A Beginner’s Guide to Quantum Computing

[Music] all right the good morning maker fair thank you for coming early and kicking it off with beginner's guide to funny computing I'm Talia Griffin I'm a research scientist at IBM at the New York office and I'm actually a beginner myself I just started learning about this six months ago so I'm here to tell you that anyone can get started learning about quantum computing and hopefully after learning a little bit about it this morning you'll be motivated to learn a little bit more so I want to start with two facts right back number one classical computers have enabled amazing things the internet they're part of how I got here I was able to fly across the country on a plane right the electronic giraffe outside blaring music right classical computers have been able just amazing things but one of the things we don't often talk about is all the things they can't do right we talk about all the stuff they can do there's so many things I can't so I want to start with two examples of things classical computers are really bad at and you know maybe they can solve small versions of these types of problems but by the time the problem gets big enough to be interesting we just run out of computing horsepower so the first example of optimization right optimization is I want to find the best solution to a problem among many possible solutions all right so here's a picture of a table this is the table at my wedding it's one of many tables but you know you can see 10 people around a table you've sent people over for dinner how many different ways are there to configure 10 people around the table all right the answer is 10 factorial right the number 10 seems so small but 10 factorial is 3.6 million there's 3.6 million ways to raise ten people for dinner right next time you guys have dinner with 10 friends share that fun fact with them right so so when we go and do this we consider some of the options we make approximations right because it's the only way we're really going to end up seating people at all right but you know the truth of the matter is every time I add one person to my dinner table the number of possible configurations grows exponentially all right so we can solve small versions of this problem on classical machines but we don't solve big problem big versions of this problem very well at all second example is chemistry this is a picture of a nitrogenase enzyme you know anyone who has ever eaten food should care about this enzyme it's an important catalyst for the creation of ammonia which is an important component of food fertilizer pharmaceutical than many other things so this is the enzyme I've called out three molecules of iron sulfide clusters in this enzyme of different sizes the one on the Left four iron atoms and four sulfur atoms believe it or not this is the biggest of those iron sulfide clusters we can stimulate on the biggest supercomputer that we have right it's so small what why is that the biggest molecule of these three that we can that we can actually simulate on a classical machine and the reason is because to actually simulate what's going on in that molecule I have to account for every electron electron repulsion and every attraction of the electrons of the nuclei and that number grows exponentially the bigger the molecule right every single electron exerts an electrostatic force on every single other electron right so when I add another one I got to recalculate all the electron energies right so these two bigger clusters they look so small we can't simulate them ok so there's actually many problems that have this characteristics and what they have in common is this this idea of exponential scaling so there's a there's a classic stable about the power of an exponential and the stable goes you know the Creator as a game of chess brought the chessboard to the Emperor and the Emperor said I love this game what can I give you as a reward right and the craftsman said ok there's 64 squares on the chess board on day one give me one grain of rice and every day after that double the number of grains of rice I get right so on the first day Emperor gives him one grain of rice the next day to the next base for the next day eight and after a week he had a teaspoon full of rice but after a month he has the rice production of a small country and after the full 64 squares it was Mount Everest right right so it grows really fast the number 64 doesn't sound that big but two to the 64 is an enormous number so why do we think quantum computing is actually going to allow us to solve some of these problems we can't solve classically right so it boils down to two fundamentally quantum effects one of the effects is superposition so classical information is basically a string of zeros and ones you know everything that classical computing has enabled is you know boiled down to a sequence of zeros and ones so quantum computing or quantum information has this property that the states can exist in a superposition of 0 and 1 right so not just 0 not just one but a superposition of 0 and 1 right and you can also have complex superpositions of 0 and 1 so you start to be able to explore much richer set of states so if one qubit can be in a superposition of two states then two qubits can be in a superposition of four states and three qubits can be in a superposition of eight states so the possibility space you can explore is much more interesting and complex in quantum information so that this diagram on the right is showing you a superposition of five cubits five cubits right so you can be in a superposition of 32 states so superposition is the first thing the second thing is entanglement this idea of entanglement is okay I've got two qubits and I'm entangling them together so measuring the first qubit can tell me something about what will happen when I measure the second Cuba okay so entanglement is the second property that gives quantum information a really unique difference so together this allows us to totally change how we run algorithms right so take the optimization case if I'm just going to consider three point six million possible ways of configuring ten people at a table classically I have to consider each one individually and then I have to compare them all right here's how quantum computing is going to solve that problem you take your cubist you go into a superposition of all the possible states all the possible configurations and then when you encode the problem into your quantum computer you're applying a phase on each of the states the phase is you know that kind of access towards the center of that sphere you saw on a previous chart you code a phase on each of the states and you know when waves are in phase the amplitudes ad and the waves are out of phase they cancel right so when you have noise cancelling headphones what you're doing is you're creating noise that's exactly out of phase with the noise you're trying to cancel right so in quantum computing you're going to go into a superposition of all these states when you encode the problem onto the machine you're applying a phase I need to the coupon each of the space and then you're using interference you amplify some answers and you cancel other answers eventually arrive at the solution so it's just totally different I just got a completely rethink how we're going to solve these problems and that's that's kind of how quantum going to do it so it's obvious why the number of qubit matter keep its matter right if I have one qubit I can be in two states in the more qubits I can I can have I can be in a superposition of two to the end States but another important factor is this error rate all right I have to be able to control what's going on in the qubit if I have really high errors and all of my operations don't work out as I expect them to then that's not going to really work so we're promoting a new metric called quantum volume where we're saying okay if you increase the number of qubits you can get to higher computational power but not if you have really high error rates so we have to both move towards lower error rates and higher cube accounts okay so how do you actually build a quantum computer right this is how it should work theoretically how do you actually go and do this in real life so first of all you have to have qubits that work in such a way that you can harness quantum mechanics so we build basically artificial atoms you know atoms behave quantum mechanically we build an artificial atom and we make it out of a superconducting Josephson Junction coupled to a microwave resonator okay so this is what it actually looks like on the chip you have these squares that are your cubist and these squiggly lines are your microwave resonators and inside of the qubit is a superconducting Josephson Junction and we got to cool this thing down to point zero one five Kelvin where zero is absolute zero right room temperature is 300 this is significantly colder than outer space and we talk to the qubits with microwaves so this is what it actually looks like this is how we talk to the qubits we have inside of a dilution refrigerator which you'll see on the next chart we have all of these microwave cables that allow you to actually go and probe the qubits with microwaves okay and this is what it looks like in the actual lab so these giant white cylinders these are our dilution refrigerators and here you can see my friend Nick working on one of the insides of the of the quantum feeder so great that we can actually do this but did you know you can play with one for free today any time you want at our booth when you go home you can go and you can play with an actual quantum computer we plugged up a five cubic quantum computer in the lab you know through those microwave cables up out of the fridge through the internet available to you at this website for free so you go to this website when you do you're going to find three different user guides one is for total beginners the beginner's guide is no mask just concept I'm going to show you visualizations of how the qubits work we're going to walk you through early examples you know what are the what are the gates you can use how do they work what how could you think about them and then you go into the user guide the user guide is going to show you some of the linear algebra you need to know it's going to show you how algorithms work we're going to give you a suggestion of five different algorithms you can run we're going to show you what sequence of operations you need to actually implement that algorithm and then when you're really interested you're going to go into the developer guide this is going to take you to our API online it's a sequence of Jupiter notebooks that are written in Python and you can actually send that job to the quantum computer free just go and learn right we're putting it out there for you to learn this is what it's going to look like when you click in to the composer tab on the previous chart you know there's three tabs there's the composer the user guys in the community I'm going to talk about each of those this is what it looks like when you click composer you're going to get to choose you want to simulate it you can choose to simulate where you actually start with like two qubits instead of five because it's easier to get started learning with two or you can simulate up to twenty cubits with a simulator up to 20 but let's click into the real device so clicking in this is the interface you're going to see this is a picture of the actual chip that we have in the lab and see those five black squares are five cubits and the squiggly lines again are those resonators that we talk to the qubits with those resonators and here we're giving you some information you may want to know about the cuba's what's the coherence time coherence time is how long does my quantum information last before it gets it gets out of out of coherence so for us you know we have 50 to 100 microseconds of coherence time and that matters because it determines how many different operations can I do before my I lose my quantum information ok so now let's create superposition we talked about superposition let's created you drag and drop the different gates that you can perform their different logic operations that would be shown on the right just drag and drop it onto the graphical user interface here I've dragged the H gate to the Hadamard gate onto the score the line representing the first qubit and I've done a measurement and actually these are results I got this morning having run a thousand experiments on the actual quantum computer this morning and you get roughly 50/50 zero and one right you're in a superposition of zero one when you measure it randomly chooses half the time to choose a zero half the time to chooses one again if I'm in a superposition of about two qubits and superpositions and I have four different possible outcomes 0 0 0 1 1 0 1 1 so if I measured a thousand times I get a distribution of each of those possible outcomes and I can create entanglement this is how you create an angle 'men you put your H gate and then you apply an entanglement gate this is a C not what it says if I measure the qubit to be 0 don't do anything if I measure the qubit to be 1 flip the second qubit to 1 so this is called a bell state and this is an entangled pair of qubit so you guys can actually go in and create an angle min on a real device in our lab in New York through the internet which I think it's pretty cool ok so now once you get through the beginners guide and you're really keen going to the school user guide this is showing you some of the algorithms that we are enabling you to learn about through through the full user guide if you click into Grover's algorithm you can get a whole derivation of what's going on in that algorithm and we show you this is how you should think of up at quantum interference that amplitude amplification right and this is a sequence of gates you need to implement Grover's algorithm on the real device you can go in and you can actually implement that algorithm on a real quantum computer so all of this is made possible by a bunch of really committed and passionate makers at IBM this is a drones eye view of some of the members of the team outside on the lawn in New York it's a lot of people that put in a lot of hours to make this available to all of you for free ok and so what are people actually doing with it we have over 45,000 users worldwide this is a snapshot of some of them the white dots represent some of the top universities with the most number of users in the quantum computer a lot of universities are using it to teach we have a teacher at EPFL who is it as part of his quantum information curriculum we have a teacher at UT Austin got teachers all over the world using it and then we have people just doing really cool and interesting fun things with it there's a woman christine moran she's at dot at the South Pole in between measurements on the South Pole telescope she's just playing around with a clinic Evita was really cool you know we had a workshop we telecast a workshop from our Zurich Lab to South Africa so at three high schools in South Africa is having a quantum information workshop which is really cool and this is what those people actually look like we had a cruise Christine in the bottom right at the South Pole this is the students at the workshop in South Africa every summer starting last summer so I guess it's only the second year in a row the University of Waterloo runs a quantum information workshop for their undergraduates there's a woman Emily who you know she has a youtube video that was I found her on YouTube she's talking about what it felt like to actually run the algorithms on a real quantum piece of quantum hardware the first time she's been learning about in a classroom she actually got to run it on a real quantum device which you know she talks about being kind of an awesome experience or so people are actually using this thing for research there's been 17 I think at this point 17 research papers that have been published where someone validates their algorithm on an actual quantum device in a lab people have made games there's a guy who created a game called quantum battleship you know you actually actually run the you actually run the game on the actual processor so you know he's got a blog about it and he walks you through how to actually program the game how he programmed the game you know at scientistic that's really cool and people are having fun right so we had the third tab that I mentioned was the community tab so you know in the community people are asking questions and people in our research team are answering them right so how what is the C na key how do I think about that from a mess have a point of view what is what how come when I have this sequence of gates that doesn't match between experiment and theory and a lot of the time you know because not every gate has perfect accuracy if you apply too many in a row then the answer starts to look more randomized because the errors add up but then you know any community is made up of people who like each other and they like so you know joke around and have a good time so that you know some of the posts are like what's your best quantum pun so you know they suggest a couple of drinks like gin entanglement which I love and you know how do you implement rock-paper-scissors on a quantum computer you know we all we've all played rock-paper-scissors there was a whole long discussion on one of the community threads how do you actually implement rock paper scissors on a quantum computer and actually you know I'm in the slag channel where members of our team you know answer the question it was like a whole discussion an actual like theoretical discussion about well what is the best way to commit Rock Paper Scissors so in is is fun you know people having a good time and as of last week we put a 16 cubed device online for beta users so this is you know this is the most advanced public quantum computer available anywhere and it's free and if you want to sign up as a beta user you can just go in and sign up and you know we also had an announcement where we doubled our quantum volume so not only did we announce 16 qubits we have another device with double the quantum volume and there's a picture of Katie in New York on Times Square next to a next to one of our quantum computers so I encourage all of you to visit our booth we have a booth in building 2 you can come and learn we have at least two different demos that explain to you how does a qubit work we have a gyroscope you know that's supposed to be able to show you what it you know what is how to think about you know mapping a gyroscope on to this thing called the Bloch sphere that you'll learn about in our beginners guys and we'll show you a demo of the experience right you can go in and you've got that v cubed interface we've got it up at the booth you can drag and drop gates you can create entanglement and a real device online and we've actually hooked up also some LEDs so that you can actually visualize the results of your experiments you can run it a bunch of times and see how that quantum randomness is that arbitrary measurement of 0 and 1 if you're in a superposition how does that actually look when you run it and you you show the results with some LEDs so I've shown you a bunch of stuff that we've made and what I like is for all of you to go off and make something you know we've opened up this quantum computer done it for a few reasons one of the reasons is we'd love to know what you want to do with it what are the things you guys want to do with it right we encourage all of you to go off and make something and share with share your ideas with us and we'd love to hear about it so thank you all for your attention [Applause] you [Music]

Channel: IBM Research

Views: 313,412

Rating: 4.889864 out of 5

Keywords: quantum computing, IBM Q, Maker Faire, IBM Research, qubit

Id: S52rxZG-zi0

Length: 18min 43sec (1123 seconds)

Published: Wed May 31 2017