I came up with a model for a new type of computer from playing a game, Counter-Strike (a Half-Life mod) when it was in its original beta phase. The system was very poorly designed, like the accuracy system for the weapons was designed that if you slow down to a walk, your guns were more accurate, but they set the parameters up so that it triggered this extra accuracy just going the slightest speed under a full run. Using +moveup which was meant for swimming in the scripting language, which is the only "language" I used, you could get half way between a run and a walk for movement speed and get the accuracy of a walk and the silence of it, with movement sound being another similar flaw they made in the game. That combined with scripting firing of the gun so it briefly made you do +moveup before actually firing the gun and turning it off immediately after firing the gun effectively gave you a more accurate gun at a running speed. There were many holes in the original CS system, I repeatedly told them about them on their message board, getting repeatedly banned. I remind you: I only used the extremely simplistic scripting language built into the game, so I was exploiting and not cheating, even though in effect it was cheating. CS 1.6 should have been CS 2.0 because they made major changes to the engine due to what I was spreading around. At least one of the hacks that I kept to myself and did not put into my script still exist in the current CS system as far as I know. It was basic to the Quakeworld original engine Half-Life is based on. The script that is part of my work, for CS 1.6, has a fully automated taunt system for giving people a hard time. I built a randomizer and relational database that sometimes spits out a taunt based on the weapon or weapon type you are using just before your gun is actually fired when you fire, only using the one command, alias. Alias just lets you create or reassign a command to an indicated string of commands, and nothing else. We are not digital and nothing in Nature is digital. Digital computers are an exact science with exact results. Nature is based on "good enough is good enough". Oxford quantum physics professor Andrew Steane wrote in his paper about quantum information systems titled "Quantum computing" at arxiv.org/pdf/quant-ph/9708022.pdf : "The new version of the Church-Turing thesis (now called the ‘Church-Turing Principle’) does not refer to Turing machines. This is important because there are fundamental differences between the very nature of the Turing machine and the principles of quantum mechanics. One is described in terms of operations on classical bits, the other in terms of evolution of quantum states. Hence there is the possibility that the universal Turing machine, and hence all classical computers, might not be able to simulate some of the behavior to be found in Nature. Conversely, it may be physically possible (i.e. not ruled out by the laws of Nature) to realize a new type of computation essentially different from that of classical computer science. This is the central aim of quantum computing." From what I understand they are forcing current quantum computers to unnaturally apply a binary state to something that has a infinite evolution of states. Think of the electron and the circle it makes around a nucleus. That 360 degrees circle it makes is infinite in precision, and that movement certainly has an effect on its surroundings. Basically, practical math is the descriptive language of the universe, and not the actual universe because it uses measurements. I propose a "Dynamic Stateless Computer" that operates on "Logic Geometry" based only on connections, or links, or pointers - a much more simple computer than the three basic Boolean logic gates operating on mathematical binary bits that is every computer out there. The shape is the logic and the logic is the shape, sort of like a truth table that is dynamic where the "truths" change as it runs. Quantum mechanics is beyond me, but if this only needs connections, ie a quantum entanglement (short video on entanglement: th-cam.com/video/z1GCnycbMeA/w-d-xo.html ), can we build a computer that operates and does its entire run instantly? Like I said, all I need is connections to perform logic... no need for information... the shape is the logic. You are best off going to Github and seeing online without downloading the paper and models. When someone looked at my calculators, they accused me of: "You're not doing math! You're emulating math!" Look at the simple calculator first, it only does addition and subtraction. Then look at the complex calculator that does multiplication and division. As you well know, if I can do those things, I can do anything mathematically. In the main model I created if-thens, complex do-whiles, a randomizer and a relational database. github.com/johnphantom/Dynamic-Stateless-Computer Through the exercise of the most complex do-while I asked a question related to that, and the answer uses the ancient Chinese/Pascal's Triangle (which millions have looked at over thousands of years) in a new way: mathhelpforum.com/threads/combination-lock.17147/ I basically had to count nothing as something to count, as in you can have different items to count the permutations of but a default state of no item is possible for each, some or all to count in the permutations, and it doesn't seem anyone else in history was able to use the really basic mathematical concept of the Triangle in that way for the solution. It is similar to the 4 hats and 4 pegs question of how many permutations you can have that is commonly associated with Pascal's Triangle, but they did not count the empty pegs as part of the permutations that they can have. The technique of the implementation is a little interesting, with it being able to reach any of the 209 possible permutations of 4 wheels with 4 numbers (don't know if I should count 0, it is special in this case - if you do count 0, it is 5 numbers) in 4 keystrokes or less - it's how it scales that is the curiosity, where if I had 18 slots and 18 items to form a permutation it would have almost 3x10 to the 18th power or 2,968,971,264,021,448,999 possible permutations, each reachable within 18 keystrokes or less. I don't have any idea as to how this would be physically built - none of the aspects of it, except for the dynamic logic that I also do not have any clue if it really is what I ask above. I just can do these things I demonstrate and in my extensive almost 50 years of digital computer experience I have not seen anything exactly like it. Maybe you wonder about my computer experience? I have always been fascinated by computers, starting in 1972 using a prototype Cogar 4 that my dad got his hands on, when I was 3. By the time I was 5, Singer wanted to use me in a commercial to sell the computer, because if a 5 yo could start it, load the OS and then load games, that proved anyone could. My first mentor helped develop Ethernet after working for my father, and allowed me to hold one of the first breadboard ethernet cards developed when I was 10. My first real program (programming since at least 5 if you count the Cogar ASM I had to type to get to the OS and games) was in BASIC when I was 11 that I learned from a manual without anything more than a small example for each command, written with pencil on paper; a rudimentary AI demonstration called "Animals". Second program I made I had another computer (we had moved and left the one at my dads company behind when he sold it) and was a dot bouncing around the screen. Third program, with a 12 year old's understanding of math, I attempted to do 3D. I first professionally programmed in 1982, started building computers and networks for a small computer company in 1986 owned by my second mentor, Peter De Blanc who lead ICANN for a period, was an official beta tester and developer for OS/2 2.0 and developed a device driver for it for the extremely complex Truevision Targa+ 64 video editing board (pic: imgur.com/a/hMe21Qe ) directly flipping bits on it in 1991. The code for the model for the dynamic stateless computer is about 640 lines and took me 6 months to complete, with the code for the Targa+ device driver being over 4200 lines and took me one 20 hour sitting that compiled and ran the first time that I have 3 witnesses for. That's almost 30 years ago. My experience has only gone up from there. This dynamic logic is something I found, that I have never seen anything like even searching for it on the Internet for the past 20 years. I think this is basic to everything and is a new science, as it only operates on one concept - connections.
Finally, there is a non-pop-science quantum computing presentation that clearly explains all the complicated stuff in a certain level of mathematical rigour and clarity. Thanks.
You can look up minutephysics' videos on the Shor's algorith and Bell's theorem. They are more simplified but still quite rigorous. And most importantly they *refrain from technically incorrect analogies.* There's also the video by Veritasium on the Many Worlds Interpretation which - although sounds like the dumb, overly used, science-fictiony trope that requires an ill-behaving spatial dimension - is actually a scary, brilliant idea about the superposition never collapsing.
@Doido do Minescraft FLOPS stands for FLoating OPeration per Second, if that is what you mean. It doesn't have much to do with quantum computing but rather with measuring power|speed of scientific simulation supercomputers.
@Doido do Minescraft I'm still pretty sure the issue here is with question. Otherwise you could just explain it. I'm answering, because you asked it in this thread. There's only me and OP here.
@Doido do Minescraft Yes, exactly. 1 qubit has 1 floop, and 2 qubits have 4 floops. However, due to the error rate and superposition effects, sometimes the 2 qubits have only 2 or 3 floops, with a probability equal to their values squared. It do be like that sometimes.
@Buck The Banjo Player I will not waste my time listing the commercially profitable software products that microsoft has built and keep developing but you literally have no idea what software engineering and computer science is all about. go ahead and memorize some javascript APIs and call it coding.
You've outdone yourself, Microsoft. This is hands-down the best video on quantum computing on the internet. To say that I'm extremely impressed would be an understatement.
This actually goes into my 'gems' collection. A resource so straight to the point, deliver things at such the right amount and makes everything clicks together. Really great talk.
@@nessbrawlaaja Yeah so I have a folder for lots of documentations, planning and whatnot that I regularly make updates. I also have cron jobs running every midnight to push changes to github, and the gems file is at github.com/157239n/Documents/blob/master/gems But there's also another folder where you might find helpful, that's full of technical details at github.com/157239n/Documents/tree/master/technical_references
Keith : This is a very clearest introduction to QC, indeed. It also gives a new interpretation to entanglement. Chinese teleportation experiment shows how advanced they are, they just might produce the first QC.
@Tarek701 : But IBM's QC cannot or does not know how to check for errors and also have difficulty computing or making the QC to calculate, for example, or how to use the QC anything besides factoring big primes. Also checking if you get the right simulation is not known. The problem I think lies in the fact that our senses, our brain, all our cells employ natural quantum computers (we do today what nature did yesterday) enabling us to survive and evolve (in room temperature), even protein production occurs at 99.99 % efficiently and at lightening speed. We need to learn from nature, just like Ibn Haytham dissected an eye to discover optics that enabled Galileo to invent the telescope. Practical QC seems a long way away.
@@saskiavanhoutert3190 We damage 50-70 billion cells daily that repair/regenerate at 99.99% efficiency and at lightning speed, due to our body functioning as a QC.
I'm studying for a Quantum mechanics final, watched this to procrastinate and honestly ended up understanding the Bloch Sphere a lot better because of it
Superposition isn't really all that weird, and actually has a really nice physical analog. How can a qubit be true and false at the same time? The same way that you can hear a song with multiple instruments playing simultaneously. Each instrument makes a sound wave, and they all get added up as they hit your ears. A "superposition" is really just a linear combination of different waves of the possible states. Instead of a sound wave, they're "probability waves". Each state has its own wave function, which you can imagine as a simple sine wave. Different states => different frequency. Now imagine adding multiple sine waves of different frequencies. That's a superposition. The catch is that only certain states (frequencies) can be used. The frequencies are quantized, and that's where the whole "quantum" thing comes from. And this whole superposition business isn't even unique to QM. It arises from wave mechanics in general and DiffEq
@@cwifrbm926 Wow that example of superposition is probably the best one I've heard! I'll probably use that in my presentation. Did you come up with that yourself or did you get it from somewhere?
@@pokepe12 In the physics of waves the wave (sound, mechanical, light, quantum, ...) constructed as a sum of some elementar waves (like pure sine waves with different frequency) is normally called a superposition. So the connection is not really that new and surprising. But as you have said, it is probably the best way to visualise it.
You are allowed to Pause the Video I believe the body language is more important then the very over simplistic sides showing the flow charts of represented matrixes. He is communicating a topic of how these gates work not how to make them and then later the end how to code for these gates again not how to make them.
I left TH-cam to play for like an hour while making and eating lunch, and ended up here, and watched for about 20 minutes and I don’t understand much because I’m not at all studied in computers, but I can already tell that this guy is an amazing teacher.
Superposition isn't really all that weird, and actually has a really nice physical analog. How can a qubit be true and false at the same time? The same way that you can hear a song with multiple instruments playing simultaneously. Each instrument makes a sound wave, and they all get added up as they hit your ears. A "superposition" is really just a linear combination of different waves of the possible states. Instead of a sound wave, they're "probability waves". Each state has its own wave function, which you can imagine as a simple sine wave. Different states => different frequency. Now imagine adding multiple sine waves of different frequencies. That's a superposition. The catch is that only certain states (frequencies) can be used. The frequencies are quantized, and that's where the whole "quantum" thing comes from. And this whole superposition business isn't even unique to QM. It arises from wave mechanics in general and DiffEq
@@zzzzzzmc Superposition is probably the least weird thing about quantum physics. It makes sense, intuitively. You start trying to measure particles and you get REALLY weird results. But, even that isn't is complicated as people make it out to be. The problem is people try to imagine particles as little balls of matter, and they are not that at all.
One of the best explanation undoubtedly. One additional info: Now a days, the IBM composer (what he has shown, while creating Q-circuits graphically) is much more robust and added more bells and whistles., which is kind of expected.
A classroom full of smart computer science students all being silent when asked if they have any questions just means they were all so confused that they didn't even know how to go about asking a question.
My impression was that they had just taken him at his word when he quoted the "shut-up-and-calculate" guy. It's kind of like the suspension-of-disbelief that people do every day when they watch TV shows about dragons or superheroes.
It's just that quantum mechanics is just such a severe level of abstraction from reality that they don't really understand it. After all, they are just e flat manilla computer scientists. No QM classes, very few with linear algebra or vector calculus background. Unless, you do research in quantum mechanics, it is a shut and calculate discipline. Its very non-intuitive. Schrödinger, Dirac, Heidelberg, et al. Were really bright guys.
Think it's more that this guy doesn't know as much as he'd like you to think. It's a common tactic to keep banging on about how everything is super simple when it clearly isn't, it makes the audience reluctant to ask questions because it'll show they don't understand. The couple of questions this guy did get asked, he was like a rabbit in the headlights and was unable to explain because it deviated from the script that was prepared for him.
Half way through the video, God, I saw so many videos about Quantun computers and none of them explained as clearly as this guy. We must find who he is and ask him to write a book or put a course series, It will benefit humanity very much.
Happy you enjoyed! I've written a couple follow-up blog posts: ahelwer.ca/post/2018-12-07-chsh/ ahelwer.ca/post/2019-12-21-quantum-chemistry/ There are a ton of good quantum computing resources out there, but a whoooooole lot of bad ones. We just hope the good ones will eventually bubble up.
As to the input/output confusion at 36:24 The words i/o refer to the names of two variables: i is *used* to store the input data, o is *used* to retrieve the result. The apostrophe ' indicates the state before or after applying the gate BB. This is different from the one qbit gate, where there was no variable name for the one qbit, and i/o described the state like the apostrophe ' in the second example.
Another way to put it: "Input" and "Output" are merely the *names* of the *lines* -- "Input" is named after what it is used for *before* the computation, and "Output" is named after what it is used for *after* the computation.
Thank you so much. This is the best intro to QC I've yet seen. At 36:00 you get into the hack to handle non-reversible functions. I'd like to offer this intuition: The motivation for adding the extra qbit is to have a place to store the information that would otherwise be lost by the non-reversible function. We don't actually care about this information, but we don't want it to be lost, so we stow it away in Input'. It's basically the same reason be rent storage units. We don't actually want all that crap, and deep down we know we will never use it, but we are unwilling to throw it away quite yet.
Joel Forsyth , this is a very curious feature of qm. My understanding is that quantum coherence is destroyed by any interaction that involves increasing entropy,. Increasing entropy means irreversible. One way to look at this is to note that the arrow of time is defined by entropy change. Physical interactions can freely ignore the arrow of time as long as entropy doesn't change. This is why QM violates our "common sense" notions of time and space, and yet just barely leaves our notion of macroscopic casualty intact. Mind blowing exercise for the reader: given an ideal quantum computer with the magic ability to perform arbitrary irreversible computations, construct a temporal telegraph capable of sending messages into the past.
@Hrithik Diwakar okay... here's another angle on it: a pretty good test for any hypothesis with respect to either relativity or quantum mechanics is causality. If you can do a thought experiment that violates causality, then you have an error in your model. A causality violation is anything that transmits information back in time (or faster than c). Physics doesn't have any causality violations. However, quantum mechanics includes phenomena that are logically indistinguishable from time travel. Why doesn't that violate causality? Because quantum entanglement is time and entropy symmetric. If one side of a quantum interaction has more information than the other, that would break that symmetry, allowing a message to be sent faster than light or equivalently backwards in time, thereby violating causality. I hope that helps :)
I agree that this is definitely the most accessible intro to QC I’ve seen, but perhaps you’ll be able to clarify my confusion at 36:00; if you were to chain the reverse operation BB to the output of the forward BB, would input’ be input for the second BB?
As someone with basically 0 knowledge about the topic this was awesome! Really good explanation, i actually feel like i learned something which is not always the case in some of these talks
AHHA, finally a good explanation of quantum computing. After having found lots of good descriptions of quantum mechanics (bra-ket notation, hermitian matrices, etc, etc, etc) that left me wondering about computing, this video is great. Many thanks to the presenter.
This presentation is fantastic. This is the first one after which I'm legitimately interested in learning more about quantum computing instead of thinking "Man, that's weird"
Thank you so much, you are awesome! I've been studying tensor calculus for at least 2 years and never could grasp how the tensor product is actually calculated. It literally took you less than 2 minutes to explain it in a way I could clearly understand and visualize!
Well, it's been some years ago, but what he calls a tensor product is not really the definition of it. It's just one example, one representation. The definition is more like: glue two vectors together, and get a new vector. This object v×w (the tensor product) can now be multiplied by scalars which is the same as multiplying with this scalar in one of the components (v, w) and can be added with other tensors of the same 'rank'. If one choose a Basis for this vector space , where v and w where 2-dim, you can define (1,0)×(1,0) = (1,0,0,0) , ... but also as matrix (1,0 \ 0,0) when using dyads. Since vectors, matrices and tensors are elements from vector spaces, they can be represented as real or complex vector spaces. After identifying (which requires the basis) this is one tensor product, but the general definition does not need a Basis at all and can be used at different vector space (e.g. complexification). The wedge product (the hat like logical and, subspaces of the Tensor product) has one more special property: v×w = -w×v, which is useful in multidimensional analysis when calculating volumes in 3D, 4D ...
This is one of the best introductions to quantum computing that I've seen. Other teachers should use this presentation as a template to expand on. Thanks to whoever that guy is, he's a good (nifty) instructor!
It was hard to follow for me but my interest in this has grown due to this. I did run through this multiple times before it started making sense. I am still not 100% there but getting there. QC is really cool. I love it. Nature is so weird and wonderful. This was a great video. Thanks!
It'll be easier if you learn classical computing first. He does a lot of comparing to classical computing to help understand quantum computing, so if you don't follow the comparisons, you're probably going to struggle to understand the subject matter.
Great video! I was totally impressed that programming a QC is so easy. The hard part is finding suitable problems and developing algorithms. We have to reinvent computing from scratch!
That's not really correct... Quantum decoherence is the biggest issue, which has to do with the device. Currently they use an electric flux threading two concentric superconducting rings. This has a 90 microsecond decoherence time! Certain topological phases of 2D systems with certain guage symmetries support nonlocal objects called anyons. These have recently been shown as an avenue for robust quantum computation using their braiding statistics. In short the underlying device is the main issue.
@@marcusrosales3344 sure. I wasn't talking about the actual hardware. I am computer scientist, they know nothing about the hardware since abstraction layers have been invented;)
@@jpt3640 There are problems which are known to be solved better on QC though... Have you read ANYTHING on the subject? If not, you don't know, so you can't state anything.
What a great video, thank you so much for that! It is also very good to see that someone so young may be so knowleadgeable about a high technology subject.
At 39:00: "Why q-function needs output q-bit as input?" type Qbit = {value: boolean}; type QuantumFunction = (input: Qbit, output: Qbit)=> void; The function doesn't return anything, so it takes an "output" argument/"pointer" which it can modify (using the input). Such a modification might also be affected of the "output" initial value, which is why it is supposed to be initialized to 0. eg. "NotOperation = (input, output)=> input? output: !output", is only a NOT-function if the output is initialized to 0. Helpful metaphor?
finally. I've watched and read so many things that just talk about the fact that superposition is a thing without actually going in to what it means for computation
Shor's algorithm can (in theory) find prime factors of very large integers. If we ever have a reliable quantum computer with MANY qubits we will be able to break military grade encryption in a fairly short amount of time. That is currently impossible using classic computers.
Great slides and great talk! Andrew needs to create a follow up presentation that goes deeper. Maybe explain a few more quantum operators, gates, algorithms. etc. Some discussion about various physical implementations of qubits (including topological QC) would also be helpful.
I thought he was presenting to an auditorium full of people and when the camera showed the audience there's like 10 people there in a small room! Pretty cool that nearly 1.5 million ppl have seen the vid and nearly 2k comments so there is definitely interest. Very well presented by this young guy.
Thank you for your mathematical approach of e,g, the difficult phenomenon of "coordination" of Qbits. Our language limitations only reflect the fact that we do not really understand quantum physics. It works, but we don't know why it works. Mathematics only define a model.
Loved seeing a quantum computing explanation that wasn't afraid to get into the mathmatics Every other explanation I had seen so far barely talked about how this could mathmatically work, all they said was "well, is both 1 and 0 at the same time" But actually understanding the complex tensor representation really helped However, I kind of got curious as to how the qbit could work in complex states I believe I should start my studies in computer science next year, since it's my last year at high school, so I believe I am going to be one of the pioneers of actually puting this new technology to use in my carrer, really excited
yes, matrix manipulation definitely helps understanding. The Deutsch Oracle case gives an idea on quantum advantage, but there are other factors including which algorithms can really benefit and how long will it take to deploy them in production. Yet I see VW , Airbus and CERN are serious about QC.
It's times like these I wish I knew more. It always takes me 2 time longer than the actual video length to watch informative videos like these. I don't know what he's drinking but it makes me want some fruit punch.
Don't worry man, I've been studying Physics since high school for about 6 years now and specifically studying quantum computing for about 4 weeks and I feel that that's barely good enough to watch the video and understand it all without stopping. Furthermore none of those guys in the lecture seemed to get it.
Imagine you have to make a decision. You can either do x, or you can do y. Then, 5 minutes later, you must also do either x or y. And so on, for an arbitrary amount of time. You won't know the factors that affect each decision until you reach each decision. But if you write each decision on to a paper, you can predict many possible paths, x-y-x-x-y-x-y-y-y etc. The superposition is the sum of all those paths. You cannot possibly know which path you will take, same for the the next person and the next person after that. But you can predict the probability you'll each end up in certain places.
Finally an explanation without the metaphors. I was so confused by all the garbage videos that ended up conveying no information to me, but now everything makes so much sense.
Awesome video! And watching all these older people, who're probably smarter and more educated than me, asking the same questions I do goes a long way toward not making me feel stupid.
Hello there's this trader you should know she's an expert in trading she as made me wealthy, Ever since i started following her strategy & techniques things as be going well for me, Contact via Gmail @hauserjolietradingoptions and thank me later..
Hello there's this trader you should know she's an expert in trading she as made me wealthy, Ever since i started following her strategy & techniques things as be going well for me, Contact via Gmail @hauserjolietradingoptions and thank me later..
I agree with you. It' the guy. he is honest and likes to share what he knows which is the opposite of what Microsoft does. Hopefully, he doesn't work there for a while otherwise he will learn their culture and he will start to explain stuff as what Microsoft did in the Microsoft helicopter joke!
slides 1:21 why learn 4:33 learning objectives 7:37 he got me here lol, started saying 2bit ops lol 10:48 tensor products of values (missed a slide oops) ... 24:30 Hadamard Gate 33:11 Deutsch Oracle 52:00 Tricky bit on Tensor Products
I'm still struggling to understamd the thing with the Deutch oracle. Yes we got one input of 2 qbits, but the samecould be done with a classical. If we simply send a 1 and a 0 at the same time all will be clear. However, if we only send one at a time, why did the algorhythm shown in the video has 2?
This is a very clear introduction to the topic. I think a useful way to describe the hack for non-reversible functions is as follows: one can't build a "block" using quantum mechanical circuitry that implements a non-reversible function f. But, one can implement a block that takes two input qubits, O and I, and produces two output qubits O' and I' with the following property: if supplied a zero for O and some input x on I, it will output f(x) on O'. One can use this building block in a larger circuit, ignore O and I', and get the effect of a block that computes f.
As far as I understand the phenomenon of entanglement collapse is only about the actual knowledge of the states. I guess a better picture to simplify the would be a + b = 1, with a and b being natural numbers. The time you set (measure) the value of one of them you automatically *know* the value of the other and that's why it's not possible to send information through them without entangling the variables again. Thank you for the excellent presentation and please correct me if I'm equivocated. 1:01:42 Ok. So he basically explained here that this is in fact not what happens at all.
that example (as almost all) fails to explain the essence though which is that the numbers a and b were not predetermined until you looked at them. It's not like some third person wrote 'a' on one paper and 'b' on the other and gave it to 2 different people to look at when far apart. It's more like I'll give you 'a' then you flip a coin and decide what to write on it and the other person will get what's needed to satisfy the equation. And the confusion comes from the part where you try to figure out how the other person knew what you flipped.Or the other person's 'b' paper rather.
That's called the local hidden variable hypothesis. It was disproven in 1972 by showing that some entangled quantum systems can violate bell inequalities which must be satisfied for local hidden variable models. So it must be that entangled quantum systems can have superluminal causal influence (nonlocal) or that they are fundamentally undetermined until measurement (no hidden variables).
Andrew, in the video you mention that it'll be a few years until we make or break it with QC. Do you have any new insights on the exponential term and the scaling of QC now?
people working at microsoft are probably the only people using edge :) Very well explaining presentation. Thats an introduction that i wanted from someone who really knows whats going on with quantum computing. Thx for sharing
1:05:26 - I'm still confused about that myself. Maybe you can't transfer data through the entangled particles but if you can collapse it by measuring it at will, if there is a way to tell when that other particle collapses then the collapse itself can be the communication. I'm guessing what he's saying is that there is no way to tell when the state collapses so you can't use it to signal?
40:00 The confusion here is that there are two qbits, one named 'input' and one named 'output', but these two are both passed into the quantum black box to be operated on. The 'input' qbit is left unscathed after the black box, but it is the 'output' qbit which is rewritten with the value of [the possibly nonreversible function f : applied to whatever the value of our 'input' bit is]
Great Video! I have looked at various articles and videos online which described pieces (generally confusing) of the quantum computing puzzle. This brought it all together and I now feel I have a solid basis to go further.
If these algorithms can be modeled with the unit-circle state machine, why can't a classical computer run that algorithm? Or can it, but just not when using complex numbers?
Joe Cleveland classical computers can run any quantum algorithm, with exponential slow-down in certain scenarios. It is actually true that any quantum circuit consisting only of Hadamard and CNOT gates can be simulated in polynomial time with a classical computer. If you add another gate called a Toffoli gate (basically CNOT with two control bits) then you can generate circuits which take exponential resources to simulate on a classical computer.
You can see that for example the entanglement, it cannot be factored into two qbits but have to be stored as a 4-vector. Thus cannot be represented by the unit circle state machine
Quantum computing also is a finite memory turing machine.... Just like classical ones.... Hence both have similar computational power but classical computers get exponentially slow for such superposition driven algorithms
Thank you so much!! I've watched so many videos on this topic, but after I've watched this one, everything became truly clearer. Great presentation! (till now, I've also had 5 courses at the University, and didn't get anything unfortunately)
Exactly! 😆 The empty can and the belittling of the audience *Okay, got it? Look at this shiny coin! Any more questions you dummies?* lol.. Yeah they all want to know why flipping a coin and writing both possibilities on two pieces of paper, then separating the two answers over a vast distance is so amazing! "Your coin was heads? No way! Mine was tails, and it collapsed faster than the speed of light!" 🤣
this speed is how i got normal computer science tbh,, the teacher would write while speaking about the ands and ors the nands and xnors etc with one hand on the white board ... and in the other hand he would have the white board eraser to make space for the next block of text lines and formula's
![News from the front in the post-quantum crypto wars with Dr. Craig Costello [Podcast]](http://i.ytimg.com/vi/HYgxhJ9HgPw/mqdefault.jpg)
![News from the front in the post-quantum crypto wars with Dr. Craig Costello [Podcast]](/img/tr.png)
Finally someone that explains how it works, and not just the "is in both states at the same time" kinda thing
Superposition
It a tug of war
Very instructional.
I came up with a model for a new type of computer from playing a game, Counter-Strike (a Half-Life mod) when it was in its original beta phase. The system was very poorly designed, like the accuracy system for the weapons was designed that if you slow down to a walk, your guns were more accurate, but they set the parameters up so that it triggered this extra accuracy just going the slightest speed under a full run. Using +moveup which was meant for swimming in the scripting language, which is the only "language" I used, you could get half way between a run and a walk for movement speed and get the accuracy of a walk and the silence of it, with movement sound being another similar flaw they made in the game. That combined with scripting firing of the gun so it briefly made you do +moveup before actually firing the gun and turning it off immediately after firing the gun effectively gave you a more accurate gun at a running speed. There were many holes in the original CS system, I repeatedly told them about them on their message board, getting repeatedly banned. I remind you: I only used the extremely simplistic scripting language built into the game, so I was exploiting and not cheating, even though in effect it was cheating. CS 1.6 should have been CS 2.0 because they made major changes to the engine due to what I was spreading around. At least one of the hacks that I kept to myself and did not put into my script still exist in the current CS system as far as I know. It was basic to the Quakeworld original engine Half-Life is based on. The script that is part of my work, for CS 1.6, has a fully automated taunt system for giving people a hard time. I built a randomizer and relational database that sometimes spits out a taunt based on the weapon or weapon type you are using just before your gun is actually fired when you fire, only using the one command, alias. Alias just lets you create or reassign a command to an indicated string of commands, and nothing else.
We are not digital and nothing in Nature is digital. Digital computers are an exact science with exact results. Nature is based on "good enough is good enough". Oxford quantum physics professor Andrew Steane wrote in his paper about quantum information systems titled "Quantum computing" at arxiv.org/pdf/quant-ph/9708022.pdf :
"The new version of the Church-Turing thesis (now called the ‘Church-Turing Principle’) does not refer to Turing machines. This is important because there are fundamental differences between the very nature of the Turing machine and the principles of quantum mechanics. One is described in terms of operations on classical bits, the other in terms of evolution of quantum states. Hence there is the possibility that the universal Turing machine, and hence all classical computers, might not be able to simulate some of the behavior to be found in Nature. Conversely, it may be physically possible (i.e. not ruled out by the laws of Nature) to realize a new type of computation essentially different from that of classical computer science. This is the central aim of quantum computing."
From what I understand they are forcing current quantum computers to unnaturally apply a binary state to something that has a infinite evolution of states. Think of the electron and the circle it makes around a nucleus. That 360 degrees circle it makes is infinite in precision, and that movement certainly has an effect on its surroundings. Basically, practical math is the descriptive language of the universe, and not the actual universe because it uses measurements.
I propose a "Dynamic Stateless Computer" that operates on "Logic Geometry" based only on connections, or links, or pointers - a much more simple computer than the three basic Boolean logic gates operating on mathematical binary bits that is every computer out there. The shape is the logic and the logic is the shape, sort of like a truth table that is dynamic where the "truths" change as it runs.
Quantum mechanics is beyond me, but if this only needs connections, ie a quantum entanglement (short video on entanglement: th-cam.com/video/z1GCnycbMeA/w-d-xo.html ), can we build a computer that operates and does its entire run instantly? Like I said, all I need is connections to perform logic... no need for information... the shape is the logic.
You are best off going to Github and seeing online without downloading the paper and models. When someone looked at my calculators, they accused me of: "You're not doing math! You're emulating math!" Look at the simple calculator first, it only does addition and subtraction. Then look at the complex calculator that does multiplication and division. As you well know, if I can do those things, I can do anything mathematically. In the main model I created if-thens, complex do-whiles, a randomizer and a relational database.
github.com/johnphantom/Dynamic-Stateless-Computer
Through the exercise of the most complex do-while I asked a question related to that, and the answer uses the ancient Chinese/Pascal's Triangle (which millions have looked at over thousands of years) in a new way: mathhelpforum.com/threads/combination-lock.17147/ I basically had to count nothing as something to count, as in you can have different items to count the permutations of but a default state of no item is possible for each, some or all to count in the permutations, and it doesn't seem anyone else in history was able to use the really basic mathematical concept of the Triangle in that way for the solution. It is similar to the 4 hats and 4 pegs question of how many permutations you can have that is commonly associated with Pascal's Triangle, but they did not count the empty pegs as part of the permutations that they can have. The technique of the implementation is a little interesting, with it being able to reach any of the 209 possible permutations of 4 wheels with 4 numbers (don't know if I should count 0, it is special in this case - if you do count 0, it is 5 numbers) in 4 keystrokes or less - it's how it scales that is the curiosity, where if I had 18 slots and 18 items to form a permutation it would have almost 3x10 to the 18th power or 2,968,971,264,021,448,999 possible permutations, each reachable within 18 keystrokes or less.
I don't have any idea as to how this would be physically built - none of the aspects of it, except for the dynamic logic that I also do not have any clue if it really is what I ask above. I just can do these things I demonstrate and in my extensive almost 50 years of digital computer experience I have not seen anything exactly like it.
Maybe you wonder about my computer experience? I have always been fascinated by computers, starting in 1972 using a prototype Cogar 4 that my dad got his hands on, when I was 3. By the time I was 5, Singer wanted to use me in a commercial to sell the computer, because if a 5 yo could start it, load the OS and then load games, that proved anyone could. My first mentor helped develop Ethernet after working for my father, and allowed me to hold one of the first breadboard ethernet cards developed when I was 10. My first real program (programming since at least 5 if you count the Cogar ASM I had to type to get to the OS and games) was in BASIC when I was 11 that I learned from a manual without anything more than a small example for each command, written with pencil on paper; a rudimentary AI demonstration called "Animals". Second program I made I had another computer (we had moved and left the one at my dads company behind when he sold it) and was a dot bouncing around the screen. Third program, with a 12 year old's understanding of math, I attempted to do 3D. I first professionally programmed in 1982, started building computers and networks for a small computer company in 1986 owned by my second mentor, Peter De Blanc who lead ICANN for a period, was an official beta tester and developer for OS/2 2.0 and developed a device driver for it for the extremely complex Truevision Targa+ 64 video editing board (pic: imgur.com/a/hMe21Qe ) directly flipping bits on it in 1991. The code for the model for the dynamic stateless computer is about 640 lines and took me 6 months to complete, with the code for the Targa+ device driver being over 4200 lines and took me one 20 hour sitting that compiled and ran the first time that I have 3 witnesses for. That's almost 30 years ago. My experience has only gone up from there. This dynamic logic is something I found, that I have never seen anything like even searching for it on the Internet for the past 20 years. I think this is basic to everything and is a new science, as it only operates on one concept - connections.
@@johnphantom What you "invented" is just a really big truth table
Finally, there is a non-pop-science quantum computing presentation that clearly explains all the complicated stuff in a certain level of mathematical rigour and clarity. Thanks.
You can look up minutephysics' videos on the Shor's algorith and Bell's theorem. They are more simplified but still quite rigorous. And most importantly they *refrain from technically incorrect analogies.*
There's also the video by Veritasium on the Many Worlds Interpretation which - although sounds like the dumb, overly used, science-fictiony trope that requires an ill-behaving spatial dimension - is actually a scary, brilliant idea about the superposition never collapsing.
@Doido do Minescraft FLOPS stands for FLoating OPeration per Second, if that is what you mean. It doesn't have much to do with quantum computing but rather with measuring power|speed of scientific simulation supercomputers.
@Doido do Minescraft I'm still pretty sure the issue here is with question. Otherwise you could just explain it.
I'm answering, because you asked it in this thread. There's only me and OP here.
@Doido do Minescraft Yes, exactly. 1 qubit has 1 floop, and 2 qubits have 4 floops. However, due to the error rate and superposition effects, sometimes the 2 qubits have only 2 or 3 floops, with a probability equal to their values squared. It do be like that sometimes.
Uuhzvhxhh fff tax a sffYRytTt as
May I just say that this presenter is excellent. Engaging, enthusiastic and very knowledgeable. I would love to see more lectures he's given.
“This is aimed at computer scientists”
Me, definitely not a computer scientist: Ah, finally, a video for me
Im not even a programmer but i understood everything
Our battle will be legendary
@Buck The Banjo Player He works at Microsoft as a Software Engineer. What are you talking about?
@Buck The Banjo Player wtf is a real software job?
@Buck The Banjo Player I will not waste my time listing the commercially profitable software products that microsoft has built and keep developing but you literally have no idea what software engineering and computer science is all about. go ahead and memorize some javascript APIs and call it coding.
What a breath of fresh air seeing an older gentleman asking a clarifying question. Kudos!
You've outdone yourself, Microsoft. This is hands-down the best video on quantum computing on the internet. To say that I'm extremely impressed would be an understatement.
This actually goes into my 'gems' collection. A resource so straight to the point, deliver things at such the right amount and makes everything clicks together. Really great talk.
it's really just sponsored and hosted by microsoft
This makes me curious about your other "gems", mind sharing? 🙂
@@quangho8120 commenting to hopefully be notified with your other gems
@@nessbrawlaaja Yeah so I have a folder for lots of documentations, planning and whatnot that I regularly make updates. I also have cron jobs running every midnight to push changes to github, and the gems file is at github.com/157239n/Documents/blob/master/gems
But there's also another folder where you might find helpful, that's full of technical details at github.com/157239n/Documents/tree/master/technical_references
This is the clearest, simplest, most practical introduction to quantum computing I have seen. Thanks so much for this presentation!!
Keith : This is a very clearest introduction to QC, indeed. It also gives a new interpretation to entanglement. Chinese teleportation experiment shows how advanced they are, they just might produce the first QC.
@Tarek701 : But IBM's QC cannot or does not know how to check for errors and also have difficulty computing or making the QC to calculate, for example, or how to use the QC anything besides factoring big primes. Also checking if you get the right simulation is not known.
The problem I think lies in the fact that our senses, our brain, all our cells employ natural quantum computers (we do today what nature did yesterday) enabling us to survive and evolve (in room temperature), even protein production occurs at 99.99 % efficiently and at lightening speed. We need to learn from nature, just like Ibn Haytham dissected an eye to discover optics that enabled Galileo to invent the telescope.
Practical QC seems a long way away.
I agree, like to see quatum computing in function, thanks.
@@saskiavanhoutert3190 We damage 50-70 billion cells daily that repair/regenerate at 99.99% efficiency and at lightning speed, due to our body functioning as a QC.
Naimul Haq fast computation on certain computing problems is not necessarily quantum computing
"So you can send entangled Qbits by laser?" "Yes" "Okay that's even cooler" love that guy 😂
@@im_mid_af what minute?
@@abisarwan20 1:23:40
I'm studying for a Quantum mechanics final, watched this to procrastinate and honestly ended up understanding the Bloch Sphere a lot better because of it
Same
Superposition isn't really all that weird, and actually has a really nice physical analog.
How can a qubit be true and false at the same time? The same way that you can hear a song with multiple instruments playing simultaneously. Each instrument makes a sound wave, and they all get added up as they hit your ears. A "superposition" is really just a linear combination of different waves of the possible states. Instead of a sound wave, they're "probability waves".
Each state has its own wave function, which you can imagine as a simple sine wave. Different states => different frequency. Now imagine adding multiple sine waves of different frequencies. That's a superposition.
The catch is that only certain states (frequencies) can be used. The frequencies are quantized, and that's where the whole "quantum" thing comes from. And this whole superposition business isn't even unique to QM. It arises from wave mechanics in general and DiffEq
@@cwifrbm926 Wow that example of superposition is probably the best one I've heard! I'll probably use that in my presentation. Did you come up with that yourself or did you get it from somewhere?
@@pokepe12 In the physics of waves the wave (sound, mechanical, light, quantum, ...) constructed as a sum of some elementar waves (like pure sine waves with different frequency) is normally called a superposition. So the connection is not really that new and surprising. But as you have said, it is probably the best way to visualise it.
@@cwifrbm926 This is honestly the most perfect description of superposition I have ever read!
I'm loving his jabs at pop science while actually explaining things as simply as possible.
We could do with more focus on the slides and less on the guy.
Here's a link to the slides if you'd like to follow along! Includes bonus appendices: ahelwer.ca/files/qc-for-cs.pdf
You are allowed to Pause the Video I believe the body language is more important then the very over simplistic sides showing the flow charts of represented matrixes. He is communicating a topic of how these gates work not how to make them and then later the end how to code for these gates again not how to make them.
He's so cute though
@Troy McQuinn You know that you can place two windows next to each other. The slides are one click away.
... and here is an other link www.microsoft.com/en-us/research/uploads/prod/2018/05/40655.compressed.pdf
I left TH-cam to play for like an hour while making and eating lunch, and ended up here, and watched for about 20 minutes and I don’t understand much because I’m not at all studied in computers, but I can already tell that this guy is an amazing teacher.
Props to Andrew for being so active in the comments and helping out people who have questions
This is the best video about quantum computers I have found so far.
This guy looks like he enjoys this, and actually explains things so I can understand.
Superposition isn't really all that weird, and actually has a really nice physical analog.
How can a qubit be true and false at the same time? The same way that you can hear a song with multiple instruments playing simultaneously. Each instrument makes a sound wave, and they all get added up as they hit your ears. A "superposition" is really just a linear combination of different waves of the possible states. Instead of a sound wave, they're "probability waves".
Each state has its own wave function, which you can imagine as a simple sine wave. Different states => different frequency. Now imagine adding multiple sine waves of different frequencies. That's a superposition.
The catch is that only certain states (frequencies) can be used. The frequencies are quantized, and that's where the whole "quantum" thing comes from. And this whole superposition business isn't even unique to QM. It arises from wave mechanics in general and DiffEq
Genuinely curious if you don't think superposition is really that weird, what things DO you find weird?
Dude that was the most amazing and simple explanation I've gotten so far, thanks a ton John you made my day.
I like this explanation, thanks! If after this, you still find superposition 'weird', then Quantum Computing, or Quantum Mechanics is not for you
@@zzzzzzmc Superposition is probably the least weird thing about quantum physics. It makes sense, intuitively. You start trying to measure particles and you get REALLY weird results. But, even that isn't is complicated as people make it out to be. The problem is people try to imagine particles as little balls of matter, and they are not that at all.
Amazing talk, even better comment
One of the best explanation undoubtedly. One additional info: Now a days, the IBM composer (what he has shown, while creating Q-circuits graphically) is much more robust and added more bells and whistles., which is kind of expected.
A classroom full of smart computer science students all being silent when asked if they have any questions just means they were all so confused that they didn't even know how to go about asking a question.
My impression was that they had just taken him at his word when he quoted the "shut-up-and-calculate" guy. It's kind of like the suspension-of-disbelief that people do every day when they watch TV shows about dragons or superheroes.
They ain't students, most are scientists
They’re comp sci people and they’re trying to remember back to their college courses.
It's just that quantum mechanics is just such a severe level of abstraction from reality that they don't really understand it. After all, they are just e flat manilla computer scientists. No QM classes, very few with linear algebra or vector calculus background. Unless, you do research in quantum mechanics, it is a shut and calculate discipline. Its very non-intuitive. Schrödinger, Dirac, Heidelberg, et al. Were really bright guys.
Think it's more that this guy doesn't know as much as he'd like you to think. It's a common tactic to keep banging on about how everything is super simple when it clearly isn't, it makes the audience reluctant to ask questions because it'll show they don't understand. The couple of questions this guy did get asked, he was like a rabbit in the headlights and was unable to explain because it deviated from the script that was prepared for him.
Half way through the video, God, I saw so many videos about Quantun computers and none of them explained as clearly as this guy. We must find who he is and ask him to write a book or put a course series, It will benefit humanity very much.
Happy you enjoyed! I've written a couple follow-up blog posts:
ahelwer.ca/post/2018-12-07-chsh/
ahelwer.ca/post/2019-12-21-quantum-chemistry/
There are a ton of good quantum computing resources out there, but a whoooooole lot of bad ones. We just hope the good ones will eventually bubble up.
He's a great presenter and it's a really good presentation. Even as a non computer scientist this was really interesting.
As to the input/output confusion at 36:24
The words i/o refer to the names of two variables: i is *used* to store the input data, o is *used* to retrieve the result. The apostrophe ' indicates the state before or after applying the gate BB.
This is different from the one qbit gate, where there was no variable name for the one qbit, and i/o described the state like the apostrophe ' in the second example.
Yes, it really makes sense if you put it that way. I think he explained this naming convention poorly.
Another way to put it: "Input" and "Output" are merely the *names* of the *lines* -- "Input" is named after what it is used for *before* the computation, and "Output" is named after what it is used for *after* the computation.
I'm impressed that you managed to get through the whole talk without saying the term 'decoherence' once.
By far this is the best presentation on quantum computing I have seen so far. I watched many times and till day I keep watching it. Thanks Andrew!
two years later, it remains the best video on quantum cumputing, Bravo and thank you
Glad you enjoyed!
Thank you so much. This is the best intro to QC I've yet seen.
At 36:00 you get into the hack to handle non-reversible functions. I'd like to offer this intuition: The motivation for adding the extra qbit is to have a place to store the information that would otherwise be lost by the non-reversible function. We don't actually care about this information, but we don't want it to be lost, so we stow it away in Input'.
It's basically the same reason be rent storage units. We don't actually want all that crap, and deep down we know we will never use it, but we are unwilling to throw it away quite yet.
Presumably, this extra information is stored to maintain reversibility. Why does reversibility matter?
Joel Forsyth , this is a very curious feature of qm. My understanding is that quantum coherence is destroyed by any interaction that involves increasing entropy,. Increasing entropy means irreversible.
One way to look at this is to note that the arrow of time is defined by entropy change. Physical interactions can freely ignore the arrow of time as long as entropy doesn't change. This is why QM violates our "common sense" notions of time and space, and yet just barely leaves our notion of macroscopic casualty intact.
Mind blowing exercise for the reader: given an ideal quantum computer with the magic ability to perform arbitrary irreversible computations, construct a temporal telegraph capable of sending messages into the past.
@Hrithik Diwakar okay... here's another angle on it: a pretty good test for any hypothesis with respect to either relativity or quantum mechanics is causality. If you can do a thought experiment that violates causality, then you have an error in your model. A causality violation is anything that transmits information back in time (or faster than c). Physics doesn't have any causality violations. However, quantum mechanics includes phenomena that are logically indistinguishable from time travel. Why doesn't that violate causality? Because quantum entanglement is time and entropy symmetric. If one side of a quantum interaction has more information than the other, that would break that symmetry, allowing a message to be sent faster than light or equivalently backwards in time, thereby violating causality.
I hope that helps :)
interesting, thanks
I agree that this is definitely the most accessible intro to QC I’ve seen, but perhaps you’ll be able to clarify my confusion at 36:00; if you were to chain the reverse operation BB to the output of the forward BB, would input’ be input for the second BB?
At only 32 minutes in, things are already SO much CLEARER! Thank you for making this!
As someone with basically 0 knowledge about the topic this was awesome! Really good explanation, i actually feel like i learned something which is not always the case in some of these talks
This is the best quantum computing video I've found so far by a country mile. Fantastic presenter!
AHHA, finally a good explanation of quantum computing. After having found lots of good descriptions of quantum mechanics (bra-ket notation, hermitian matrices, etc, etc, etc) that left me wondering about computing, this video is great. Many thanks to the presenter.
This presentation is fantastic. This is the first one after which I'm legitimately interested in learning more about quantum computing instead of thinking "Man, that's weird"
Glad you enjoyed!
This is the best explantation / introduction of QC I've seen, and indeed very well targeted at computer scientists. Thanks for that.
Thank you so much, you are awesome! I've been studying tensor calculus for at least 2 years and never could grasp how the tensor product is actually calculated. It literally took you less than 2 minutes to explain it in a way I could clearly understand and visualize!
Well, it's been some years ago, but what he calls a tensor product is not really the definition of it. It's just one example, one representation. The definition is more like: glue two vectors together, and get a new vector. This object v×w (the tensor product) can now be multiplied by scalars which is the same as multiplying with this scalar in one of the components (v, w) and can be added with other tensors of the same 'rank'. If one choose a Basis for this vector space , where v and w where 2-dim, you can define (1,0)×(1,0) = (1,0,0,0) , ... but also as matrix (1,0 \ 0,0) when using dyads. Since vectors, matrices and tensors are elements from vector spaces, they can be represented as real or complex vector spaces. After identifying (which requires the basis) this is one tensor product, but the general definition does not need a Basis at all and can be used at different vector space (e.g. complexification). The wedge product (the hat like logical and, subspaces of the Tensor product) has one more special property: v×w = -w×v, which is useful in multidimensional analysis when calculating volumes in 3D, 4D ...
The RIGHT APPROACH!!! Congratulations!
The most helpful, effective and fast tutorial on Q-computing I've ever read.
the feeling this guy gives you is just phenomenal, such a great explanation and such amazing classroom charisma.
This is one of the best introductions to quantum computing that I've seen. Other teachers should use this presentation as a template to expand on. Thanks to whoever that guy is, he's a good (nifty) instructor!
I fell asleep and found myself here, i don’t think I need to know this, I’m an animator wtf
The beauty is that you can learn it and apply it to your field
learn it tho its super fun
I bet quantum computers could speed up computer animation?
It was hard to follow for me but my interest in this has grown due to this. I did run through this multiple times before it started making sense. I am still not 100% there but getting there. QC is really cool. I love it. Nature is so weird and wonderful. This was a great video. Thanks!
It'll be easier if you learn classical computing first. He does a lot of comparing to classical computing to help understand quantum computing, so if you don't follow the comparisons, you're probably going to struggle to understand the subject matter.
You should probably start with learning some linear algebra. Would be a big help
Gotta give this kid credit for his effort at explaining the unexplainable. Good job!
He is the fastest speaking lecturer I've ever listened to !
This presentation was pretty nifty!
pretty neat... ooookay
@Doido do Minescraft First, what is a "floop" to you ?
Best explanation I’ve seen so far. Nice job
Great video!
I was totally impressed that programming a QC is so easy.
The hard part is finding suitable problems and developing algorithms. We have to reinvent computing from scratch!
That's not really correct... Quantum decoherence is the biggest issue, which has to do with the device.
Currently they use an electric flux threading two concentric superconducting rings. This has a 90 microsecond decoherence time! Certain topological phases of 2D systems with certain guage symmetries support nonlocal objects called anyons. These have recently been shown as an avenue for robust quantum computation using their braiding statistics. In short the underlying device is the main issue.
@@marcusrosales3344 sure. I wasn't talking about the actual hardware. I am computer scientist, they know nothing about the hardware since abstraction layers have been invented;)
@@jpt3640 There are problems which are known to be solved better on QC though... Have you read ANYTHING on the subject? If not, you don't know, so you can't state anything.
@@jpt3640 Sometimes haters be hating
@@marcusrosales3344 "big issue" =/= "biggest issue"
This guy teaches everything a lot better than instructors in my university. Love you man, you saved my semester.
This is literally the best video on quantum mechanics, how could 275 people dislike it?
What a great video, thank you so much for that! It is also very good to see that someone so young may be so knowleadgeable about a high technology subject.
The lecture has been accessible but not oversimplified, great work!
At 39:00: "Why q-function needs output q-bit as input?"
type Qbit = {value: boolean}; type QuantumFunction = (input: Qbit, output: Qbit)=> void;
The function doesn't return anything, so it takes an "output" argument/"pointer" which it can modify (using the input).
Such a modification might also be affected of the "output" initial value, which is why it is supposed to be initialized to 0.
eg. "NotOperation = (input, output)=> input? output: !output", is only a NOT-function if the output is initialized to 0.
Helpful metaphor?
Yes, that is a great metaphor! I wish I had thought of using an "out variable" when I created this presentation.
finally. I've watched and read so many things that just talk about the fact that superposition is a thing without actually going in to what it means for computation
At ~40:00, it becomes very obvious when you call "Output" as "Output holder" instead -- makes understanding it a lot easier
I hope to watch a video on shor's algorithm by this guy! This is the best video I've seen so far!
Shor's algorithm can (in theory) find prime factors of very large integers. If we ever have a reliable quantum computer with MANY qubits we will be able to break military grade encryption in a fairly short amount of time. That is currently impossible using classic computers.
For a young guy he sure knows a lot, He must've been studying quantum mechanics in kindergarten
He was and he wasn't
@@mmd1080 oof
This stuff is fairly basic tbh. You learn way more than this in an undergraduate Quantum course.
@@brandonberisford stfu
@@5ithofnov159 you good?
Great slides and great talk! Andrew needs to create a follow up presentation that goes deeper. Maybe explain a few more quantum operators, gates, algorithms. etc. Some discussion about various physical implementations of qubits (including topological QC) would also be helpful.
I thought he was presenting to an auditorium full of people and when the camera showed the audience there's like 10 people there in a small room! Pretty cool that nearly 1.5 million ppl have seen the vid and nearly 2k comments so there is definitely interest. Very well presented by this young guy.
Perfect Video start learning quantum computing, the actual math behind it, rather than fancy statements. Well Done!
Thank you for your mathematical approach of e,g, the difficult phenomenon of "coordination" of Qbits. Our language limitations only reflect the fact that we do not really understand quantum physics. It works, but we don't know why it works. Mathematics only define a model.
Loved seeing a quantum computing explanation that wasn't afraid to get into the mathmatics
Every other explanation I had seen so far barely talked about how this could mathmatically work, all they said was "well, is both 1 and 0 at the same time"
But actually understanding the complex tensor representation really helped
However, I kind of got curious as to how the qbit could work in complex states
I believe I should start my studies in computer science next year, since it's my last year at high school, so I believe I am going to be one of the pioneers of actually puting this new technology to use in my carrer, really excited
yes, matrix manipulation definitely helps understanding. The Deutsch Oracle case gives an idea on quantum advantage, but there are other factors including which algorithms can really benefit and how long will it take to deploy them in production. Yet I see VW , Airbus and CERN are serious about QC.
It's times like these I wish I knew more. It always takes me 2 time longer than the actual video length to watch informative videos like these.
I don't know what he's drinking but it makes me want some fruit punch.
Hey it's quantum computing, not the Kardashians' ;)
lol. I couldn't watch them for more than 5 minutes. Watching them is like watching monkeys argue about which shape fits in what hole. It's irritating.
Instead we are trying to understand quantum physics, isn't that a lot to be proud of???? 😍
Don't worry man, I've been studying Physics since high school for about 6 years now and specifically studying quantum computing for about 4 weeks and I feel that that's barely good enough to watch the video and understand it all without stopping. Furthermore none of those guys in the lecture seemed to get it.
Imagine you have to make a decision. You can either do x, or you can do y. Then, 5 minutes later, you must also do either x or y. And so on, for an arbitrary amount of time. You won't know the factors that affect each decision until you reach each decision. But if you write each decision on to a paper, you can predict many possible paths, x-y-x-x-y-x-y-y-y etc. The superposition is the sum of all those paths. You cannot possibly know which path you will take, same for the the next person and the next person after that. But you can predict the probability you'll each end up in certain places.
Finally an explanation without the metaphors. I was so confused by all the garbage videos that ended up conveying no information to me, but now everything makes so much sense.
Awesome video! And watching all these older people, who're probably smarter and more educated than me, asking the same questions I do goes a long way toward not making me feel stupid.
@30:30 "Its nice to know that if you use complex numbers our diagram is a shpere" HOLY SHIZZZ. dude that really blew my mind.
A 3-dimensional sphere (unit vectors in 4D): www.wikiwand.com/en/Special_unitary_group#/Diffeomorphism_with_S3
This is excellent! This is exactly what I was looking for to gain a basic understanding.
Hello there's this trader you should know she's an expert in trading she as made me wealthy, Ever since i started following her strategy & techniques things as be going well for me, Contact via Gmail @hauserjolietradingoptions and thank me later..
She's available on Whats app:
+ 1. 2. 5. 3. 5. 5. 9. 0. 6. 8. 8. ..
Still one of the best explanations out there, well done!!
Hello there's this trader you should know she's an expert in trading she as made me wealthy, Ever since i started following her strategy & techniques things as be going well for me, Contact via Gmail @hauserjolietradingoptions and thank me later..
She's available on Whats app:
+ 1. 2. 5. 3. 5. 5. 9. 0. 6. 8. 8. ..
This by far the *best* introductory course on quantum computing! If this guy teaches at a college or university, he should be getting very high marks.
Hands down, the clearest explanation of quantum computing there is.
The can is empty yet he keeps drinking.
If it was a snake it would have bit me!
There is a dealive cat inside.
m.th-cam.com/video/ramTL-k5Vpk/w-d-xo.html
It's both half-full and half-empty at the same time!
I once drank too much from a half-full, half-empty beer can, felt like I was in a superposition, became entangled for a while, then collapsed...
I never thought i would say this in my entire life but Thank you microsoft for this vid
I agree with you.
It' the guy. he is honest and likes to share what he knows which is the opposite of what Microsoft does.
Hopefully, he doesn't work there for a while otherwise he will learn their culture and he will start to explain stuff as what Microsoft did in the Microsoft helicopter joke!
slides
1:21 why learn
4:33 learning objectives
7:37 he got me here lol, started saying 2bit ops lol
10:48 tensor products of values (missed a slide oops)
...
24:30 Hadamard Gate
33:11 Deutsch Oracle
52:00 Tricky bit on Tensor Products
I'm still struggling to understamd the thing with the Deutch oracle. Yes we got one input of 2 qbits, but the samecould be done with a classical. If we simply send a 1 and a 0 at the same time all will be clear. However, if we only send one at a time, why did the algorhythm shown in the video has 2?
This is a very clear introduction to the topic. I think a useful way to describe the hack for non-reversible functions is as follows: one can't build a "block" using quantum mechanical circuitry that implements a non-reversible function f. But, one can implement a block that takes two input qubits, O and I, and produces two output qubits O' and I' with the following property: if supplied a zero for O and some input x on I, it will output f(x) on O'. One can use this building block in a larger circuit, ignore O and I', and get the effect of a block that computes f.
This video is, by far, the clearest explanation that I have been searching on TH-cam. Thank you very much!
Very very good tutorial (I am a computer science Ph.D. candidate). One interesting piece of information, the instructor of this video left Microsoft.
Correct! Decided to try the independent contractor life.
41:00 He should have just labeled the real input as input and the used output as output. The incoming 0 and outgoing the input don’t need labels.
As far as I understand the phenomenon of entanglement collapse is only about the actual knowledge of the states.
I guess a better picture to simplify the would be a + b = 1, with a and b being natural numbers. The time you set (measure) the value of one of them you automatically *know* the value of the other and that's why it's not possible to send information through them without entangling the variables again.
Thank you for the excellent presentation and please correct me if I'm equivocated.
1:01:42 Ok. So he basically explained here that this is in fact not what happens at all.
that example (as almost all) fails to explain the essence though which is that the numbers a and b were not predetermined until you looked at them. It's not like some third person wrote 'a' on one paper and 'b' on the other and gave it to 2 different people to look at when far apart.
It's more like I'll give you 'a' then you flip a coin and decide what to write on it and the other person will get what's needed to satisfy the equation. And the confusion comes from the part where you try to figure out how the other person knew what you flipped.Or the other person's 'b' paper rather.
That's called the local hidden variable hypothesis. It was disproven in 1972 by showing that some entangled quantum systems can violate bell inequalities which must be satisfied for local hidden variable models. So it must be that entangled quantum systems can have superluminal causal influence (nonlocal) or that they are fundamentally undetermined until measurement (no hidden variables).
Its nice to see a technical presentation on QC with real content that one can learn from.
A very nice, easy to follow and exhaustive introduction. Really love the speaker!
can you tell if he has is done with the drink or not? I bet he is in a superposition...
gg
Aw, Sarcasm
Quantum cola. It’s both empty and full
@@overseer3072 Schrödinger's Coke ! :-D !
excellent presentation! I thought it would be quite terse and boring, but turns out fairly easy to follow and very interesting.
Andrew, in the video you mention that it'll be a few years until we make or break it with QC. Do you have any new insights on the exponential term and the scaling of QC now?
Not really. Google claimed quantum advantage but others dispute it.
One of the clearest explanations I've ever seen
Thank you TH-cam for recommending this. Been trying to understand this whole QC thing the past month...
If you made this into a series that would be amazing!
7:43 this is so funny to me for some reason, i thought he was talking to a room full of people
he is but the room is in superposition
@@zzzzzzmc whoa
It will be when it's near and theres a business opportunity
Just like me during my presentation, I tired to be funny but no body laughing. Keep it up
people working at microsoft are probably the only people using edge :)
Very well explaining presentation. Thats an introduction that i wanted from someone who really knows whats going on with quantum computing. Thx for sharing
1:05:26 - I'm still confused about that myself.
Maybe you can't transfer data through the entangled particles but if you can collapse it by measuring it at will, if there is a way to tell when that other particle collapses then the collapse itself can be the communication.
I'm guessing what he's saying is that there is no way to tell when the state collapses so you can't use it to signal?
Yes correct, you cannot tell whether the state is collapsed.
32:22 - "Imagine I show up on your doorstep and I give you a package, its just a black box that has a function on one bit, what a horrible present" 😂
43:24 I think the output input thing could be more easily understood by C programmers, since we have the concept of returning through parameters.
I think they were just having problem understanding because he didn't clearly state that the outputs and inputs were physical qbits
Good stuff, wish he'd covered more than 50/50 probabilities. Would love to see more from this lecturer
Excellent presentation!!. This is one of the clearest webinar about this topic that I have ever watched. Thanks so much!!
40:00 The confusion here is that there are two qbits, one named 'input' and one named 'output', but these two are both passed into the quantum black box to be operated on. The 'input' qbit is left unscathed after the black box, but it is the 'output' qbit which is rewritten with the value of [the possibly nonreversible function f : applied to whatever the value of our 'input' bit is]
Great intro to quantum computing. I finally understand WHAT IT IS!. The really distracting part is his drinking from an empty can!
i know right. started great but then became a bit insecure. good talk i agree
The can is in superposition and is both empty and full at the same time that’s why
Watching this on valentine's day. Freaked me out for a moment:D
this kid is actually very funny without knowing it, as well as an excellent lecturer
yea the kid effectively communicating quantum computing doesn't know when hes being funny
It's Microsoft 😂
Great Video! I have looked at various articles and videos online which described pieces (generally confusing) of the quantum computing puzzle. This brought it all together and I now feel I have a solid basis to go further.
Take a shot every time he takes a sip. Helps understanding by 5000%
He's so young, yet so adult at explaining and reasoning. Wow.
Even Microsoft can't get a competent cameraman? Point. At. The. Slide. Not. His. Face.
If these algorithms can be modeled with the unit-circle state machine, why can't a classical computer run that algorithm? Or can it, but just not when using complex numbers?
Joe Cleveland classical computers can run any quantum algorithm, with exponential slow-down in certain scenarios. It is actually true that any quantum circuit consisting only of Hadamard and CNOT gates can be simulated in polynomial time with a classical computer. If you add another gate called a Toffoli gate (basically CNOT with two control bits) then you can generate circuits which take exponential resources to simulate on a classical computer.
thanks for the nice reply, i just this answer whispered in my mind while reading above question.
You can see that for example the entanglement, it cannot be factored into two qbits but have to be stored as a 4-vector. Thus cannot be represented by the unit circle state machine
Quantum computing also is a finite memory turing machine.... Just like classical ones.... Hence both have similar computational power but classical computers get exponentially slow for such superposition driven algorithms
Thank you so much!! I've watched so many videos on this topic, but after I've watched this one, everything became truly clearer. Great presentation! (till now, I've also had 5 courses at the University, and didn't get anything
unfortunately)
This is the best lecture I've seen from this channel
Arbnora Miftari is he drinking beer 🤔
He goes through this fast, but look at the participants. Some of those guys likely have PhDs with long shaggy beards.
Tim Volk "Alright! Next is superposition, any questions? GOOD."
Exactly! 😆 The empty can and the belittling of the audience *Okay, got it? Look at this shiny coin! Any more questions you dummies?* lol.. Yeah they all want to know why flipping a coin and writing both possibilities on two pieces of paper, then separating the two answers over a vast distance is so amazing! "Your coin was heads? No way! Mine was tails, and it collapsed faster than the speed of light!" 🤣
And I answered "yes" since the first "is anyone having trouble with this so far?" :'(
this speed is how i got normal computer science tbh,, the teacher would write while speaking about the ands and ors the nands and xnors etc with one hand on the white board ... and in the other hand he would have the white board eraser to make space for the next block of text lines and formula's