another. Thus we frequently can't speak meaningfully of information conveyed by a single neuron, let alone a single action potential.

So every time we see a dog the neural pattern is familiar to our conscious mind. It's the like the brain says "Oh yeah, I know this path. I've walked it before". When it is something new to us there is a different neural pattern that is established. But once that pattern has been established multiple times the same type of recognition occurs.

There is something of this, yes. But a key thing is that the pattern isn't constant and isn't unique to "dog". Parts of that same pattern are there for cat, animal, treat, pet, fur, etc. That's what allows a kind of limitless memory.

The more that route is traversed the stronger, more durable the memory of it.

Sort of yes. Also key, though, is how the route is traversed in tandem with other things. One simplification might be that of a dictionary. When we don't understand what a word means, we look it up and we get synonymous words and phrases. We understand the new word only in relation to others. The closer two ideas are to one another the more they tend to overlap in the way they are represented in the brain. If I don't know about guns, the first time I hear about the IWS 2000 I won't be able to have the conceptual framework to really remember about it. But if I know about anti-materiel rifles, armor piercing sabot rounds, steyr, caliber, etc., I can associate the rifle with more. The more I can associate it with the better I can understand it because the "neural space" traversed to represent it overlaps with so much else.

Is that anything close to what you are saying?

Yes.

 

[idav]

Also key, though, is how the route is traversed in tandem with other things. One simplification might be that of a dictionary. When we don't understand what a word means, we look it up and we get synonymous words and phrases. We understand the new word only in relation to others. The closer two ideas are to one another the more they tend to overlap in the way they are represented in the brain. If I don't know about guns, the first time I hear about the IWS 2000 I won't be able to have the conceptual framework to really remember about it. But if I know about anti-materiel rifles, armor piercing sabot rounds, steyr, caliber, etc., I can associate the rifle with more. The more I can associate it with the better I can understand it because the "neural space" traversed to represent it overlaps with so much else.

It isn't just all over the brain though. Certain areas of the brain have certain functions and we all store certain numbers letters and colors in the same places. So its far from random.

 

[LegionOnomaMoi]

we all store certain numbers letters and colors in the same places. So its far from random.

None of us do. I've watched on neuroimaging studies which areas are active when words, numbers, objects, etc., are processed.

 

[idav]

None of us do. I've watched on neuroimaging studies which areas are active when words, numbers, objects, etc., are processed.

There are some commonalities and numbers are in ccertain regions vs. colors.

While it is extremely unlikely that any two synesthetes will report the same colors for all letters and numbers, studies of large numbers of synesthetes find that there are some commonalities across letters (e.g., "A" is likely to be red).

Grapheme

 

[LegionOnomaMoi]

There are some commonalities and numbers are in ccertain regions vs. colors.

There are. And these same regions are correlated with other concepts. Hence the overlap I was attempting to explain to nazz.

 

[PolyHedral]

Because there is no "object" and no semantics. All this is computer science 101. Computers are syntactic processors. We know how computers work. Programs are procedures. We store semantics and relations on a computer in a way that generates meaningful structures to us.

In my course, Computer Science 11001 (as it was designated) was the basics of Java programming, including the semantics of objects, instances and static data. Of course, they were teaching the semantics to the humans, and so didn't use that particular word, but that's what they explained - what Java programs mean, in terms of operations inside the JVM.

In the second semester, there was a module on basic data structures, like linked lists and maps. (I can't remember if we did trees then, or the next year.) Data structures, IMO, are the basic building blocks of semantics inside an OOP language - they define what is a meaningful operation.

What you're effectively doing is denying compilers and interpreters. You're saying that computer programs do not mean anything to the computer... and yet somehow, the compiler can resolve the names within the program into the abstract entities that those names represent. That's the definition of denotation. It's the exact same process I would go through if I needed to turn a Java program into JVM bytecode. Further, it represents this correspondance between names and entities in its memory.

However, I suspect I know the reason you're doing that. You're saying that computers don't understand anything and just process arithmetic because it's true - if and only if you look at the computer at the level of voltages in transistors. Individual neurons don't understand anything either. When you look at the machines in terms of the logic of higher-level operations, i.e. what needs to be true for specific JVM operations to work, and what becomes true afterwards, then you see the process we call "understanding" emerge - namely, knowledge extraction, deduction and abstraction. These two views are of the same thing, but they are not compatible with each other, and so we must carefully differentiate between them.

 

[LegionOnomaMoi]

In my course, Computer Science 11001 (as it was designated) was the basics of Java programming, including the semantics of objects

I've tutored in college courses on intro programming with java, subed in on teaching, and even sat in on one session for planning an intro OOP course. I also have about 100 textbooks relating to OOP. Yet I've yet to hear anybody but you refer to the semantics of objects in OOP.

Of course, they were teaching the semantics to the humans, and so didn't use that particular word, but that's what they explained

What did they explain?

what Java programs mean, in terms of operations inside the JVM.

Java programs mean one thing to a human and another to the things that run them. Can you specify what you are talking about?

IMO, are the basic building blocks of semantics inside an OOP language - they define what is a meaningful operation.

Meaningful to humans. Not computers.

What you're effectively doing is denying compilers and interpreters.

What I'm doing is so basic it is standard in any undergrad computer course. I've been asked to guest lecture on this. It's so basic that CS students get bored by it.

You're saying that computer programs do not mean anything to the computer

Duh.

... and yet somehow, the compiler can resolve the names within the program into the abstract entities

They can't. I'm surprised at you. Really surprised.

 

[PolyHedral]

I've tutored in college courses on intro programming with java, subed in on teaching, and even sat in on one session for planning an intro OOP course. I also have about 100 textbooks relating to OOP. Yet I've yet to hear anybody but you refer to the semantics of objects in OOP.

How many have you read? "Semantics" of programming is incredibly basic. It covers things like copy-by-value vs. copy-by-reference.

 

[nazz]

It's important to understand that the connections aren't so much between neurons as they are between firing patterns among neurons. neurons are constantly firing. The information is conveyed both in the rate of neural firing and in the ways that neurons fire in time with one another. Thus we frequently can't speak meaningfully of information conveyed by a single neuron, let alone a single action potential.

OK, so wouldn't that be a specific pattern then? If differentiation of perceptual/conceptual data is not tied to different patterns in the brain then what IS it based upon? There must be something otherwise how could we distinguish one thing from another?

There is something of this, yes. But a key thing is that the pattern isn't constant and isn't unique to "dog". Parts of that same pattern are there for cat, animal, treat, pet, fur, etc. That's what allows a kind of limitless memory.

I can't wrap my head around this. Written words are patterns. If I write "dog" it means "dog" (unless something in the surrounding context changes the meaning). It doesn't ever mean "cat" because that is a different pattern.

 

[LegionOnomaMoi]

How many have you read? "Semantics" of programming is incredibly basic. It covers things like copy-by-value vs. copy-by-reference.

Weisfeld, M. (2008). The object-oriented thought process. Pearson Education.
Beginning C# 3.0: An Introduction to Object Oriented Programming
Weiss, M. A., & Hartman, S. (2006). Data structures and problem solving using Java. Addison-Wesley.
Gold, J. (2004). Object-Oriented Game Development. Pearson Education.
Meyer, B. (1988). Object-oriented software construction New York: Prentice hall.
Beginning C Sharp 3.0: Introduction to Object Oriented Programming
Java, Java, Java - Object-Oriented Problem Solving
A Concise and Practical Introduction to Programming Algorithms in Java
Heaton, J. (2008). Introduction to neural networks for Java. Heaton Research, Inc..
Steeb, W. H. (2008). The Nonlinear Workbook: Chaos, Fractals. Neural Networks, Genetic Algorithms, Gene Expression Programming, Support Vector Machine, Wavelets, Hidden Markov Models, Fuzzy Logic with C++, Java and SymbolicC++ Programs:
Lau, H. T. (2010). A Java library of graph algorithms and optimization. CRC Press.
and more.

Please.

 

[LegionOnomaMoi]

OK, so wouldn't that be a specific pattern then? If differentiation of perceptual/conceptual data is not tied to different patterns in the brain then what IS it based upon?

The patterns aren't constant. That's the problem with brains. We overlap patterns to a significant degree which allows us so much but also means imperfection. Also, there isn't always a distinction between perception and conception. When we try to perceive we often mess up our conceptions.

There must be something otherwise how could we distinguish one thing from another?

We aren't always able to do this. The lines are fuzzy. However, we can usually do it because the overlap between firing patterns isn't complete. When I think of "dog" a lot of the areas involved overlap with, say, "cat". However, not all of the pattern overlaps. The difference allows me to distinguish.

Written words are patterns.

Try the stroop test. It's written words. Can you read of the colors the words are written in as easily as you can the words themselves?

 

[Slapstick]

If anyone wants to talk to a jabberwacky or Cleverbot they are both AI.

 

[nazz]

The patterns aren't constant.

Yes, if I am understanding correctly you are saying having the perception of a dog won't always involve the same set of neurons firing. So does that mean that in some cases that exact same set of neurons firing would cause the perception of a cat?

We aren't always able to do this. The lines are fuzzy. However, we can usually do it because the overlap between firing patterns isn't complete. When I think of "dog" a lot of the areas involved overlap with, say, "cat". However, not all of the pattern overlaps. The difference allows me to distinguish.

I'm not totally sure I understand what you mean by "overlap". Going back to words the pattern "dodge"' overlaps with the pattern "dog". It's just got a little more information to it that distinguishes it from "dog". Is that what you mean?

Try the stroop test. It's written words. Can you read of the colors the words are written in as easily as you can the words themselves?

No, it's confusing because I'm getting two sets of information at once.

 

[LegionOnomaMoi]

Yes, if I am understanding correctly you are saying having the perception of a dog won't always involve the same set of neurons firing. So does that mean that in some cases that same set of neurons firing would cause the perception of a cat?

Yes and no. There will be overlap.

I'm not totally sure I understand what you mean by "overlap". Going back to words the pattern "dodge"' overlaps with the pattern "dog".

It can but it is likely not to. Think of "overlap" more in terms of similar concepts not similar spelling.

No, it's confusing because I'm getting two sets of information at once.

Those "sets" involve overlap. You can't easily distinguish the color from what you are reading. Likewise, when you hear the word "flower" you are more likely to process words like "rose" "sunflower", even "red" faster then you are "block", "brick", "seasoning", etc.

 

[nazz]

Yes and no. There will be overlap.

As in cats and dogs are both animals, mammals, pets?

It can but it is likely not to. Think of "overlap" more in terms of similar concepts not similar spelling.

So like dog, canine, descended from wolves, that sort of thing?

 

[LegionOnomaMoi]

As in cats and dogs are both animals, mammals, pets?

More or less, yes. Overlap in what we think of concept typically translates to overlap in neural activation.

So like dog, canine, descended from wolves, that sort of thing?

Only if you know that sort of thing. Wolves and dogs don't necessarily overlap as much as dogs and cats. That's because dogs and cats are both pets, but not everybody sees them as such and not everybody who does sees them as belonging to the family of dogs vs. cats. Typically, people group cats and dogs closer (in terms of neural space) than dogs and wolves.

 

[idav]

Yes and no. There will be overlap.

So if a single nueron cant hold a memory how does the brain hold the memory, with the neural network?

Kinda reminds me of various RAID systems used in servers. One disk can go out and the lost data is retrieved by rebuilding from the remaining disks. which is what I mean by redundant, fault tolerance. But even worse like google servers sending the stuff overseas for redundancy and never knowing which server we will actually hit. More of your favorite type of comparisons.

 

[LegionOnomaMoi]

So if a single nueron cant hold a memory how does the brain hold the memory, with the neural network?

Correlations between neural firings.

Kinda reminds me of various RAID systems used in servers. One disk can go out and the lost data is retrieved by rebuilding from the remaining disks.

The idea of "going out" but still retaining function was the basis for the development of the internet by DARPA. It's not how the brain works. We have the opposite redundancy - overlap.

 

[nazz]

More or less, yes. Overlap in what we think of concept typically translates to overlap in neural activation.


Only if you know that sort of thing. Wolves and dogs don't necessarily overlap as much as dogs and cats. That's because dogs and cats are both pets, but not everybody sees them as such and not everybody who does sees them as belonging to the family of dogs vs. cats. Typically, people group cats and dogs closer (in terms of neural space) than dogs and wolves.

Well okay, I can grasp all of that. But it just seems where the patterns don't overlap that is where the distinction comes in. So there would be specific patterns for specific things with overlaps with similar things. But it seemed you were saying there was no specific pattern for a specific thing--"dog" might have one pattern one time and a different pattern the next and the pattern used for "dog" might be later used for something else--and that is what still confuses me.