Memory without trace

Another long one (100 pages) that I've been trying to make my through.

On the Possibility of Direct Memory
Unfortunately I can't get round to reading this immediately, but I am glad to see this thread revived. I hope he tackles the issue of computer memory vs human/animal memory. I think this is the crucial point - we can create things with 'memory traces' of a sort, but they don't seem to function like human memories.

David
 
Unfortunately I can't get round to reading this immediately, but I am glad to see this thread revived. I hope he tackles the issue of computer memory vs human/animal memory. I think this is the crucial point - we can create things with 'memory traces' of a sort, but they don't seem to function like human memories.

David

How do you think computer memory is different from "writing things down" memory? Does the paper remember what's written on it?

If not, why does a computer "remember" anything?
 
How do you think computer memory is different from "writing things down" memory? Does the paper remember what's written on it?

If not, why does a computer "remember" anything?
Well the difference is, of course, that software can interrogate a computer 'memory' and AI enthusiasts will claim that this is precisely analogous to what we do with our memories. I don't accept that, but it would be nice to see a crisp formulation of that point.

The previous article made the point about the open-ended use of memory well with Bethoven's Fifth, but it could still do with really pinning down.

David
 
Well the difference is, of course, that software can interrogate a computer 'memory' and AI enthusiasts will claim that this is precisely analogous to what we do with our memories. I don't accept that, but it would be nice to see a crisp formulation of that point.

The previous article made the point about the open-ended use of memory well with Bethoven's Fifth, but it could still do with really pinning down.

David

I guess I don't see what it is in software that has intrinsic meaning? In the end any program is a bit-string of 0s and 1s.

There's nothing in a machine that has fixed representation. It's a human that projects meaning onto the machines as the same saved bit-string can be compiled into different programs which do different things.

There seem to be utility arguments for fixing meaning (the program in connection with the world it's supposed to do things in), but if that's the case my fridge and car engine are also conscious entities?
 
How do you think computer memory is different from "writing things down" memory? Does the paper remember what's written on it?

If not, why does a computer "remember" anything?

I think you've got to improve the definition of what is meant by memory first... although a spatial pattern stored in a computer is no different from a spatial pattern stored on paper, the computer is allowed to manipulate both energy and matter using a very narrow window of regularities.Thus it can move these patterns around and manipulate them using a feed forward process that produces a result which is dislocated in space-time... i.e. some time has to pass, some energy has to be used, some space has to be crossed etc... What it can't yet do is directly listen to the same patterns of itself in the past, and sum these patterns to the present... i.e. coherently interfere with itself. You could think of a computer chip with a few more dimensions?

Until the hardware (spatial pattern matter) is made plastic, to accommodate interference (temporal pattern energy)... interference being the very thing that everybody currently want's to design out. We're not going to see the current type of computers do much more than they already do. One would have to free them to evolve, and give them a plastic substrate, to allow them to use fields, which would allow them to be interfered with, so that past states would directly interfere with their present states (but they would start interfering with each other too). But I suspect we're stuck in an age of rules and controls, a clockwork age where everything has to be to be rigidly controlled (including us), and the designers can't yet think beyond where we are either.
 
Last edited:
I'm not sure drawing parallels between computer memory and biological memory is useful as their functions are so different. The very features that make human memory what it is (inaccurate, imperfect and malleable) would render computers useless.
 
I'm not sure drawing parallels between computer memory and biological memory is useful as their functions are so different. The very features that make human memory what it is (inaccurate, imperfect and malleable) would render computers useless.
Interesting comment. There is a human ability which is much closer to that of the computer. But we don't call it 'memory'. I'm referring to our ability to read and write. In some circumstances, such as when doing a calculation, we use this in the same way as the computer. But in others, a few written words can remind us of some event or occasion, in this case the words are not the memory, but are either a key to an actual memory, or perhaps a trigger to a creative but fictionalised imagining of something that is not remembered.

So though reading and writing gives some small overlap with computer memory, it rapidly broadens into something very different. For example, sheet music can give a precise record of a piece. Yet each performance can differ, particularly where improvisation or expression is involved. In this case, spontaneity takes precedence over memory.
 
I'm not sure drawing parallels between computer memory and biological memory is useful as their functions are so different. The very features that make human memory what it is (inaccurate, imperfect and malleable) would render computers useless.
Well, Artificial Neural Nets (ANNs) are used for some tasks like face recognition, and they are supposed to be inaccurate, imperfect and malleable!

Generally speaking, they are still digital, but information is stored in floating point numbers. Alternatively, if they are analog, the data is already effectively in the form of floating point numbers.

David
 
While I don't believe computers remember anything I do see some usefulness in drawing a connection to human memory.

You have the hard disk where information is stored, but also the CPU cache that allows the computer to "remember" what is was doing mere moments ago.

Similarly, for a human every meaningful action involves memory of what happened a moment before. Speaking a sentence involves memory of the previous words is the obvious example.

This is why the question of where/if memories are stored ends up being so critical.
 
The idea of human memory, and for example the process of using human language may have value in its own right.

However, it risks discarding what it means to be a conscious, thinking, feeling being and focusing instead on the mere facade, the superficial processes of existence, such as eating or breathing. Words and language to me are a necessity of physical existence, much as are food and shelter. But I don't see the depth there, the thinking which goes on behind the construction of a sentence is in a language without words, ideas and thoughts flow rapidly in a rich inner world, which may have no relation to the mechanics of using language. If the topic is memory, then the memory used in to participate in this outer world may be something unlike that used for the inner experience, and there may not be any way to relate the two.
 
Atom by atom memory:

http://www.sciencemag.org/news/2016/07/scientists-make-single-atom-memory-copper-and-chlorine

In today’s issue of Nature Nanotechnology researchers report using a scanning tunnelling microscope (STM) to store data at the atomic scale.

Also here, with a little video:

http://www.livescience.com/55437-tiny-atomic-memory-device.html

A new "atomic memory" device that encodes data atom by atom can store hundreds of times more data than current hard disks can, a new study finds.

Now this is being hailed as possible big leap for technology etc, but it made me wonder if this begins to address the "memory trace" problem. Any thoughts?
 
It doesn't, unless someone can explain how the structure/materials of the storage exclusively determines the content.

Unless you're saying the Pattern has some intrinsic/ontological reality?
 
Does it need to have "a reality" beyond and outside the system that it's in?

Depends on what you mean by system?

You can grant Forms ontological reality without a Platonic Realm of Ideas, if that's what you mean.
 
Atom by atom memory:

http://www.sciencemag.org/news/2016/07/scientists-make-single-atom-memory-copper-and-chlorine



Also here, with a little video:

http://www.livescience.com/55437-tiny-atomic-memory-device.html



Now this is being hailed as possible big leap for technology etc, but it made me wonder if this begins to address the "memory trace" problem. Any thoughts?

It's still just a shared spatial pattern, as a way of moving access to information forward in time, so we can re-access again in the future... the actual processing of the spatial pattern would need to be temporal...

...well at least that is how we'd currently understand whatever is going on within what we call space-time.

You're going to have to get smaller than an atom, before we begin to see the interference effects that lie behind these stable patterns... I mean if they didn't have some stability across time, they would be useless for sharing and moving access to information forward in time within space-time.
 
Depends on what you mean by system?

You can grant Forms ontological reality without a Platonic Realm of Ideas, if that's what you mean.

Ok. We may be in agreement. I'm not saying this explains everything but atomic memory does massively explode the possible configurations and patterns available.

In terms of the more plastic nature of animal brains perhaps "familiarity" is a better word than "memory", which has come to mean something quite different to what we have. A familiarity trace may give a more realistic picture of our experience at the level of the pathways, structures, and now possibly, atoms involved
 
Last edited:
Ok. We may be in agreement. I'm not saying this explains everything but atomic memory does massively explode the possible configurations and patterns available.

In terms of the more plastic nature of animal brains perhaps "familiarity" is a better word than "memory", which has come to mean something quite different to what we have. A familiarity trace may give a more realistic picture of our experience at the level of the pathways, strucures, and now possibly, atoms involved

Maybe there's hope for homeopathy here.. :)
 
Back
Top