Is information a physical quantity?

[lalbatros]

I have the feeling that yes, it is.
But what are the examples?
What are the foundations?
What are the derivations from physical laws to information theory?
What are the conditions?
What are the consequences?

Any book, reference?
Any suggestion?
Any comment, contribution, any thought?

Thanks to boost me on this topic,

Michel

 

[3trQN]

2 joules per bit comes to mind, don't know from where...

Maybe the more enlightened could erm... enlighten us.

 

[Loren Booda]

A black hole embodies the simplist physical example of information. Once it was accepted that black holes have entropy propotional to surface area, one could represent units of entropy topologically on the horizon of the hole. These bits of information were - in essence - reduced mass, charge and spin. There is still great controversy whether black holes eliminate information carried by quantum numbers other than those of mass, charge and spin.

 

[lalbatros]

Can physical interaction be associated with a flow of information?

Of course, it is the case when we ear sounds,
when fiber optics make systems interact at a distance and exchange information,
when bits flop in chain reactions in our computer,
...

The second principle of thermodynamics, I guess, tells us that information decreases, by the irreversible interactions.

Do the basic law of physics, say electrodynamics, lead us in some way to understand what information physically is and its importance that is so obvious in daliy life?

 

[tehno]

. There is still great controversy whether black holes eliminate information carried by quantum numbers other than those of mass, charge and spin.

I believe that problem is that they still don't know how preceisely to define it and what at all completely the physical information is.

 

[marlon]

I have the feeling that yes, it is.
What are the foundations?

Well, that ofcourse depends on what criteria (spin, momentum, quantumbits) you use to define information. Then, one needs to find mathematical expressions that link such criteria.


If we start from quantumbits, then indeed the answer is yes. I want to refer to the Shannon entropy and quantum error correction codes. Just take a look at the famous course of http://www.theory.caltech.edu/people/preskill/ph229/. You will be surprised how analoguous "quantum" thermodynamics and QIT can be.

regards
marlon

 

[lalbatros]

marlon,

Indeed, Preskil is quite interresting, I will take the time needed to read it.

But what strikes me now is the emphasis on quantum information.
I understand that information plays a big role in the understanding of fundamental aspects of QM, like the measurement 'problem', or probably the EPR paradox too. The potential applications are also motivating.

However, I would like to have an overview that -if this makes sense- covers the whole physics, avoiding a too strong attraction to these fascinating QM topics. Would you know of some such references?

In the limit, since physics is represented essentially by PDEs, I could even imagine that information theory had something to teach me about PDEs. I see at least an example from control theory and 'observability': parameters from a (process) model can be 'observed' with a precision determined by the properties of the model (and the measurements).

In a still farther limit, theories can be related to bits and bytes in their numerical version.

Finally, if I can get this (more classical) overview, I would be pleased to go back to QM: it seems to me that a specificity of QM is that it integrate information theory somehow, while classical physics don't.

Thanks for any suggestion, reading, papers, ...

And my best regards,

Michel

 

[marlon]

However, I would like to have an overview that -if this makes sense- covers the whole physics, avoiding a too strong attraction to these fascinating QM topics. Would you know of some such references?

What exactly are you saying here ?

Too me, it appears to be very logic that we use QM to "caracterize" information since we need to start from tiny bits to build up a bigger entity.

In the limit, since physics is represented essentially by PDEs, I could even imagine that information theory had something to teach me about PDEs.

Huh ? The PDEs are just a mathematical tool we use to describe reality, so...what do you want to say here ?

marlon

 

[lalbatros]

Marlon,

Indeed, I have some difficulties to express my questions correctly.

I am guessing (!) that a big difference between QM and CM is that the 'wavefunction' does not represent the reality but -somehow- the information that we can get on the physical world. I have read that here and there, but I did not take care for writing down the references (!).

For this reason I would like to concentrate my attention on how and why this transition from CM to QM could have brought this radical change. And therefore, I would like to start first by understanding (learning) the "physics of information" in the classical world. I guess again (!) that this transition is related to the unavoidable interaction necessary to observe a system. This makes interaction and information totally linked in contrast with CM.

Concerning PDEs. If you take the classical example of lossy transmission lines, it is clear that the capacity to transmit information depends on the distance, because of the losses. So, there is a link between the PDE description of a system and the "information transfer" inside the system. For sure this can also be the conclusion from a reductionist model of the transmission line, based on Maxwell's equations and the laws of motion.

But all this is (not too) easily said but I would need go from words to the language of mathematics and physics. And these are the kind of references I am looking for.

Thank already for 'Preskil',

Michel

 

[lalbatros]

Hello,

I just started the reading of "Science and Information Theory" by Leon Brillouin. I learned that Szilard showed in 1929 that the entropy of a unit of information was equal to k ln 2. (see http://www.dannen.com/szilard.html) .

Michel

 

[rcgldr]

A bit confused here, is there a physics based definition of "information" that differs from the more common abstract concept of information (like a thought process, image, feeling, memory, ... , in someones mind?).

 

[lalbatros]

Jeff,

The point made by Brillouin in his book and by other famous names in the 50's was precisely to restrict the concept of information to a scientific concept. Therefore it is all about measurable things, not about the content and its subjective meaning. Brillouin, in his introduction, precisely points out that an amount of information can have a high meaning for some people and no meaning at all for others.

However, I thing this is merely a practical limitation of physics. Human aspects are often too complex. It is clear that an information can also have a huge effect on a certain computer and no effect at all on another one, because of the configuration. In a sense this is computer-subjectivity. And on this computer toy-model, it is clear that the 'importance' of an information can also be quantified, just by the same tools as those devised in the 50's before computers where available everywhere.

Michel

 

[meemoe_uk]

Good question lalbatros.

>But what are the examples?
One pure link between infomation and reality is the uncertainty principle.
- Knowing extremely precisely what state a system is in now reduces the infomation you know about it in the future.

Quantum wave-particle duality,
-If you have no infomation on the state of a system, it's behaviour will be wave-like with respect to you.
- the more infomation you have on a system, the more particle like it and it's subcomponents will behave.

note that macro objects all have particle like behaviouor ( classical )because you and everything else in the environment has so much info on them ( i.e. countless photons are being bounced off 'em ).
It's only at the molecular size and less that objects can be considered wavelike. Yes, they are still interacting with the rest of the enviroment, but relative to the amount they interact internally ( i.e. electron with the nucleus ) they don't interact much, so there internals appear wavelike.

It's a general rule.
If the internal goings on of a closed system are not transmitted to the outside enviroment, they will appear wavelike.

(More interaction = more photons = more infomation transfer ) => less wavelike behaviour.

A single photon transfer will 'collapse' a wave into a particle.