Experiments and information

[SuperSlim]

When I took Communication Theory (quite some years back), I learned to my surprise that information has entropy. A message that's been received has an information content, and an expectation that the information will be received.

Then I learned that sending a message has an expectation of being received; this is something that also intersects the domain of engineering--you want the sender and receiver to have a reliable channel and an agreed protocol. So that there is the connection: an experiment is in both domains, you need to do some engineering (after doing some design), then you get some information and then . . . analysis. You hand in the lab assignment and hope you get at least a C.

So where I'm trying to go is, when we do the experiment, what information isn't then in the analysis, what do we ignore, and does it matter? Why would or wouldn't that be true?

I know it sounds trivial, but so does the reason Maxwell's demon doesn't get to "see" gas molecules and violate the 2nd law (you know, the one that also says more or less, time doesn't go backwards).

 

Edited February 17, 2022 by SuperSlim

[studiot]

14 hours ago, SuperSlim said:

When I took Communication Theory (quite some years back), I learned to my surprise that information has entropy. A message that's been received has an information content, and an expectation that the information will be received.

Then I learned that sending a message has an expectation of being received; this is something that also intersects the domain of engineering--you want the sender and receiver to have a reliable channel and an agreed protocol. So that there is the connection: an experiment is in both domains, you need to do some engineering (after doing some design), then you get some information and then . . . analysis. You hand in the lab assignment and hope you get at least a C.

So where I'm trying to go is, when we do the experiment, what information isn't then in the analysis, what do we ignore, and does it matter? Why would or wouldn't that be true?

I know it sounds trivial, but so does the reason Maxwell's demon doesn't get to "see" gas molecules and violate the 2nd law (you know, the one that also says more or less, time doesn't go backwards).

 

 

Perhaps you have no takers, because, like myself, others are confused as to where you are coming from or going to with this question.

Please clarify.

[SuperSlim]

Suppose you're doing an experiment that involves the use of an electronic circuit.

Where do the electrons come from, and where do they go to? If there wasn't a "bath" of available electrons that can be pushed around a circuit then dumped back into the reservoir, electronics would be a lot more esoteric, like particle accelerator experiments with hadrons maybe.

Electrons are light particles and they're easy to gather together in classical amounts. The experiment focuses on the circuit and the voltages and currents in it; the fact there is a reservoir of available electric charge is sort of a given; ignoring it has no effect on the experiment.

But in quantum experiments, the electrons have to go back into the reservoir; if this didn't happen they wouldn't leave any marks on a screen, they wouldn't be detected.

I'm trying to examine this idea of having a reservoir, part of which is given a known state (230V, 60Hz), does work, then returns to the same unkown state it was in. In particular it seems to be an important detail in quantum experiments--particles that need to leave evidence behind also have to vanish, so nothing further can be known. What's the common denominator? Information has to be "written" somewhere, then it has to be "erased", at a cost.

Edited February 19, 2022 by SuperSlim

[SuperSlim]

I guess what I'd like to do is verify Landauer's principle holds for any experiment, including those that violate Landauer's limit for the erasure of information.

This erasure must be based on a choice of what the information is, this appears to be unavoidably related to a choice of gauge. That is, choosing where 0V is, is something you have a few degrees of freedom for in the average electronic LRC circuit, including those with active elements. This choice determines the values (the information content) of other voltages. A string of voltages/currents that add to zero should obey Kirchoff's laws (a linear algebra over V and i).

Edited February 19, 2022 by SuperSlim

[exchemist]

6 hours ago, SuperSlim said:

I guess what I'd like to do is verify Landauer's principle holds for any experiment, including those that violate Landauer's limit for the erasure of information.

This erasure must be based on a choice of what the information is, this appears to be unavoidably related to a choice of gauge. That is, choosing where 0V is, is something you have a few degrees of freedom for in the average electronic LRC circuit, including those with active elements. This choice determines the values (the information content) of other voltages. A string of voltages/currents that add to zero should obey Kirchoff's laws (a linear algebra over V and i).

You can do plenty of experiments that don't involve anything electrical at all. So it doesn't seem to me that trying to relate the information obtained from an experiment to voltages is a particularly helpful line of enquiry. Arfa?

Edited February 19, 2022 by exchemist

[SuperSlim]

5 hours ago, exchemist said:

You can do plenty of experiments that don't involve anything electrical at all.

Which experiments don't involve anything electrical? Or do you mean the ones that don't need to plug anything into a mains supply?

Quote

So it doesn't seem to me that trying to relate the information obtained from an experiment to voltages is a particularly helpful line of enquiry.

Unless the experiment involved the creation of some voltages, if that was the aim of the experiment, you mean? Say the experiment was about mixing reagents together and observing what happens. Say you also know what reagents are being mixed together.

How would you relate something like Kirchoff's laws to the chemistry in the reactions? Or would your opinion be it isn't something anyone would bother with?

Would someone who tries to formulate a linear algebra of reactions, be just getting lost, you think? You know how easy that is, sometimes people don't manage to connect a post with the one preceding it !! And it's from the same person, and it's about the same thing !!!

Edited February 19, 2022 by SuperSlim

[studiot]

There are lots of physical processes that are characterised by the expression A = Blog_p(C)
Where A and C are variables and B is a constant (which may incorporate a negative sign) and p is a given base for the logarithm.

Why are folks so determined to dream up a physical link beteen two such processes, one in thermodynamics and one in Information theory ?

Should we be adding say a link to chemical pH ?

[exchemist]

6 hours ago, SuperSlim said:

Which experiments don't involve anything electrical? Or do you mean the ones that don't need to plug anything into a mains supply?

Unless the experiment involved the creation of some voltages, if that was the aim of the experiment, you mean? Say the experiment was about mixing reagents together and observing what happens. Say you also know what reagents are being mixed together.

How would you relate something like Kirchoff's laws to the chemistry in the reactions? Or would your opinion be it isn't something anyone would bother with?

Would someone who tries to formulate a linear algebra of reactions, be just getting lost, you think? You know how easy that is, sometimes people don't manage to connect a post with the one preceding it !! And it's from the same person, and it's about the same thing !!!

I'm sure that after a moment's thought you will be able to think of plenty of experiments that don't involve electricity. 

Trying to apply Kirchoff's laws to an organic chemical synthesis, or to an experiment on the behaviour of insects, would strike me as pretty crazy, at any rate. 

[SuperSlim]

9 hours ago, exchemist said:

Trying to apply Kirchoff's laws to an organic chemical synthesis, or to an experiment on the behaviour of insects, would strike me as pretty crazy, at any rate. 

Why do you think that's what I'm laying out here? I was talking about an experiment with an electronic circuit. You've pointed out how in chemistry you can do non-electronic experiments. Something I know already, but, thanks anyway.

Obviously something has wooshed straight over that head of yours. I'm trying to examine what information there is in any experiment and how you would encode that information, not necessarily on paper or an electronic circuit. I know how scared some people are of Shannon entropy, but it is what it is.

So I point out something about how you can choose 0V in a circuit, and this doesn't seem to you, to be something you can do in every experiment. Well, you could be right about that.

Edited February 20, 2022 by SuperSlim

[SuperSlim]

13 hours ago, studiot said:

Why are folks so determined to dream up a physical link beteen two such processes, one in thermodynamics and one in Information theory ?

Look I'm sorry but that seems to be a mischaracterization. "The folks" are trying to explain why thermodynamics and information theory are connected.

Because they are, I mean it's just one of those things. Take information entropy and multiply it by the total energy of a system, then divide that by the average energy per particle, and you get a number that has no physical dimensions. It's a ratio.

So say information is also physical, it has the same kind of thermodynamic entropy when considered as a source of heat, surely?

In a computer, the information is a resource, so is heat but heat is a "waste product". Turn that around and consider the information being processed by a cyclic heat engine, and say what the computer is doing is wasting a resource. Why doesn't that work? Or maybe, it does work.

So that's the memo I guess. In any computation there are chosen information-bearing degrees of freedom, and so there are, although we almost don't bother with it, non-information-bearing degrees of freedom. I think this is some kind of fundamental principle, perhaps, so must apply to any computation.

If an experiment is such a thing, and I can't see how it isn't, there we have it. Except for the choice part because most of them are made for us.

Edited February 20, 2022 by SuperSlim

[exchemist]

2 hours ago, SuperSlim said:

Why do you think that's what I'm laying out here? I was talking about an experiment with an electronic circuit. You've pointed out how in chemistry you can do non-electronic experiments. Something I know already, but, thanks anyway.

Obviously something has wooshed straight over that head of yours. I'm trying to examine what information there is in any experiment and how you would encode that information, not necessarily on paper or an electronic circuit. I know how scared some people are of Shannon entropy, but it is what it is.

So I point out something about how you can choose 0V in a circuit, and this doesn't seem to you, to be something you can do in every experiment. Well, you could be right about that.

Well, Arfa, I freely admit I find you confusing, but then you do seem to be flip-flopping between an experiment in an electronic circuit and experiments in general. For instance, how am I to reconcile the two passages I have highlighted in red in your previous post?  

[studiot]

10 hours ago, SuperSlim said:

So say information is also physical, it has the same kind of thermodynamic entropy when considered as a source of heat, surely?

By all means demonstrate that this is the case but consider this

 

On 2/19/2022 at 12:36 AM, SuperSlim said:

Information has to be "written" somewhere, then it has to be "erased", at a cost.

So let us consider computer memory chip, A RAM chip.

So I agree it takes a specific amount of energy input to 'write' to one cell of the Ram chip.

But there are several ways to effect erasure, each with their own energy cost.

Note I have started a new thread, specifically to examine any correspondence between Shannon entropy and Thermodynamic entropy.

 

Edited February 20, 2022 by studiot

[SuperSlim]

2 hours ago, studiot said:

But there are several ways to effect erasure, each with their own energy cost.

Yes absolutely. Erasing information means you erase the information-bearing degrees of freedom of what has been defined (before anything is written or erased) as having those IBDGs. You map those to NIBDGs, also pre-defined. The mapping is always a physical, dissipative process.

12 hours ago, exchemist said:

Well, Arfa, I freely admit I find you confusing, but then you do seem to be flip-flopping between an experiment in an electronic circuit and experiments in general.

That's quite a coincidence. I find you confusing, but then you have jumped to the conclusion that since I started talking about an example--electronics--that's what I have to stay with, I can't generalise it. I can't talk about a chemistry experiment or coupled pendulums, or anything else, because I have to talk about electronics.

[exchemist]

24 minutes ago, SuperSlim said:

 

That's quite a coincidence. I find you confusing, but then you have jumped to the conclusion that since I started talking about an example--electronics--that's what I have to stay with, I can't generalise it. I can't talk about a chemistry experiment or coupled pendulums, or anything else, because I have to talk about electronics.

How do you propose to generalise from an electronic example to chemistry? And what is the principle, or insight, you want to take from electronics and apply to other cases? 

 

 

[SuperSlim]

32 minutes ago, exchemist said:

How do you propose to generalise from an electronic example to chemistry?

By finding a common algebra, preferably one that defines linear relations (such as the addition of voltages).

 

32 minutes ago, exchemist said:

And what is the principle, or insight, you want to take from electronics and apply to other cases? 

I believe it's called information. Or, what it is you know, and what it is you can't know. The reality of information.

[SuperSlim]

I think what the (so far, confusing) storyline might be, is that Baez conjectured whether a theory is like an algorithm, and an experiment is like  a path or trajectory through an (abstract) machine. I guess I've picked up the baton, here.

In any experiment, what can you write down that is representative of the state of a system (of, yknow, particles), and then what do you do with the recorded information? Just to throw a spanner in, what if someone else steals your notes and burns them, they're a pile of ashes now? (The machine has run an unexpected information-bearing algorithm).

I'll add the observation that Kirchoff's laws are based on ideal elements, they say nothing about temperature. An ideal resistor has a fixed resistance, a real physical resistor has a resistance that depends on temperature. The trick is to restrict the ambient temperature by actively or passively cooling the physics down.

So is there something like a "Kirchoff's algorithm" for chemistry? If you can define an abstract machine and paths through it, and algorithms, well, maybe.

Edited February 21, 2022 by SuperSlim

[studiot]

10 hours ago, SuperSlim said:

I think what the (so far, confusing) storyline might be, is that Baez conjectured whether a theory is like an algorithm, and an experiment is like  a path or trajectory through an (abstract) machine. I guess I've picked up the baton, here.

In any experiment, what can you write down that is representative of the state of a system (of, yknow, particles), and then what do you do with the recorded information? Just to throw a spanner in, what if someone else steals your notes and burns them, they're a pile of ashes now? (The machine has run an unexpected information-bearing algorithm).

I'll add the observation that Kirchoff's laws are based on ideal elements, they say nothing about temperature. An ideal resistor has a fixed resistance, a real physical resistor has a resistance that depends on temperature. The trick is to restrict the ambient temperature by actively or passively cooling the physics down.

So is there something like a "Kirchoff's algorithm" for chemistry? If you can define an abstract machine and paths through it, and algorithms, well, maybe.

Kirchhoff (with two h's and two f's) was a 'polymath', who also contributed to Chemistry, particularly in the domain of chemical thermodynamics.

And yes there is such a law, though it was not due to K but a gentleman named Hess.

Hesse's Law and Hess Cycles.

https://en.wikipedia.org/wiki/Hess's_law

+1 for interesting developments of your musings.

I have also mentioned them in my thread on entropy and information.

Edited February 21, 2022 by studiot

[Sensei]

On 2/19/2022 at 1:36 AM, SuperSlim said:

But in quantum experiments, the electrons have to go back into the reservoir; if this didn't happen they wouldn't leave any marks on a screen, they wouldn't be detected.

Electrons are accelerated by an external electric field and have a certain maximum kinetic energy that depends on the strength of the electric field. An electron hitting a screen loses its kinetic energy, the atom in the screen becomes excited, and as it transitions to ground state, it emits photons. The nearly motionless electron is then captured by the electrode.

On 2/19/2022 at 1:36 AM, SuperSlim said:

I'm trying to examine this idea of having a reservoir, part of which is given a known state (230V, 60Hz), does work, then returns to the same unkown state it was in.

230V (230 Vrms!) 60 Hz are properties of AC, not quantum states..

 

[SuperSlim]

21 hours ago, Sensei said:

Electrons are accelerated by an external electric field and have a certain maximum kinetic energy that depends on the strength of the electric field. An electron hitting a screen loses its kinetic energy, the atom in the screen becomes excited, and as it transitions to ground state, it emits photons. The nearly motionless electron is then captured by the electrode.

Are you trying to say the position and momentum of the electron that lost its kinetic energy to a screen is known? It's been captured?

 

21 hours ago, Sensei said:

230V (230 Vrms!) 60 Hz are properties of AC, not quantum states..

Captain Obvious rocks, eh?