Journal Club #6

"I don't want to be a pie. I don't like gravy."

Out of the 279 papers in the January 2009 quant-ph section of arxiv, three of them were downloaded onto my computer. The criteria for download being 1) I understand all the concepts/terms in the title, 2) I understand the first line of the abstract, and 3) I can see how it relates to my humble quest for 5-dimensional glory. It will become ridiculously apparent in just a few minutes why I don't attempt to discuss papers like this more often.

Of the three papers that I downloaded, the winner is a tasty morsel that grabbed me with the second line of its abstract - "The greater the information that is gained, the less reversible the measurement dynamics become." This sounds familiar, though I'm starting to forget exactly where I wrote about certain ideas... I'll repeat the gist of it. The more well-anchored an observation is in memory - that is, the more supporting and/or dependent observations that are also encoded in connection with an observation, and/or the more connections to previous memories that are generated with respect to a particular observation - the less-likely you are to be able to UNDO the observation.

With pen in hand, I delve into the introduction, hopelessly curious about the nature of the 'uncertainty' relation between reliability and reversibility that the authors will be deriving with respect to quantum measurements. "Can we find a useful role for the idea of dynamical reversibility in the context of quantum physics...?" The way I'm reading this, the question could be restated 'Can the presentation of a specific outcome state be undone, allowing the object to settle into the same quantum (smeared) state that it maintained prior to the measurement?' Presumably from there it could subsequently be observed in a different state.

It's immediately apparent in the following sentences that the aspect of information retention is causing a bit of a problem. "To measure is to create information; and information is a state - in a machine or an organism - which extends from a certain time into the future." This quotation is actually from Otto Frisch's 1965 paper "Take a Photon." (Brief pause while I google for a copy of this paper. Significant pause to appreciate the irony of finding, among the top google results, Raymer - one of the author's of the paper I'm currently reading - discussing rational atheism with John C. Garrison in online story posted today. Unable to find the Frisch article though.)

By page 3, the first few lines of the section titled 'Unitary Evolution', I'm lost. Measurement as a four-step process? What are 'kets' and 'qutrits'? Hey, isn't there a big game on that I could be watching?

I've now given up all hope of complete comprehension of this paper within a reasonable time frame, and am skimming the rest of it for key concepts that will enable me to have a punchy ending to this blog post. "A key element of our treatment is to consider... what constraints are placed on this reversibility by virtue of leaving a permanent trace (information) in the probe." If I were to argue that the only state of 'information' that matters is the state of the neurons post-observation related firing, then the question of reversibility becomes one of undoing or overriding the changes to the neurons as a result of observation-related firing. If we accept the experience of information becoming 'undone', then the most likely site for such an 'erasure' of information should be one that evidences the possibility of such modification. Such an erasure wouldn't be permanent, but would it be sufficient for experiencing inconsistent observations that suggest such an erasure? (Thusly do I retreat into the familiar to salve my wounded ego.)

"[T]he degree of reversibility decreases with each newly added observer k..." Ah, multiple observers! I understand this! But wait - why does the degree of reversibility decrease? Each observer is treated as "an additional probe oscillator coupled to the counter as in Figure 1." So, physical interaction with the system? Which leads me to wonder how exactly the authors are suggesting that the information from the system is dispersed across interactions with observers... I see shades of relative and cumulative entropy in this section.

"The very act of Bob deciding to remember permanently the result is sufficient to make the measurement irreversible... Furthermore, the greater the information that is gained by Bob... the less reversible the dynamics become." This is almost consistent with the concept of 'memory as an anchor.' "[O]ther observers necessarily change the state of the system, making it less amenable to reversal from Bob's viewpoint." This seems to be a definitive statement in support of the existence of multiple observers, whose existence and effects I have previously pondered.

Even though I'm missing out on a great deal of the detail/depth of this paper, I enjoyed being exposed to an attempt to model relationships that resemble (to me) the ones I have previously discussed. Perhaps you will see something in this paper that I missed.

(So, how was that for an attempt to leap interdisciplinary boundaries? ;)