Quantum Light Harvesting Hints at Entirely New Form of Computing | Hacker News Reader

Quantum Light Harvesting Hints at Entirely New Form of Computing

technologyreview.com

89 pointsdigisth28 comments

This is a really cool article! I've never heard of plants doing something like this. However, the article is written kind of sloppily.

> Quantum states are highly fragile — sneeze and they disappear in a puff of smoke.

No, quantum states never disappear. They just evolve in time. However, it is hard to isolate a system that you are assigning a state vector from its surroundings.

> They say the processes behind light harvesting are a special blend of the quantum and the classical.

This kind of misleadingly suggests that reality has separate quantum and classical components. That's not the case; everything is quantum. It's just that some things that occur cannot be described accurately with classical physics.

> Because energy can exist in a superposition of states, it can travel a variety of routes around the network at the same time. And when it finds the correct destination, the superposition collapses, leaving the energy at the reaction centre.

Light travels. Energy is just a number that's conserved in non-relativistic QM. Also, while I suppose you could describe light as traveling through every path at once (and I know a lot of physicists like to describe it this way too), it sounds much more mysterious than it needs to be to me. I would just say: "as the wavefunction evolves in time, its probability amplitude spreads throughout space".

> Another is the quantum zeno effect, the paradoxical phenomenon in which an unstable state never changes if it is watched continuously.

This makes it sound like its conscious or something. There's no paradox here. "Watching" is just a series of measurements. Measurements collapse a state into an eigenstate. If you perform these frequently enough, the system doesn't have time to decohere, so it just keeps collapsing into the same eigenstate. It would be less like watching the pot and more like pulling it off the stove each time its about to boil.

Anyway, I sound nitpicky, but I remember that before I learned QM I was always bothered by the fact that people tried to make it sound mysterious and vague for some reason when its actually quite objective and concrete.

> This makes it sound like its conscious or something. There's no paradox here

IANAP, but I don't quite get the intuition behind the difference here. I thought the whole point of decoherence is that the quantum superposition collapses into an eigenstate?

I've seen some explanations that make it seem like subjective recognition of a quantum system in superposition is what caused it to decohere. The word "observed" doesn't imply interaction (to the lay-person, at least), so a term like "interacted with" would make things much clearer.

>Energy is just a number that's conserved in non-relativistic QM.

What happens in relativistic QM?

It's not really conserved. Here's a better article about it than anything I could explain:

http://motls.blogspot.com/2010/08/why-and-how-energy-is-not-...

ow... I tried reading, and largely failed. The only thing I got out of that was "Energy density of universe is conserved. universe is expanding. Therefore, total energy is increasing." But not sure if even that part is correct.

My turn to nitpick:

> as the wavefunction evolves in time, its probability amplitude spreads throughout space

  sed 's/probability//'

Probability is a real number embodied in a state of mind[1]. Amplitude are (as far as current physics believes) a complex number out there in the world. I know, squared norm and Born statistics, but still.[2]

http://lesswrong.com/lw/oj/probability_is_in_the_mind/

http://lesswrong.com/lw/r5/the_quantum_physics_sequence/

Oh god who writes those?

It's a convention in physics to call it probability amplitude because, you know, it makes sense, as a direct analogous to energy.

(http://en.wikipedia.org/wiki/Probability_amplitude)

If you want to learn quantum mechanics there are far better places than sketchy, hand waving articles. I'd start here:

http://www.youtube.com/watch?v=hUJfjRoxCbk

> It's a convention in physics to call it probability amplitude because, you know, it makes sense, as a direct analogous to energy.

Wait, what? Could you please flesh out the analogy between probability and energy?

---

From the Wikipedia article:

> Probability amplitudes provide a physical meaning of the wave function, a link first proposed by Max Born, and this is a pillar of the Copenhagen interpretation of quantum mechanics.

Oh, my…

I'd say the real pillar of the Copenhagen interpretation is the word "probability". Which is rather weak, compared to the fact that the collapse postulate of the Copenhagen interpretation is non-linear, non-unitary, non-differentiable (even discontinuous!), non-local, violates CPT-symmetry and Liouville's Theorem, non-deterministic, and faster-than-light. Oh, and it's strictly more complex than a many-world interpretation, since it adds an extra postulate on top of existing equations. http://lesswrong.com/lw/q6/collapse_postulates/

Seriously, how many arrows to the chest does the Copenhagen Interpretation must endure before we declare it dead, and stop saying "probability" before every occurrence of "amplitude"?

---

Fire a photon through a half sieved mirror, with 2 sensors to catch the 2 possible paths, such that no matter what, one of the sensors will go off. What do you think happens?

The equations (you know them more than I do) say the amplitude will flow and decohere into 2 non-interacting blobs. The blob where sensor 1 goes off, and the blob where sensor 2 goes off. (Note from Captain Obvious: the equations are deterministic. They do not say that God plays dice with the universe.) Now say you fire the photon, and observe sensor 1 going off. What do you think happened to the other blob, the one you don't interact with? Do you think it went away just like that? It would be like believing that a colonization spaceship going beyond the observable universe disappears, not only from your view, but for good. http://lesswrong.com/lw/pb/belief_in_the_implied_invisible/

---

Of course this sequence is sketchy and hand waving (it doesn't even try to hide it). Its purpose is not to teach us Quantum Physics. Only certain aspects of it, to serve a philosophical agenda (most notably questions around personal identity and Free Will).

What makes sense about a probability that can take on negative or complex values and collections of which don't sum to 1? Convention is no excuse for bad terminology or notation, especially for beginners to the subject who may not have the math or physics background to appreciate the connection with the continuity equation. You lose nothing by removing the word "probability" and avert possible confusion. "Amplitude" and either "squared modulus" or "amplitude density" work fine.

I think Feynman's QED is mandatory reading for an intro to quantum thinking (and Feynman in general is fantastic). I also think his several comments on wave v. particle indicate another abuse of conventional terminology. Sometimes he'll insist light is not a wave but a particle, but it's actually a screwy kind of particle and you need to think about arrows and paths. ("Collapse" is another word that'd be better off going away...)

My favorite beginner-level approach to QM nowadays, at least for beginners who are familiar with linear algebra, is actually Scott Aaronson's (edit: who is an expert in the field, and I believe has even commented on some of Eliezer's quantum posts...) (quantum) computer scientist approach: http://www.scottaaronson.com/democritus/lec9.html It introduces QM as "What if we tried to make something like probability theory but allowing negative or complex values?" It wouldn't be probability, but it would be like probability. Once you set up the math, the usual "weird" quantum effects of experiments can be shown as a direct consequence of the math.

I'm gonna have to agree with darkmighty here: learn your physics from physicists, not from Eliezer. No matter how well or badly he actually understands quantum physics (which I'm not qualified to judge), we should be learning from actual expert practitioners rather than from a fellow amateur who happened to "ace the class", so to speak.

If physicists say probability amplitude, it's called probability amplitude.

I'm willing to believe experts when I know nothing about a subject. Given the apparent current reception of MWI[1], I concede that things are not looking good for Everett.

On the other hand, arguments trump authority[2]. And Eliezer produced specific, very convincing arguments in favour of Everett. To this day, I don't know of any specific strong objection to his writings on the subject. Therefore I believe him. (For now. Please let me know if you know of any objection he did not address.)

Physicists say "probability amplitude"? I don't care. I'll go on saying "amplitude" alone. The term is less confusing, and physicists will understand me regardless.

[1]: https://en.wikipedia.org/wiki/Many-worlds_interpretation#Pol...

[2]: http://lesswrong.com/lw/lx/argument_screens_off_authority/

I'm taking a QFT class at my university right now, and it's very different than any discussion of quantum mechanics I see online. It's pages upon pages of math and linear algebra. Very little text or discussion.

Physicists don't really care whether you call it "probability", "probability amplitude", or "fish". What you end up with is an equation that you can take to the lab and use to predict experimentally measured results.

I'll add that at some point the math just becomes so bizarre that you can't (or at least I'm unable to) come up with some kind of textual analogy for it.

(This is what QFT is like for me: Manipulate symbols according to mathematical rules until you end up with an equation. Realize the equation can't be solved analytically. Realize exponential computing power would be required to solve it numerically. Make an approximation and modify the equation. Solve numerically and see how well it agrees with experiment.)

Oh God,

Your post seems like nothing but that "genre" of hn post someone pointed out recently. It's the "let me show you I'm smart by pointing the article is technically incorrect in how it states things even though its statements are how both informed laymen and experts tend to express thing" Post

OK, for "Quantum state", substitute "states requiring a specifically probabilistic/quantum-mechanical description" and for "classical states" substitute states amenable to a classical description.

And someone else can explain the other things for you if you need it.

> Your post seems like nothing but that "genre" of hn post someone pointed out recently.

Yeah, sorry about that. I was kind of worried it would come across that way.

I was just thinking though of what kind of clarifications I would have liked to have had with this kind of article before I knew anything about the subject and thought some others might as well.

I think you added something to the conversation, and being the top comment it would seem so do other users. The comment made by 'joe_the_user, however, is shallow and pointless and advances the conversation in no direction but down. I think it's often good to have someone point out the "editorialisation" (for want of a better word) of an article, because after all this material has been designed for (somewhat) casual consumption, not as a how-to. But we're hackers and we like to start sentences with the word "Technically..." ;)

It would be better if for each statement, you said how it should have been phrased rather than just why you thought it was wrong.

My opinion is that you really are nitpicking, in the sense that the author is making the best explanation possible, but you wouldn't accept any verbal explanation.

This is a very interesting comment indeed. I am far too poorly informed to be able to tell if anything you say here is reasonable, but I must admit that your comments all sound very reasonable. But lets please get instead to the substance of the article that you are commenting upon: tell us more about what do you think regarding the implications of the central claims of the article itself?

I wish I knew more about quantum physics.

With most science-related articles, I know enough to be able to tell what's journalistic hype and what contains an element of truth. But with anything about quantum physics or quantum computing, I am completely at a loss as to whether the ideas presented are even real, never mind having a hope of understanding them.

Here is an understandable and serious 40.000 feet bird's eye view of quantum physics: http://lesswrong.com/lw/r5/the_quantum_physics_sequence/

Having reading it, I'm confident this post is crap. The abstract of the scientific article behind it looks much better, though.

I started reading the sequence a while ago. I'm confident I understood everything, up until the dialogue between the aliens who split their brains. There were simply too many streams of consciousness for me to follow. At that point, I decided to quit and come back to it after a few months. I find that sometimes, looking at a problem with a fresh mind is enough to solve it instantly.

the writing smacks of pseudoscience sadly... the idea of 'a mix of quantum and classical' implies a lack of understanding of what quantum mechanics is and specifically the correspondence principle.

when did these two things become separate in anything other than some categorisation of our theoretical understanding of things?

The entire article is coloured with it which makes it hard to glean any valuable information from. The analogy (?) about light bouncing around proteins, superpositions and collapses is especially confusing.

If you look at the paper it does seem genuine and without the hype and bad analogies. (although there is some quite bad grammar, and i am far enough from subjects i am comfortable with that this could be psuedoscience).

What I get from this is 'we have done work on a very specific example of quantum computing in nature - that work provides evidence that such approaches are viable'

This article is very exciting, it almost makes you wonder if the universe is set up to keep adding new possibilities to how much information can in theory be processed by biological systems... e.g. the brain. Anyway if this research is true it's incredible that nature has worked out such a clever hack.

I've not seen many other examples of biology using quantum states - given the article there must be quite a few I assume?

Well, some scientists propose that smell uses quantum physics:

http://www.ted.com/talks/luca_turin_on_the_science_of_scent....

(as I understand, Luca Turin's explanation isn't complete either - again it's a mix of different systems)

There's also the theory that birds detect magnetic fields through quantum effects:

http://www.newscientist.com/article/mg19826544.000-do-birds-...

Actually, I wonder if it's a coincidence that both of these are sensors.

I can't think of any other examples at the moment, but why would't life evolve to exploit quantum effects where it can?

Quantum Bees: http://discovermagazine.com/1997/nov/quantumhoneybees1263

I know it's Discover and probably total crap, but I just love the idea.

This is actually interesting. If their theory is correct plants have essentially loss less and super fast way for photons to traverse a random network^1.

1: They claim it is random network, I am not sure if we can call specific protein truly random. There is also question of size, scale of this effect might be small enough that there will be no applicable technology that could utilize it yet.

in the original article they clear that it isn't a photon, it is exciton that does the traveling:

http://arxiv.org/pdf/1311.4688v1.pdf

"In these systems excitons initiated by the incoming photons should travel really fast throughout a chain of chromophores in order to reach the reaction center where they are converted to chemical energy."

>loss less and super fast way for photons to traverse a random network^1.

My understanding is it isn't exactly loss less. It is a trade-off. The "natural" QM way of simultaneous multi-path travel is the super fast way of finding the reaction center. The issue is that many of these multi-paths may interfere and as result some degree of "entrapment" happens significantly delaying the arrival at the reaction center (Anderson localization - on a lattice draw all possible ways from A to B, including loops, etc... and whenever paths pass through the same edge with different direction double the edge time on each path and given lattice complex enough the propagation would almost stall - very-very rough illustration). To minimize the delay, the nature seems to came up with workaround - [slightly] breaking the phases on different paths and it seems that coherence is still preserved so we still have multi-path simultaneity while the above mentioned Anderson localization is decreased because of phase mismatch. This phase breaking is result of interaction which causes some energy loss/dissipation.

I think this article is auto-generated or pseudoscience.

Reply on news.ycombinator.com