Hacker News new | past | comments | ask | show | jobs | submit login
What does it mean for light to be stopped or stored? (askamathematician.com)
108 points by ColinWright on July 30, 2013 | hide | past | favorite | 18 comments



I think that was the easiest to understand/intro to quantam information handling that I have read.

I also enjoyed how they went from: Heinze, Hubrich, and Halfmann -> H and H and H -> the three H's. I kind of expected it to go to "triple H"


H^3 perhaps. There is a precedent. The Lenstra-Lenstra-Lovasz algorithm (LLL) is often called L^3, admittedly because the complexity is O(L^3) in the bit size L. Later a quadratic algorithm was discovered, cheekily called the L^2 algorithm. Later still it was done in quasilinear time. Unfortunately, the authors of those papers did not have the sense to have names starting with L.


I was unaware of an L^2 (or better) algorithm - can you provide a reference? Even just a term to search for - my Search-fu is failing.

Thanks.

Added in edit:

OK, I've found these - hoping I'm on the right track:

    1982    LLL    Lenstra Lenstra Lovász
    1987    BKZ    Block Korkine Zolotarev
    2002    RSR    Random Sampling Reduction
    2002    PDR    Primal Dual Reduction


I definitely like your reasoning for H^3 better than my reasoning for "Triple H," the wrestler :)


An actually readable explanation that can be understood. However, it seems to me that the actual accomplishment, as the author mentioned, is our newly acquired ability of "storing quantum information" and not actually "stopping light". I like the "H's" references...how interesting that they all have H last names - perhaps there is a German stat of last names starting with H.


The article discusses at the outset that the distinction between "storing info" and "stopping light" is subtle, I not nonexistent in this case. When light travels through a material, one can treat it either at a macro level of a light wave slowing down, or at a micro level of light travelling at vacuum speed but being continuously absorbed and re-emitted. The former is a much simpler picture and captures the essence of what is happening. So, if you use a material (as H^3 have done) in which the effective speed becomes vanishingly small, this is "stopping light" in the only sense one can talk about.


We are tidy people:

http://www.datendieter.de/item/Buchstabenhaeufigkeit_von_deu...

(In german, but the picture speaks for itself)


I've tried to read the paper (note IANP, I'm not a physicist) and some additional descriptions, remembering what I've read in Feynman lectures. The principles really sound like reading about the DRAM (with the pulses to write, to read and to "refresh" the content of memory) but with photons in the input and output instead of the electrons. Inside of the crystal, it's the electrons that get to move to higher levels when the photons hit the atom they orbit and then to the lower level emitting the photons out. That's how a plain glass window works. The difference is that the researchers add carefully controlled electromagnetic (light is also electromagnetic emission) signals (additional controlling inputs and outputs) which gives them the possibility to "write" at one point in time and "read" sometime later the same "image." The new achievement is that they reach 40 seconds or more between the write and read, which is a new record. There's even some application of genetic algorithms described in the paper used for preparing the conditions to make such write and read possible:

The loop starts with a random set (‘‘generation’’) of preparation pulses (‘‘genetic individuals’’). Each individual is described by a temporal array of intensity and frequency values (‘‘genes’’). The self-learning loop applies the pulses for EIT and determines the individuals with the highest fitness, i.e., the best light storage efficiency. The next generation is built by imitating concepts of evolutionary biology: The best individuals are copied into the next generation (cloning). Other good individuals are modified by variations of their genes (mutation) or combination with other fit individuals (inheritance). The loop goes through several hundred generations, until the gene sequences (i.e., pulse shapes) converge toward an optimum. Figure 3 shows the progress of the self-learning loop, i.e., the increase of signal pulse energy after light storage vs number of completed generations. As expected, the fitness increases monotonically with the generations.

I'd appreciate comments from anybody who knows and understands more.


Could it also conceivably be used to store energy?


Light is energy, so yes.

A slightly better question might be whether EIT could store energy in amounts large enough to make it useful as a capacitor. At present, this is not likely. This technology is striving towards optical RAM, not caps.


By definition, it can. I think it wouldn't be useful for batteries though, if that's what you had in mind. The energy density is awful (99.9...% of the mass will be that of the crystal, which is just used as a media, not converted to energy).


What I wonder is if this could be used to bridge across decoherence.

Current problem: a quantum system is used to do a calculation, but it decoheres to fast to be useful, or to scale.

Possible solution: do part of a calculation, shove the result in one of these light storage thingies, decohere, reboot and re-establish coherence, feed in result from light storage thingy, continue calculating.


Are you saying the light would somehow allow you to easily reconstruct the lost state of the quantum computer? Or that it's easier to read out and store the quantum state as light than another way?

I don't think either of those things are true, so I don't think this would work to solve that problem.


(De)coherence is a property of your data, rather than a property of your hardware. Hardware quality just affects how quickly your data will decohere.



Is this a possible way to store information in quantum computers?

EDIT: Note to self, reading is tech.


FTA: "What makes the experiment most exciting is that this experiment has proven to be an extremely long term method for storing quantum information, which has traditionally been a major hurdle. Normally a quantum computer (such as they are) has to get all of its work done in a fraction of a second."


Take note, all you commenters who asked "what's the practical value?" when discussing that experiment last week.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: