Quantum information...

Date: Sat, 11 Nov 1995 03:24:26 -0800
From: Paul Wilkens <paul@reed.edu>
To: quantum-d@teleport.com
Subject: QUANTUM-D: Quantum information...

        these comments were stimulated by an excerpt from "Quantum 
communication moves into the unknown" by David Deutsch and Artur 
Ekert, which is available at

http://eve.physics.ox.ac.uk/QCresearch/communication/communication.html

	i'll first summarize the material, which discusses two
interesting possibilities. bandwidth could be doubled through the
process of communication by transmitting a single spin-half
particle. a particle, which represents one bit with two possible
states, can be be put into an "entangled" (sic) state with an isolated
reference particle. the receiver of the message then sends this
particle to the sender, who performs a polarization operation on it,
putting it into one of four states. when sent back to the receiver, it
can be compared to the reference particle to extract one of four
messages - two bits. two particles are needed actually, but only one 
is sent. 

	the article also mentions transmission of a particular
particle's state information (controverting the uncertainty principle)
by a similar scheme. this one is a bit less practical, ludicrously
difficult actually. Harald Weinfirter and Anton Zeilinger from the
University of Innsbruck are developing an implementation of the first
idea.

        meanwhile Rhett had asked

>           I have a question for you: can we think of a biological
> use or analogue of the quantum communication scheme that these
> people are writing about? if there is a way for secure or high
> density communication to happen across space, which uses unique 
> aspects of physics, then it seems it is not out of bounds that 
> natural selection might result in some biological utilization - 
> or do you disagree? i am not saying that this is a promising 
> problem in biology but just that it should be interesting to 
> evaluate the tangibility of physical information by situating 
> it that way.

	well, i can think of biological near-analogues, such as endocrine
systems that transmit information by modifying a hormone, released by
a gland, that then has a different effect on the receptive tissues and
a feedback effect on the gland that produced the initial unmodified
hormone, which in turn influences the production of modifying
factor. in a sense, this is a multiple party system which transmits
information in more than just the on-off (hormone present/not present)
state - it has three states, hormone present, not present, and present
in second form. unfortunately for this analogy, most cells
differentiate to pay attention to either form a or form b, reducing it
to binary. two things though - biological systems pay attention to
varying amounts, and the process of differentiation secures systems
against unwanted crosstalk.
	it seems the methods outlined in the paper are somewhat
awkward - each particle must be synced, sent, modified, received and
interpreted (bigger bandwidth but at least twice as slow) and for each
particle sent, you have a reference that must be kept in total
isolation! it's that reference particle that makes things really
tough. even with an organ that is receptive to quantum states, the
reference state cannot persist.
	binary encoding is, for the most part, unused in physiology. 
binary is an attempt to make communication as simple as possible, and
biology finds complexity useful. even in systems that systems
theorists would call binary, the communication lines are stochastic
and concentration-dependent.

	quantum computation is a far more interesting issue,
biologically. the most notable feature of neural processes is their
massive parallelism. the brain does nothing quickly (ms not ns speeds
predominate) but it does so much at once that the total represents
incredible processing power. in this system, any single participant is
unimportant, although it contributes to the outcome. in some sense,
you would expect the processing of quantum-level information to be
quite similar. therefore, individual "facts" are nonsensical, and the
brain must have a method of integrating many "facts" such that it can
produce a conception of what they might imply. not only that, but a
neuron is an isolated system, and communicates with other neurons
through electrochemical signals only (excluding gap junctions, which
aren't actually that prevalent in the brain). so for the brain to
react to a significant event, of which it has only quantum nonlocal
knowledge (say through the microtubule device) enough neurons have to
be excited by their individual microtubule networks to produce enough
signals to actually cause the brain to do something, and the brain has
to figure out what it's reacting to along the way. 
	it is considerably aided in this task by the individual nature
of the events responded to by the microtubule networks in each cell -
each neuron has a different set of reasons for firing. "tune" the
neurons (and their respective microtubule networks) to different
nonlocal stimuli (how would this be done?), and the pattern of firing
is interpretable by the brain in its capacity as a symbolic processor,
though it would have a tough time telling it from noise. even though
those events, individually, are silly and unimportant, enough of them
together could actually give information.
	maybe we should write a simulation to see if artificial
organisms can gain useful information from their environments if their
bit inputs are numerous but consistently trivial.


paul