Quantum Computing ‘Magic’ – Computerphile
Articles,  Blog

Quantum Computing ‘Magic’ – Computerphile


>>What’s the problem with quantum then? OK, I think we probably… Probably we should talk about the opportunity I mean what, what’s the idea? Why would we want to build a quantum computer? We discovered… the first observation was that it’s actually quite difficult to simulate a quantum system. and this could be illustrated with this famous double slit experiment where you send electrons, or photons or whatever through two slits, and then you notice that you get this interference pattern. So it gets dark and light, dark and light which we can understand if we think of electrons or photons as waves ’cause the waves interfere with each other but at the same time we know that they are actually particles because we never have half a photon, we never have half an electron so the question is, if each time we send only one electron through these slits Which way does it go? Does it go left, or right? and now you think you’re clever, you put some detectors on the slits You have to ‘trick’ nature, find out where it goes and as soon as you do this they behave like particles, there’s no interference pattern any more. and that’s called a measaurement, so all quantum systems er they have this dual behaviour, they behave like waves as long as you don’t look at them and as soon as you look at them, they behave like particles behaviour where this electron goes through, seems to be going through both holes at the same time it’s really doing something in parallel the one explanation is, er, one understanding is Everett’s er, multiple worlds interpretation that there are parallel universes and in one universe the electron goes through this and the other one through this and you don’t determine which universe you are unless you look at it, and then you are either in this one or in this one OK, now, quantum computing tries to exploit this very weird nature of quantum physics and using the mathematical models we have from physics we can develop new computational paradigm, new… programming language so to say, quantum programming and in this language we can exploit this quantum parallelism to have certain computations run faster it’s not completely easy because everything you do has to be reversible, every computation you do you also have to be able to do backwards so everything needs to be symmetric>>Is it a Qubit? Yes, a Qubit is a bit, and it has an amplitude, it can be somewhere between zero and one, yes? and a Qubit, which is maybe like, in the middle. is like this electron going through the one slit and the other slit and we don’t know which one so while the Qubit is not observed it’s at the same time zero and one, so to say. and all its interaction are based on this “it’s two things at the same time,” like Schrödinger’s cat famously (alive and dead) so as long as you don’t open the box the Qubit is in some super-position so to say it can be just one but it can be also in a super-position and that’s the interesting case but the rule is, you’re not allowed to look at it while you do this so it has to be unobserved because once you observe it you determine which universe you are then this multiple universe idea disappears, this paralellism disappears>>The magic is let out of the box? The magic yeah, once you look at it the magic is out you have to close your eyes for as long as you…. then at the end you can look because you want to know what the outcome is and then you measure but not inbetween There are some, very famous algorithms which work on a quantum computer faster than on any, known, classical there’s no known classical solution a famous problem is the problem of factoring a number so for example 15 can be factored at 3 and 5 and this problem is actually quite important for cryptography. The RSA algorithm is based on this idea that you have the product of two prime numbers but you don’t tell anybody what the prime numbers are so, it’s called a one-way function so it’s actually surprising, it’s easy, computationally to find out whether a number is a prime number but it’s hard to factor a number it’s hard (this by the way is not proven) nobody knows a way to factor a number efficiently on a classical computer but since Shor (The Shor Algorithm) we know on a hypothetical quantum computer, that using this quantum paralellism there is a clever way to do this it uses a certain number theoretic function which has a period (repeats itself) and using this super-position of Qubits you can actually, with a good probability measure, this period, the repetition, and from this period you can then, find a factor that’s quite clever, yeah. So it seems that quantum computing can… can be much much faster than classical computing now here’s the problem we haven’t yet been able to build a quantum computer of reasonable size, where there’s anything interesting happening and why is this a problem? why is it difficult? the challenge is to do computations without looking at it while you have… Your Qubits are represented by some physical object some ions or whatever some particles, yeah and you want these particles to interact with each other to do the quantum computation But, they should not interact with the rest of the universe So you basically have to make them interact with each other, without touching them and, that could be possible. It may be “an engineering problem” as one says But we don’t know whether whether it’s actually possible in principle I think there is a mistake to extrapolate, we know in small systems we have this quantum behaviour, and quantum er, theory gives us very, very good predictions on how these systems behave but we don’t know whether it actually scales up. Whether I mean, there is this problem that you have involuntary measurements, or involuntary observations it’s called decoherence so when this system loses its quantum magic and becomes classical and what we don’t know is whether we can actually avoid decoherence in a large quantum system because in the end, in physics it’s always like this: “You should never assume an outcome of an experiment before you have done it” so far, nobody has done this experiment which shows that large scale quantum computing is actually possible so it may be that there is some hidden law we haven’t yet been able to test ya? which basically says “Once you do too much quantum parallelism, nature… …shakes its head and says no that’s getting too complicated I’m not doing this, I’m going to put some decoherence in otherwise it gets too complicated it could be, I’m not saying it is like this but we basically don’t know So I think research in quantum computing in this area is very interesting, very exciting, because we have actually an open question and either way the answer will be interesting The answer is either, OK we can build a quantum computer, we can do all these cool algorithms, great. Or the answer could be, Actually it’s impossible we have discovered a new law of nature which actually says that nature in the end is classical and this quantum stuff is only on a small scale which would be very very exciting as well Either way it’s exciting, we should find out, but we shouldn’t assume now that we know the answer without having done the experiment>>There’s supposed to be some kind of computer that’s been built Yeah>>Is it that we don’t know if it is doing what we guess? Yes there’s this D-Wave and there are other projects, but as far as I know we don’t know yet whether really get quantum paralellism out of it there are all sorts of claims and counter-claims and I don’t think it has been decided yet>>Is it possible to simulate a quantum computer on a classical computer? Yes indeed, er, I’ve actually developed a library doing exactly this, called Quantum IO Monad Yes you can do this, the problem is, to simulate this quantum paralellism you have to really go through all the possibilities and you have an overhead exponential overhead so yes you can simulate it, but it’s inefficient>>So could you do a very simple sum or something like that on this? Yes we did this algorithm, the Shor algorithm to factor a number and to factor 15 and it turns out, the answer is either 3 or 5… [laugh]>>Is that something we could link to, people could look at or is it err…. yes, yes>>Cool well we’ll put that in the description if anybody is interested in it it’s actually implemented in the programming language Haskell, the functional language I hope you can still download the code!

100 Comments

  • tada1295

    Instead of scaling it up why dont we make many small isolated systems. Each one doing its small thing that is eventually observed at the right moment and then all added up.

  • No Fate

    The challenge of building large-scale quantum computers was a roadblock until just 3 days ago eurekalert org / pub_releases / 2016-12 / uos-cop120216.php

  • No Fate

    Correct the D-Wave Quantum computer is an adiabatic quantum annealing device, not a full-blow quantum computer. (Universal quantum computer's now within reach posted a link below, it will be coming soon. Exciting times ahead…)

  • Jon Smail

    I think once we look at the outcome, we've caused a predetermination of the outcome. Perhaps, we shouldn't look at the outcome. Maybe, we should let it be….

  • Nazgul Linux

    Please do a video on the Michaelson-Morley experiment that was re-performed by the U.S. Navy to detect Aether!! Nature Magazine Vol.322 pg.590 Peer-reviewed and Aether has been verified! Aether is the conduit that connects observation by consciousness to the waves that form into particles when observed! There are no parallel dimensions/universes, but rather, the AETHER!

    Tesla was right. Einstein was largely wrong.

  • Powder

    heisenberg said everything is connected and called these particles not particles but "wirks".. so seperating them from the rest of the universe sounds impossible.. but we really dont know yet 😀
    since i worked on p over np problem for several years i really cant w8 for q computing to happen on large scales 😀 maybe this would change the p np drastically

  • XenoContact

    I don't know…I just don't think quantum physics is correct in the way we think about it. Something about how results change according to our observation seems very off, as if it's not science anymore. I think It all comes back to some simple principle eventually which explains the situation rather than the current probabilistic approach.

  • Jess Stuart

    I still get confused about how you get useful information out of a quantum-computer instead of just a pretty interference pattern. I think it will come down to system-noise being the primary limitation. Physicist will classify the system noise as an "engineering problem".

  • LexorTheUber

    I'm only 2 minutes in and I am pretty high but this makes so much sense if you think of the two parallel universes as the one that exists regardless of life to experience it, and the one we experience based on the sum of our combined senses. Since we're limited by our senses, our version of the universe isn't reality, and that could explain why they act like particles when we look at them and waves when we don't. They really are in a parallel universe when we observe them. Since this universe is in our mind, and our brain is made of particles, mustn't there be some limiting factor there? And if we could make a machine that observes them in a way that our senses don't how could we interpret that information? That sounds to me like trying to understand how a cat thinks or something. I'm gonna finish the video and hope some of this is answered, I'm just here because he looks like THE ONE AND NO LONGER ONLY PIE!

  • Michael McDonald

    This guy might be my favorite person of all time. Chomsky? Russell? Scrubs. Give me some of that Altenkirch.

  • Ritwik Mishra

    We shouldn't look at them while they are computing the answer. Then how do we make sure the answer is correct? how do we make sure it isn't rubbish? does anybody knows how do those qubits "computes" the answer?

  • MM!!

    "Nobody knows a way to factor a number efficiently on a classical computer, but … we know on a hypothetical quantum computer, that using this quantum parallelism, there is a clever way to do this."

    If you understand how our modern encryption systems work you will realize just how fucking scary this is.

  • Tony Anytime

    Only way to do quantum computing is to look only at a copy or shadow of result, never at actual value. Large scale quantum computing probably won't work. Quantum works only until it doesn't.

  • Brian Liotti

    What's this mean for SHA256? are Crypto Currencies prepared!? Requesting a Quantum Encryption video Siraj style, please…

  • Aaron Clark

    This reminds me of the old Far Side comic where some cows are standing around on two legs talking in a field. Then one says "car!" and they are all on all fours eating grass as the car drives by.

  • Marciano Padilla

    The universe is the product of consciousness. Dualism seems to be the fundamental purpose of Superposition and Quantum entanglement.Once you observe one particle in a state the other particle is in the opposite state.

  • John Davis

    So, I keep listening to people talk about quantum computing. They have been repeating the same thing for the last years. I'm waiting for something new. How is it used in a real world example. Still waiting and hope to hear something before I grow old and die… I don't say this as a put down. I just hope some day someone will use some real explanation that I can understand.

  • Marc van Leeuwen

    I like the honesty to say that it is unknown whether there are any fundamental obstructions against creating large scale coherent quantum systems. It might be there is some such thing, maybe of statistic nature similarly to the second law of thermodynamics. If say it were to require, in order to prevent de-coherence, an energy that grows proportional to the exponentially increasing number of superposed states in large systems, then that would probably put any practical application beyond our grasp. But I think there is vastly more effort invested on the assumption that quantum computation will turn out to be feasible than to research the possibility of a fundamental obstruction against it.

  • Jake Harders

    so you are saying all of my print statements littering my code for error checking are not going to be forward compatible in quantum computing.

  • truthseekers666

    Guy looks bonkers… fancy being shut in a bunker with him. It would be ok as nobody looks – both of you would probably be alive… but if someone looked in. Eeeek.

  • charles poirier

    To be honest at first I was like finally .
    Then the stumble of the same script being told over and over again .
    At least there was honest clarity of a simulating factor and the announcement of the quantum still just being a theory and not yet proven .

  • Andrew Gervais

    So, its like, quantum computing lets you converse with the Universe, but the Universe refuses to communicate on anyone's other than it's OWN terms…?

  • Mark LaPolla

    What about the no-teleportation theorem? Quantum state cannot be converted into classical bits. So, what does this say for your classical simulation of a quantum computer if you cannot convert quantum state into classical bits?

  • nightmisterio

    The double slit conclusion is wrong observation does not make a diference the experiment is bad design, if it takes more actions for 1q action tha a normal cpu how can it ever be faster?
    Quantum is a scam. Stock pump?
    They probably will use normal processig with AI functions and call it quantun.

  • MrStevetmq

    May I put to the person in the video and idea that I have had. At the beginning he talks about how particles move in the duble slit experiment. He talks about how they behave like waves but as soon as you put detectors on the slits they behave like particles. My thought is what if the wave like behaviour is in fact a manifestation of the way Space/Time is effected by the movement of the particle. So in fact when we observe the wave like effects what we are seeing is space/time moving like water moves when a boat moves though it.

    Is this and interesting thought?

  • Ali Eser

    Damn I feel really lost right now. I am really interested in this stuff but I haven't even finished studying my basic algorithms and data structures

  • doppler effect

    sir thanks for the info,am new to quantum computing, since quantum computer is like shrodinger's cat, ie, info is can be 0 and 1 at the same time, sir is linear algebra still very much applicable there, because classical computers use the laws of linear algebra (de morgan's law,……….) to minumix their circuits ,which are just the properties of vectorspace in linear algebra.thanks

  • Luka Korosec

    Look it is simple the coolest dude I have seen for years. Electrons are energetic bubbels therefore you see it with double character…. I hope I could help. I would look at the tau lepton ;)! ♥! *huuuuggg 2 the fluffiest person alive!

  • Nate Wynd

    If I could trouble the better-informed on quantum mechanics with a question: I understand that the act of observing a quantum system causes the superposition property to essentially be lost (Is this waveform collapse? Decoherence? Resolution? I'm still learning the terminology)–the wave now behaves like a particle. What does it mean to "observe" a quantum system in these scenarios? What specific actions are taken upon quantum system? I saw from IBM's videos on the topic and it was explained that (in their systems, at least) microwave pulses apply constructive interference that change qubit state as well as entangle two or more qubits. Is something similar happening in order to measure/observe the quantum system?

    Please feel free to correct anything I may have gotten wrong. 🙂

  • Rich M

    Isn't simulating a quantum computer on a conventual one the same as simulating reality…in other words…makes me think a quantum computer is just reality?

  • Timothy Siwula

    can not view code: "No access to directory.

    The Computer Science web server is available to current Computer Science Staff and Research students.
    If you believe you are seeing this message in error, please contact the support team."

  • allmhuran

    This speaker is fantastic. A wonderful voice for explanation, and always extremely precise in what he says. Everything I hear him say is what Descartes would have called a "clear and distinct idea".

  • 00UncommonSense00

    The fact that something behaves one way and just the act of measuring it forces it to act differently is mind melting. You know something is black and when you look at it, it becomes white. Or it is not black or white, but when you measure it the act of measuring forces it one way or the other. That almost is proof of intelligent design IMHO. Think about that, and you will agree. If you don't agree, I will be upset. Be nice and agree! lol

  • Aida Rodriguez

    Hi. this is snow crash from evolution. NSA bled me but still won't hire me to advertise the use of Neuromoprhic Computing and Building 8's obvious neeed to be policed. Wannacry victims should be ale to call 911, Ellis Holdings LLLC. And MArs will need System Shock 4 or an adult film titled Access Talent Keyhole codeword Orgasm Research. Its coming to a HEAD aint it

  • Muuip

    A particle surrounded by a wave. The wave of the particle going trough the double slit creates an interference pattern and the particle follows it randomly. Trying to "see  or measure" it disrupt the wave and the particle acts as a particle.

  • Ella Blun

    why do all quantum computer videos end up just talking about quantum physics? yes, we all know double slits experiment and schrodinger's cat and pavlov's dog… tell us how they are using, or not using it to build a computer. all I gathered so far is that it has to be really cold, which doesn't seem practical. but does it still uses cpu and ram, and power supply? does it still have a mother board? if 1 is true, and 0 is false, what is the superposition, e.i. how do you code this thing? it's like with so many other things. when someone first tries to get into blockchain, it's all high level conceptual discussion, it's really hard to get it. then you finally find someone coding blockchain, and it's actually really simple. just show us the thing and explain what is happening or not happening and how that translates or doesn't translate to things we know from regular computers. if you know.

Leave a Reply

Your email address will not be published. Required fields are marked *