The Quantum Zone - Dr. Anne Ghesquiere and Dr. Freya Wilson

first of all um marks just done a fantastic job of introducing me there and I my name is is Freya afraid and you can get me on Twitter here although I don't tweet very often but I keep trying to tweet more and I'll go through fits and starts of it so feel free I'm currently a penetration tester for SEK 1 or clarinet as we're now part of as well they let me come here on the proviso that I made my slides look nice and read and also that I say we're hiring so if you're interested comment let me know and I I feed it most of the stuff when I was at Leeds University so I did physics

degree and then I stayed around and didn't get a job for a while longer under the PhD and that was in this sort of stuff from that I kind of did a sidetrack I worked for Hawaii for a while doing security research there and then I came back up to Leeds to do pen testing at said corn we got questions about any of those things feel free to ask me at the end and so I do have the University of these logo there very briefly so that and also EPSRC I should probably credit them slightly because they funded my PhD but that's why the slides are all right I'm okay so I did a lot of stuff in quantum physics

and I find it really hard to understand like I we were talking about this in the break like there are bits of quantum physics where and I'll read it and I'll understand it and I'll stay in my brain for about an hour maybe two hours something like that and then it slowly kind of exponentially leaves my brain and I have no idea what it is anymore and then I have to read it all over again so actually this is quite frustrating when you're doing a PhD when you have to keep rereading and like reassuring yourself that you are doing the right thing and you haven't gone crazy in between maybe I did go crazy in

between that also could be a valid thing but it's it's really really tough stuff whether or not you've been doing it for a term at degree level or for four or five years and for PhD and beyond as well and so what even is quantum area what does quantum actually mean so the best way to describe it is we'll start with what the word quantum itself actually means and quantum just means the smallest possible finite amount of stuff okay so a quantum is a small smallest possible finite amount of something robbed of whatever it is so a quantum of beans is a smallest possible finite amount of beans you could have okay and and the word quantum gets used

really really badly and it makes me really mad and actually some of my friends send me pictures of this sort of thing all the time because they're now how wound up I get about it if you come across any at the end of this talk feel free to tweet them to me I'm debating making it Twitter account dedicated to just bad quantum so here a few examples here so this is the smallest possible finite amount of pan if you can't see it as it's written there don't worry know what that means the smallest possible finite amount of dry and if any of you you sure and you might you know want to be aware that evidently they're not then

it's not that dry it so I'm making that dry it's the smallest possible finite amount of dry and I don't even know what this means I'm not really sure that and that means small as possible amount of soft toilet roll and this is my personal favorite because I think they're probably actually right scientists makes small as possible towards a security kind of internet and so actually this is all essentially rubbish people have taken this word and I've got excited about it and I like to put it in front of things because they think it makes them sound clever or their products and sunshiny and a new and in some way with of some kind of

technology involved that actually isn't all that relevant and what does I actually mean in the context of physics so we start with what things normally are so classical is what we call it but when things are completely normal and we're talking here like if I hit this table I'm expecting this table to stop my hand unless I you know suddenly super strong and I break it but I don't want to go run breaking things I think mark might kill me and you know classical things they they behave the way we're expecting them to it's in the macro world like we're talking about things as big things tangible we can see them we can touch them we can feel them and then

when you take that down a little bit and start to see what actually makes up this stuff what is this table made of what when we take it right down to the smallest possible finite blocks of matter we can things start to act really weirdly and they start to do crazy things and we don't really know why they do the things that they do but they do and if you put lots of them together then they create this macro effect that's classical and we can understand it quite well Newton was very good at it he gave us lots of equations and and things like that but and yeah so when we're looking really at the small stuff it goes a bit weird and

so quantum and the physics sense just means we're looking at the smallest possible finite amounts of matter and there is a limit to the smallest possible amount of matter that we can have as well based on I know how much energy it's gone that kind of thing and so we can sometimes try and use the really weird things that stuff does when it's really small I'm tired Vantage and at that point we might start developing new technologies so there are certain kinds of particle that you know smallest finite amounts of particles that do you kind of weird things and we can put use that in coatings and stuff like that for example silver particles I have certain

antiseptic properties and at that point is it's it's a quantum technology it's based on the quantum effects and we can do the same thing with computing and with cryptography as well so talking about that a little bit more I just say here that there is very little maths involved so hopefully you don't need a degree in maths or physics to actually follow what I'm saying but if anyone does get a lot of it lost feel free to shout out because I can't remember what it was like this yeah just just let me know and if we imagine that we've just got some water say and we we send it through two slits we would expect the

waves that that water has so we're assuming this is maybe like the sea or something like that and there's a wall and there's two holes in that wall and the waves would go through but they would do something a lot bit weird like bits where the waves crash together they would make a big wave and if you had the dip down part of a wave and the high up part of a way meet at the same time that would just disappear to nothing and you would get a sort of pattern like this so the white bits are where the waves have come together and made a big one and the dark bits are where kind of the down

part of the wave and the up part of the wave have come together to be nothing at all this is called an interference pattern and you know this is kind of what we expected for too something like water in a classical sense and if we were to take a load of part of tennis balls and we were to throw them through the same slits the same holes in that wall we would expect it to hit the wall the two points where the holes are right so tennis balls are kind of like particles and just throwing them through and they just hitting the wall and you're getting these two bits build up whether the tennis balls hit

and there might be a little bit of variance than that it won't be exactly in line with where the slit is because some of the tennis walls are lovely bounce off a little bit but it's gonna be in these two separate blocks that you see them here and if you were to do this with light um you know if you did it one at a time and you measured it in the right place and especially if you only took like if you were only looking through one one of those holes and you blocked the other one off and you're just firing your light through that one hole you would get this kind of thing you might get something else in there

called a diffraction pattern but we won't worry too much about that and what we're saying is you'd get this one big blob but if you took that if you opened up that second hole in the wall you'd start to get the same thing that you see when you see waves so you kind of you've still got these things but if you're sending individual tiny bits of light called photons so photons are the smallest possible finite amount of light and you'd still see the individual blobs hitting the screen but it's in the same sort of pattern that you see waves I'm so it's kind of got this weird nature where it's doing something that's a little bit of both and we don't really

understand why we can describe it quite well with equations and math so you may or may not have heard of the Schrodinger equation and and things like that we can we can talk about it quite well and we can predict it based off that as well because we have this these really good mathematical models that say we can predict if we cover one slit then yes it will it would only give us you know one blob and and whatever every opened it up then we would get these its interference pen we can describe it doesn't mean we actually have a clue why that's happening and so yeah if you if you do it with lightly and it's going

through both you can see how it's kind of going out here into that kind of pattern and this is actually from some real experiments that been done and you can see that you've got starting to build up here I'm so here a there or actually it's labeled B that one is where you know there's only a couple that have been sent through but as you send more and more through you start to see this pattern that's like this interference pattern but each of the individual white dots on there is a separate individual photon and and it's kind of interesting is you know you can you can see here this is kind of describing both of them I'm so you're

sending your individual source of particles through and if you were to hold your telescope or your detector over one of the slits and it's quite close you would you would see a kind of this sort of things if we cover that one over and just measured it here see this and vice versa if we just still up at the bottom one but if we open them both up and we can see kind of this mis'able pattern and interestingly if you just take the photons one at a time and you don't fire lots of them all at once and you just fire it and then you wait a little bit and then you fire another one

and you can set some kind of trigger here to detect whether or not it's gone through the left side or the right side and and you would see that it's gone through only one of them but if you waited long enough and you did enough you'd still build up this pattern where it's got all of the different printers and it was kind of weird and we can use this to our advantage in quantum and that sort of thing so I'm gonna sort of build up a concept here where if we're looking at all of the ones that just went through one of the slits we're going to call that X 1 or 1 and if we

look at all of the ones that go through the other slope we're going to call that X 2 but are we if we look at the combination of them we can combine those together and say that they are and I think the right word like entangled in some way so here we've got like um this is how we would this is what we'd call them so this is the ones I've gone through the first slip we pulled in zero in this instance and we put the funny little brackets around them and just to define that that quantum and and if we look at those specifically that's you know it's just going through one of the

slits and we can imagine that there are those are the ones that are kind of having this kind of effect and if we look at the ones that have combined and we've added them both together they're the ones that are kind of having this sort of effect so when you look at it and you're seeing this and you're looking at it from that perspective it's this sort of description that we're looking at and when we're looking at them through the perspective of it just being discrete separate lumps that's when you've just got the zero and you've just got one and that's kind of a little bit like bits so you know you're nearly all computer e-type people you have a 0

and a 1 bit and we can have the same thing we have a 0 bit and our one bit we've just put funny brackets around them because we're using quantum physics and then we have this other concept as well where we've we've added them together and we can call that a qubit as well I'm so if it's if you're using quantum physics in this way we call them qubits m so I'm gonna talk a little bit about what quantum computing actually is and what sorts of things we can do with it a man's gonna talk a little bit later about how good it is or it isn't and what sorts of things are out there so

you've got that to look forward to with an so we've talked about how we've got stuff going through one slit and we can call that zero and if we wanted to we could and describe any particle any photon as having some properties that you could look at it from the perspective here it's got the 0 or a 1 and here we can say it's pointing up or it's pointing down or we can look at it from the perspective of it actually being a little bit in all of the different places it just has a probability that it's going to look in that interference pattern and within that we can kind of describe that and by this funny symbol

here that so the funny symbol there that's just that 0 plus 1 stuff that we were looking at in a certain combination and and we can use these qubits to build up gates as well so you probably are familiar with the concept of a not gate while you take a bit and it's a zero you turn it into a 1 that's a warn you turn it into a zero and gates are essentially ways to manipulate bits in the same way quantum gates can manipulate quantum bits as well so we can have the same thing where we transform our proton we can do some stuff on it we can put some actions on it and move it around and

make that that movement into a gate as well so for example the pearly gates takes your something that's pointing in that direction there's a zero and make it point in the other direction as its point is that the uncertainty principle yes yeah same person Polly yeah and what does it do on on non integer values um that is quite a good question it is quite a complicated one and do you go into that all in your pants no you don't okay and it's essentially it's a matrix transformation so if you've got like non integer values or you've got a combination of the two and they're going to have different values at different points in the matrix and the paragate

will transform that matrix into something else and a sort of slightly more complex way we can talk it a little bit more at the end if you want yes yeah cool awesome and so as well as that ex-gay that not gate we have these others and they they don't really translate into the same kind of like and Gates or gates that we have in like classical computing but they can do slightly weird things and we can do something called the Hadamard gate which is where we look at it from a completely different perspective and we take two and we can combine them you can build these up to make different circuits and different circuits can do

different kinds of things different kinds of algorithms and I think it's going to talk a little bit about the difference or so I'll go there and you can have and I'm gonna quickly some talk about we've got this which is it's great but it doesn't mean that we can make computers really really really fast it doesn't mean they can open Microsoft Word in a blink of a second and it doesn't take forever to load anything that that is of what quantum computing does it it's quite a common misconception is that quantum computers that means you can do stuff really fast and it means that we can do some things really fast because we're doing them in

a different way but it doesn't mean that we're just gonna make computers better in general and it's because we've got these different kinds of algorithms Rob well doing up these gates in different ways and those of you are familiar with like complexity theory and stuff like that you have I don't know a mathematical problem that is really easy to solve with classical thinking and classical computers and you have ones that are really hard to solve with classical computers and we don't know how to do them so factoring prime numbers is a common one you might have heard of before but these new algorithms can potentially make those work really really well we can calculate prime that

there we can calculate prime factors really really quite quickly using a special kind of algorithms called Shor's algorithm and and it kind of some of them will overlap with classical ways of doing things at a minute and some of them are just completely new and they're a whole load of algorithms out there that haven't been invented yet but quite early stages here and there's maybe about Shore's Grover's the ones that can think of the term ahead and a couple more but their whole load of problems out there waiting to be solved that people just haven't looked at yet yes you still haven't worked out whether some problems are hard lovers repeat with enemies mm-hmm has it been proven

the quantum computers can solve some problems faster than classical finishes is that so something like shows other than and does solve their prime factor problem quicker in terms of the number of steps that it takes it has some classical steps in there as well and then there's one quantum bit that basically just makes it and all it you solve it in fewer steps and that has been proven whether or not it takes less time it depends on how good whether they're like thanks because if if like if you could show that computers couldn't do that because I think it's Apollo's something level if you shown that that was significantly possible to like classical finishing sort of solving it's a very good pumpkin

yes essentially so I'm I'm I'm not a massive expert in complexity of things like that by very certain that like the quantum algorithms have been proven to solve in in far fewer steps than than like the classical solutions to them as well which is why people really care about this sort of stuff because if we if we can make it work then we solve a whole lot of issues that people have been worried about for a while like how can we actually solve these problems and it's only like certain problems that we've kind of thought about the answers to so um protein folding is a really interesting one that is really really really hard we

can only model like a couple of proteins like small ones but you know there's potentially quantum algorithms out there that automatically better and for that you don't think you have something to add okay this algorithm actually i'm i pre prime number that you passed they did yes it does and in fact that's very much kind of the next bit where I'm where I'm gonna go with this there is someone else that I hand up over there so can you clarify what your complexity class q is supposed to be didn't call anomalies probable it with with quantum war yeah and in this instance I'm just using it to vaguely mean like solvable within and like the

time of the universe so it really depends on the problem yeah yes so complexity and in the time that is you know not necessarily polynomial or not but just better than the current version better than like polynomial or whatever is that you regard a minute and okay so talking a little bit about how good quantum computing is or isn't and I'm kind of gonna defer this to an later but some of you may have read to the press race from Google or rather the leaks about Google and that potential quantum computer that they've got and okay so exactly what you say so Shor's algorithm this is a way that we can factorize prime numbers and

potentially quite quickly and that does have applications for you know people in this room you'd probably deal with crypto at various points in various levels depending on what it is that you do at the minute you know we've got a a yes three five six we've got two three Hellman key exchange things like that these kind of rely on concepts such as the fact that prime numbers of fact the big numbers are really really really hard to factorize into their individual prime factors that has a really tough problem that computers can't solve within you know before the he doesn't the universe where as whether quantum computer we can do it fairly quickly potentially depending on how good your

computers so this is where it starts kind of getting interesting a little bit and just in a sort of a bit of a naughty way here so the idea of how like cryptography kind of works is you have your secret you put a lock on it you pass over to your other person they put a lock on it and you bring it back to yourself you take off the lock that you've got and you give it back to Bob and they can take that lock off because they've got a lock look the key for their padlocks and that way you've never left this secret unlocked at any point and we've managed to you know by this over and and you can describe

the same sort of concept with kind of colors so the idea that you've got a common thing and then so that could be like your public key and then you add your sort of private bit to it and you can shake it up mix it together and and it's got to be really hard to unmix that and then you can stop them over and you can add your own colors to it again and then you both end up with something that looks the same but at no point have you revealed what your secret color is you just know that it looks the same at the end and and the thing about that is the lock in between or that bit where

separating the colors is really really difficult at the minute that's based on this thing of prime factors being really really hard so that makes the separation difficulty is you know it is really really difficult to take this really big number and know what the prime number I started with was and the prime number that the other person started with was and go from that and so obviously that's that's great as it is but if you've got a quantum computer and it works really really well then you can break that and you can figure out what people's private keys are and and that sort of thing and if you can really care about this sort of thing I don't know maybe you're a

bank or your own nation-state or you're just really ahead of the curve and you're really nervous about when the quantum revolution is coming and you've got a couple of different options you can think of you can think about the a quantum solution so this is a problem that's been brought about by by quantum existing and and we can we can think of a quantum solution to this so we can again use the weirdness of physics to try and and come up with an answer can we use the weirdness of physics to do cryptography so there's another weird concept it's called entanglement where basically if I take two quantum particles and I mix them together in a certain way I can entangle

them so that when you do something to one of them the same thing happens that the particle and I could make some entangled particles I could keep one I could give my other one to an and you know I could rotate that particle around she would see hers rotating and through that we could transmit a message and there is nothing but I can't put anything in the middle I can't do anything else to it I you know someone might be able to measure what's going on but then the whole thing would deteriorate and we would know so it wouldn't work but but in that sense like that is one way that we could use quantum physics to find a solution to

this or we could think about doing it in other ways I don't know can we find an inherent property of mathematics or can we just use classical ways of thinking or out-of-the-box ways of thinking to try and defeat this problem rather than coming up with an ever more complicated solution to to what is already quite a confusing problem and and a lot of people really care about doing it the complicated way I have they they like to sort of make things as hard as they possibly can for themselves and for everybody else and I don't know why they do this but um that they do it and it's because it's interesting stuff right like quantum physics is really

fascinating and it would be really cool if you could come up with an answer that used all this really fascinating stuff but it doesn't necessarily have to be all we really want is something that can be transportable we can use it in a lot of different situations ideally we could just integrate it with computers as they are now rather than having to put in some crazy quantum chip box that has particles and lasers and stuff flying around in it because that's gonna look very weird and be really hard to integrate and and because these algorithms are always evolving we don't really want it to be relying on the idea of stuff that is really hard to solve

for a computer because we have no idea what someone's going to wake up with tomorrow and their heart but this algorithm that will solve this computationally difficult problem so we want to try and and find something else that is you know separate from this kind of idea as our new way of doing crypto and maybe the answer to that is using randomness somehow so randomness is everywhere it's readily available it's you know all over the place it's completely inherent it's built into the universe as it is now it's inherent even within quantum physics itself things that you know will not necessarily predictably appear where you want them to be that will be random and and it

doesn't rely on any complex problem-solving or anything like that it doesn't rely on it is really hard to factorize numbers or or anything like that it's really hard to sort this or it's really hard to separate these colors it doesn't rely on any of that and so how could that work so ok and I've got a coin and I'm gonna give mark another coin you go cut it fat okay right I'm gonna flip my coin several times and I'm gonna get a series of heads and tails at the top and I'm marks gonna do the same thing with this and he's gonna get this the second row of heads and tails here just gonna move my

mouse all the way there we go and every time we match we could use the times when we match but I get a head and markets a head or when I get a tail and Mark gets a tail we can take the bits where they match and use that to create our key potentially no one else if someone else said I know and is wanting to eavesdrop on our conversation she's got another coin she's going to match in different places and so if we can find a source of randomness that we can use in this kind of way maybe we can communicate we can come up with some system to describe when we've got things that are the same

and when we've got things that are different this is where like the only little bit of maths kind of comes in to it and if you want to you can show your ears down and I'll give you a wave and let you know when the masse bet is finished but and we can use an algorithm so the algorithm I'm describing here is called advantage distillation and we can use that to communicate about when our heads are the same or we both got heads and we both got tails and the way we can do this is I I'm gonna do it four times and I'm Mark's gonna do his four times and then if I get the same thing four

times in a row and Mark gets the same thing four times in a row in a row there's a reasonable chance or some of the time they're gonna be the same so I'll have got four heads in a row and Mark will have got four heads in a row I'll have got four tails in a row mark littell got four tails in a row and we can write that down and then we can do some kind of error correction for the times when when one of us was wrong or when we all the times that that was different all we need is for that to happen not an even number of times not an equal number of times so I get

matches out of five times I get much as four times and the same goes for mark that's all we kind of need rather than it being we do it four times and twice we're writing twice we're wrong we just need that slight bit of correlation in there and it's this ever so slightly correlated that we need to be able to take advantage of it and in order to do that and if we had an eavesdropper that's trying to do the same sort of thing what we need is for her to not be right at the same times that we're right as well and you kind of think you might think are yes but but something that's

random it means there's a 50/50 percent chance of it happening if we've got a coin and it's gonna land on heads 50 percent of the time it tails 50 percent at the time the chances are that we're not going to be right more than half of the times except randomness is random - so that means that it's never gonna be 50/50 at any point it's always going to be 48 and 52 or 49 and 51 or 40 and 60 or something like that it's always going to be ever so slightly skewed the Amanda that it's skewed is going to be random but it will be skewed one way or the other and and and there we go we've won

we've we've managed to get that slight weird bit of a correlation and that's that's all that we kind of need and yeah there are times when and when Mark will be wrong you know like where I live I'll be expecting something and I'll expect it I'll have got a much and Marvel have got a much but I'll have got heads and Markel have got tails and this is where we do our error correction protocols and maybe it's a turbo coding error correction or something like that and we can apply that and kind of make it so that we have something that's the majority of the time that's right or every time he's wrong we could just get

rid of it ignore that bit and only use the bits where we know what right um yeah cool yeah life is weird that's why randomness is random - and we're quite lucky because randomness is everywhere and this correlation that is sort of everywhere as well and short noise is something that you cannot get away with short noise exists in everything and uhm short noise if anyone is an aware kind of looks a little bit like this so if you are expecting to measure something at a certain voltage some of it will be slightly out to one side and some of it will be slightly out to the other side it's because quantum physics actually why this actually happens and we can

kind of say every time we measure it slightly to the left we're going to call that a zero and every time you measure it slightly to the right we're going to call that a one or a head or any tails a head and a tail and we can use that to start building up our secret key between each other as well yeah so this is just essentially every time it lands to the left to recall a head every time it lands right we can call it a tail and we can go through the same process of finding our matches as well and I spent a really long time during my PhD figuring this out proving that we could

do it and whether or not it was true I did a lot of simulations I did some experiments as well some of them went terribly wrong but some of them did actually mean that we could create this key and and share this key between between two people and and that's all well and good I'm at the minute though we don't know necessarily if there's another weird way of thinking about it that means that this is all wrong and we've we've all been Ryan I just wasted five years of my life thinking about this but at the minute I know so far so good and this is potentially like a viable solution to this and there are

other things as well to think about as well I just put tapping Faraday cages there but that's kind of you know how secure is as actually receiving this can anybody else get in at that point at the point where I'm receiving it do I need to be in a completely secure Faraday cage room for this to work and kind of thing the question of whether this is relevant now and very shortly is going to talk a little bit about where quantum computers are actually at and how good they are but it's entirely possible that we can use this algorithm in classical computers as potentially on IOT devices and things like that as well we're actually crypto is really really hard to

do and that's not it from me if you've got any questions you can potentially ask them now or I don't know if you want to pass straight and and I'll have a bit of a chat dirty mind yet different please um how do you make the results against obviously how do you do this mixing so decisions so basically he started take a number I would XOR it with my same same number four times like a zero four times and I would XOR that with my random stream of heads and tails and I'd wait for the bits in my series that were the same so that would mean in my heads and tails I had four heads in a row or four tails in

a row as well as like having the bit that I knew about the beginning and then they're gone I'm also how to authenticate it on the other side because you've got some decrypted fellow you know be always a demand yeah so this is a solution that we don't well this is a problem you don't actually have a solution to yet is there's got to be some pre authentication that happens you know I I don't know what the best way to do that education would be like you know how do I know that Marcus mark I don't really know usually because he's lied to Scouts but like yeah that is absolutely like a completely valid question because that's

the bet we don't actually know yeah quantum resistant yeah and so I'm I'll be completely honest have not had a massive like routinization of the candidates that I've gotten to missed I think we were slightly too late to put this - NIST and we thought about it and then the deadline was very like and you know again this cut personally for me it kind of comes back to this question of are you trying to find a really complex solution to something that doesn't actually need to have necessarily a complex solution - it's got to be something that is distributable as far as I'm concerned it's our only gonna stick if it's easy to spread around -

goodness for me personally that's what my main criteria is it may well be that there are other solutions in there in there candidates for nest that are you know far better than the one that we've been coming up with and and that's absolutely fantastic there are and but off the top of my head I add not like been through their submissions really because we mentioned kinda like this as a communication channel yeah required to store the samples particles yeah is there any reasonable it is is it reasonable to do to to do to believe that that will ever be applicable to household devices to do anything less than mass scale industrial application did applications yeah within

this millennium so that's that is a really really interesting question no one really knows at the minute one miles away to accompany knows I there's so many hurdles we've got to come over one of them is getting entanglement to last longer than like a few minutes so at the minute that's currently like a an issue certainly on the scale that we're talking about we can't get it to last too long before it starts to degrade and and you know the other thing is that to be able to create these things at the minute you know you're potentially using massive laser chambers laser tables and ion chambers and and things like that so currently no but then if you think about

back in like the 50s a computer took the size of a room and that's kind of what this sort of thing does of a minute so potentially at some point someone will come up with some idea of yeah actually no I figured out how we can do this on a chip and which would be you know absolutely crazy if it did but we currently don't have that and yeah I don't see any particular candidates on the horizon but then I don't know all of this do we know exactly why things will entangle and howling over the distance how how is it and they moved what happened how they happened how do they changed at the same time and

especially when we've got things like Large Hadron Collider explosion you know pastas are so different we're now looking to be able is there a possibility that you know that communication is I sell all its currently but you know in the future actually if at all it's just a white half of this thing that yeah and it comes back to that thing of we can describe quantum physics really well we have really strong mathematical equations that are very reliable in producing the right side answers but we don't have a clue why it does the thing and you know potentially there is there have been various people that have postulated yes there is a some sort of

carrier particle that goes goes between but so far any mathematical reason why that would exist and any physical evidence of that existing housing hasn't come to fruition well uses this or such a product would be general relativity and it's yeah the existence of such a thing is currently not really compatible with any of the mathematical models that we've got to describe quantum physics at the minute and we don't really know why it does that it might not be it might not be a particle egg I don't know maybe it's something to do with daf motto or whatever like all of these things are physics that we actually have no idea about when a completely unexplored but

we have absolutely no idea it could be strings vibrating and like string theory or whatever if depending on whatever it is you subscribe to in your sort of metaphysics and philosophy kind of levels but we had tails thing where the quantum oh that is because it's not a quantum solution so it's a classical answer using something that's inherited inherent in the properties of the world and the properties of mathematics and rather than it actually being specifically content and that's kind of why this is actual we're going away from like the idea of something really complex in quantum and entanglement and things like that to the idea maybe we can just use mass as it is

maybe I don't know go to completely got it wrong but yeah like yeah the quantum kind of comes into it a little bit if you're trying to think how can we extract run randomness so shot noise for example that is is quantum and we can use that to give us the randomness but the concept itself isn't content at all which is why CNN Charlie's recently from satellites conference room familiar with that Kunis is any relation to the and do you know what my third as the

[Music]

[Music] [Music]

surely shining a laser dump light has exactly the same on the inception something I always listen

it's feasible to detect have we said the same rules you have to take that we said the same so basically a single was the first to sort of Leslie you mentioned using entitled as a communications is it just that I've heard this sort of inner psyche like context the loitering and they kind of like faster-than-light communication does that informations don't obey the speed of light so in the entanglement since it is instantaneous possible speed of light the speed of light kind of implies that there is something that's communicating between the two of those and reality it's just an inherent thing like they just know what is happening so it's actually been losing the information if you open a

Newman doesn't violate special values we look if information has to trouble and information can travel faster speed so I have communicated to your customers what are you looking already been made as well eight okay so we haven't actually lost Lloyd yes this you know delivered if you have if you if you have to untangle 10 particles and it's your oscillating them at it at a fixed frequency and then you start moving one of them at relativistic speeds yeah so how does that clock sync up with the time dilation this is all instantaneous communication I think at this point was starting to move a little bit away from PI who am I talking about in terms of

quantitative of faith for the benefit of everyone else in the room so um do you want to do it yeah I think we do that do it do you shall we do five minutes of people can quickly go to Lou and come back or do you want to go straight into it okay all right okay in that case I mean if you wanted to talk to Reza more detail at the end you just just come on I mean tap when we're done the wonderful days where everything just stops working because of course it means but there's so many tech people around that there's no reason to be sorry it's just a number one thing that

happens at a conference where those sheep goes wrong like regardless of how it prepares you wrong you've been doing to talk with my HDMI presentation my pen is not that it sometimes takes like 30 seconds for it's going for about 4 boxes

yes that's expendable emeritus okay so what do I need to do so this is the first time I'm using this

[Music]

[Music] yes should I yeah so as you can see everywhere my name is da cunha because my last name is a difficult I currently work at the experimental column information room at the University of Leeds this is my twitter handle but then again my twitter these days is mainly between you know nice pictures of cats or something so that's not basically you know show augment my following and then boost my ego but otherwise you know it might not be worth bothering anyway what I am going to talk to you about today um hello computers we are here because of their Google announcement well I'm certainly here because of that good announcement so let's run through some pull on computers

different chains of criteria because those are important for quantum computing and you'll see why in a minute chorim and returns you've seen at least two or you've heard of two so let's go through that why not then I'm going to explain to you why encryption is going to die at which points you should please let me run away at first I would make sure that the second head starts to make it in the sense and then maybe for score to encrypt all which not really both show by this point somebody would have to make yourself feel better right I credit I'm references I usually forget or I don't get to my last slide or something so I just saw that put it up

now um you'd see a lot of very nice pictures they are skin grabs there you go and there is one who is not the screen right the head fit of Wikipedia and it's on Creative Commons or whatever license with here's the weight fight it's listen and shrine that is so known as one of the Bible's that we have for anybody working in the fields it is also at this point almost painfully old Nelson entering was written in 2001 if memory serves just so you know the infrastructure behind Google and IBM and we didn't is not in it so have fun however it does have a some it does go through the physics and it does tree

goes through algorithms pretty well so it's a good text for at least those two chapters all the references so I do there on the signing prompt from shooting course online couple years ago just to have an idea of what's being said in the field etc if you find it running it might be a good point it's not very long and it's free so might be wanting to detach so here's why we are here Google announcement consume supremacy and everything right so easier they have quantum computer they're fifty Oh checked 54 qubit chip on computer so let's play game the game is known as find the chip right okay now sorry this thing is very confusing

I haven't actually gone anywhere is yond the picture of you see so i come on estimate by of the picture but i think and it's this really bad picture but this is why you should go to the new york times i've had a much better resolution on the chip is yeah it's a bit of not a great size square but it's square about this big plugged in here so what's all the rest this bit here that is literally the outside of the fridge the Google architecture I like p.m. I think Microsoft is when their means of superman' superconductor technology right a superconductor only works very very very cool temperatures we're talking within a low degree of absolute

zero or as close team absolutely we can it's very very difficult to get to those savages which is why you need we know what we call a crier stat and this year that is the outside of the higher self it's painted red because show well it might as well have sweaty colors um it is very heavy hence why these things we think um you have all sorts of vacuum pumps over there because it starts by creating a very very high bite you you have all sorts of temperature controls here okay so we have found the chip and we know what's on your eyes the rest of it here these are cables for communications with your qubits ah sorry

it's like the last time I went driving [Music] no one heard that anyway so you need to you know set your qubit you need to be able to talk to you keep it you need the qubit chief do you think etc do you this is what all the cables here now yeah yeah here it's pretty much the same thing except that we're not seeing any of the outside so there's no off none of the red bit here but you can see at the top the whole column goes um and then you have a link to chips somewhere right here and here again you have all those cables communication yeah now again I don't know if whether they

were the first but they were amongst the first to UM allow you to code on their system this so this is you go through IBM wanna mix the IDM form experience and you can actually close on those five cubits okay so you can learn how to use all those gates that we saw earlier we jelly is the same similar thing so literally has about thirty three bits of cotton bits arranged this way and it also has the fridge which are not showing because that you've have it on the website and it's got it has also got apple pie awesome languages I could and everything to allow you to just codes the quantum computer and simulate stuff on the

computer okay now d-wave is a little bit of a special case because he um arrangement of the bit is different and TV was one of the first if not the first are actually claimed to have built a portal computer um it's fantastic and it was fantastic at the time with a little caveat which is all you need to be able to write the problem in a way that digi-wave computer will understand d-wave solve problems using the annealing technique that doesn't solve that doesn't suit all computing problems out there because the main problem with a wave is I need to be able to write you go you need to be able to write your problem in the way that the computer

will understand essentially I think at the moment it's got two thousand units so it claims to have two thousand units it has also got its own simulation environment so you can log in what you should register first anyway the register login input your problem get your solution ah recently if it has also um hello wait for the last five thousand cubits so this is the new next-gen chip it's not just next-gen because it's five thousand cubits they've also better to the way Oh increase the number of connectivity and connections that they have between that it's so so far the previous two thousand cubits not Pegasus one has qubit connected to six of the qubits this time they have each you be

permitted to seek 15 of the chivitz so they've got that much more processing power so they can also deal with much larger problems yeah you've mentioned the arrangement of qubits are different for different computers how does that like why does that affect I connection between qubits means that you have some to dekes operations I think we'll see a to gate to gate operation in a few minutes so how many operation that Freya showed you is a one qubit and it's one qubit because it only affects one bit it flips that bit that's all it does but you have some operations that work on two qubits where you have one qubit that controls the state of the other you can also have

some gates that work of three cubits etc but if you don't have connections in between those qubits as an actual physical connection then you can't apply those operations so for instance on the IDM we do have some cubits where you're not allowed to put gates in simply because infrastructure those bits with actual cubit are not physically connected that means they need to rearrange the order if you doing some operations in which is why it's very nice to have a lot more cubits that are interconnected and connected with each other because do you have more freedom when it comes down to you coach Senator Mike you remember I told you that Google tree Getty and IBM or

superconducting qubits these guys thank you our building the computer based of I am trapping so I'm trapping is very good for bidding qubits because it keeps the qubit exactly where you want them okay you can do a lot of really really precise control with fighting the problem of ion trapping is that it also requires lasers a requires a big empty vacuum chamber like the infrastructure itself is really non-trivial I mean you think Google if bad honestly I don't know how these guys are doing it but they are doing it and they are doing it particularly well and they've got single qubit gates on 79 qubits they can pair up to 11 qubits together so they do you

have like they've also got like really really good fidelity which means that they can distinguish stage really when those guys are I think we're going to need to watch out for them as well I [Music] won't go over the whole cubed thing because Freya went over the qubit and explain to you what the qubits is just again to repeat clásico bit is a 0 of 1 a quantum bits has 0 the 1 and any combination of these okay if you want to see it on the sphere as we did earlier the qubit the zero is a normal Northern Hemisphere the one is the southern hemisphere and then you're like everything else in between and because

your nads all of that extra space you have a whole mole I'm sorry the word is so you haven't one hole to play with basically and to encode the information in etc so obviously the more qubits you have the better but you also want to make sure that they stay cubits long enough okay there's one very big problem in physics called decoherence is when stuff stop being quantum and becomes classical so decoherence means that this goes away all you're left with is these much okay which brings me to the defense the defense sensor criteria I'm sorry I must have left my addiction at home when I got up this morning these are the criteria that we need to make sure that

we have a quantum computer and I have quantum computer that works so obviously scalability with very practice qubits remember I told you about iron Q and iron trapping an image of resources it takes for just one well that was why superconducting qubits we're chosen at some point is because it was the fast the fastest way to escape the infrastructure and to keep qubits well characterized now collateralized means that you can actually identify each qubit which you need to do now the ability to initialize states using or feed also state I mean that there is a point in a competition at which you need to make sure that all your kids are at state zero right

oh you know zero zero one or something that is basically what this means if you can't control your cubit so make sure that they all get to zero at the same time or when you want them it's a problem long the coherence time you need to make sure you cube it remain cubit - otherwise that defeats the whole purpose of having a golden conchita it also influences memory okay she wants you keep stuff in memory you need to make sure that those two beads don't suddenly disappear because when they do the information disappears with universal quantum gates I think this one is pretty self-explanatory yeah IQ be specific measurement capability that means that you need to

know with what each qubit is so this right here that's for the quantum computer itself right but the console computer needs somebody to input and output stop at one so these three that concern communication in an H okay communication in a night so that G the stuff that you put in gets in there and the stuff that gets in there you can take X as well and then communication within it okay those qubits have to be able to talk to each other they are not in a vacuum well yeah yeah but you get the point right so when I quantum algorithms hey no that's all right he knows Grover Wow you guys know some very very smart

people I need your dress books I said smart anyway so what why do we have quantum algorithms well you have a quantum infrastructure you have a whole bunch of transformations you really want to just limit it to classical algorithms you want to use those X you want to use the qubits to their full potential right you want to use all of what quantum physics has to offer and this is where the whole people research a quantum algorithm solid ok we're not just inventing stuff for the sake of any physicist inherently there is an amazing laziness about to a steady couch tonight so especially theoretical physicists this is what mathematical approximation is all about time yeah go on disfrute anyway and so the

two that are most interesting really useful when it comes to practical cryptography all the filler paper came out is the quantum fluid any algorithm based on the quantum Fourier transform one of which is Shaw and the condom so tripodo the best known example of it is over ok as something you can also have a quick look at this some Steven Jordan from Microsoft has started writing a comprehensive catalog of quantum algorithms so if you want to have it I invite you to go have a look it is also depressing because he tells you that you know there are some algorithms to having a bridge diffie-hellman I did say I was gonna scare you right so

quantum Fourier transform why do we bother ever with before you can load it into go it's really painful a take it takes a hop follow me turns into something that's resolved oh that's what we like before you transform except for the Fourier transform the history toy frontal passage it also is very slow and very difficult to put into practice so you have also to failure of those a class for computing and our event making this faster this is why the best known one is the fast Fourier transform and everybody uses the pathway or because it was a genius piece of work one computation takes that parlament it pish-posh I can do this so much quicker you see why we

have a destroyed finishing issue a compensating I mean fun but yeah this is it's brilliant it is an instead of spending my whole career on one problem I can just eat the computer and blue solution so again both pointed here in person the one that really is a problem is your defining algorithm which is show we'd like to show it sleep on all the fighting and all the finding works on please estimation so he remember you blocks here you have your the Bloch sphere is the sphere that you keep it is on right so you have your qubit which rests on the surface but you want to know the angle that it's making basically and that's your face it's big

rocks negative so we want to know the vector that describes the state because that's where our qubit is which means that we need to know the angle lines native right and this is known as the phase estimation problem the angle we pulled out the page physicist so we couldn't be bothered making it safe so single gate okay this one is known as the harder mode it does something it always has either it bits interference and because if it's interference it also means that you can take one two bits on that which force it to do more things that if you just leave it on its own this table here means that there's actually quite a lot of these okay

I just because they're all the same and they all do the same and they all interact the same with this one I'm just running a sync single line now let's say I want to know what this one is where this one sits on the sphere okay I'm going to take a whole bunch of she Benson Oh same to be like expanded something and then I'm going to take a phase estimation protocol or phase estimation gate okay and then that it's going to act as a control qubit and post these ones it to one particular states that I can then measure okay I have these stands for inverse Fourier transform at the moment it's a black box

because is fairly complicated to know what's it just like yeah I don't want to know what's in it right now I mean I could on the you know if there isn't an actual inverse Fourier transform black box on the IDIQ that would be too easy so I'm gonna have like if I want to be in it but I've got that about 25 minutes left and over the goal but the black box okay so inverse Fourier transform and then you measure so um when you're doing the Moriya transform you get a very good approximation of what your problem is meant to be because you get taken a half problem and you making it to something that's easy to solve

yes but you making it something that's it's still a bit of an approximation it's very very very close to what you want but it's not an exact solution okay so this here gives you a really good approximation of the angle then you looking for well why did that leave a really good generally good enough that if you especially if you try to do this several times you're going to get away with it that is his estimation okay now what about all the fighting all the fighting put some number theory at the start okay and then all it does is that it takes some of these a takes some of those puts this here x2 je

moet sorry here measures and then get something which is quite close to what you're looking for what you're looking for for the funding is all so you're getting something that is very close to s overall okay you have some number theory that you can then apply to actually get the value of our do you see why what computers can become a problem with planning it is to take something that is difficult to do and then turn it into something that actually is not really without able to do this we do usually market me here I'm always mixing those two it might be right there is also a discrete logarithm way of not a hundred percent sure sorry yes I still

run the algorithm only once might have a felony pursuit so factorization it takes it basically checks what the norm is in at the sounds the price all the finding there's some whole number theory at the end and that's it right that is short I present in previous slow but it's the reason that you use the separate given that if you measured it directly you changed the difference with that phase change the result yeah so you're basically using this is the one this is the one you want identified yeah so you need these functions yeah so you're using those as sort of infer what the value is without directly measuring it independent yeah this one interacts directly with the

others so the point when we talk about show as being a compliment rhythm there is a lot of classical into it it's just that this here he older finding that stupid such action difficult really long to do my computer can't adapt very quickly so my RSA yeah and by eg headsman which you have one of these that's called the discrete logarithm shows an rhythm which basically breaks if you help me keep curve diffie-hellman fantastic you guys scared well no you should indeed again yes and this is the biggest number that's so far been factorize using for okay we're not talking about 21 digit 21 as in 7 times 3 that's it so at the moment we safe yeah safe ish there that

is why it's also a good idea to keep a lookout for some of the algorithms that come up but I go upstairs in a minute and also some of the new architecture brother is so Grover Grover 20 start by flipping around a bit trying to find and then it applies the transformation to everything and it forms the answer this way the approver in zelf does not look for the answer it looks at where the answer can be okay so it's not look from a classical computing point he's not looking at the value is speaking at the pointer to that value however that's already quite nice to have especially if you know you're doing a sieve problem finding thing to find

out your factors and you find you know it can also stop at where the first one is going to be the problem I can see personally this is my own opinion here way over and while not finding it that super helpful at the moment but again that's my personal opinion is because you do need to build you need to have an idea for the addresses before you can flip it which kind of isn't so there are ways around this but at the moment I'm not finding that algorithm all that useful specifically for that again my own opinion um budget yeah that's a full on um so going back to this one see that's here we're looking at yep

it's true there is a point at which I'm confused so I'm going to be and stay off fast advancing I mean policy three or four years ago we had only like nineteen cubits and now we have huge for thank you is actually like doing even better and I and Q has got some memory as well which should scare all of us I mean who wants to remember what we did before but I'd moment quantum computer is a giant millipede it is investments in millions our I mean millions its millions it takes a whole team to run very specific set of skills hahaha required to run one of those which means that the likelihood of a

hacker the way I see a hacker which is probably not quite because those guys are like 95 I'll probably go to the office to work yeah um it takes a lot of money to buy one it takes a lot of money to house one it takes a lot of money to turn one on to run I mean looking at the electrical bill of Google should scare anybody seriously I mean we're not even talking an environmental impact here but we're just talking about actual costs of infrastructure the cause of the pumps the costs of the electronic here doesn't that make whoever's well a point of failure in cryptography wherever it gets applied yeah but you know that's called

social engineering and my understanding was that that is the point of failure for any cryptographic schemes regardless of whether it's on a classical I'm sorry but in any cryptography is like hell I mean seriously is like the communication between oh I understand your point exactly but I think that the moment to what we have to save ourselves is that the the people using the computers of a distance at the moment they have to register to go on IBM Chu to go on leave so there is some transitivity here and they're also students they all I think IBM prefers when you have an institution and email or when you're using your login in through an institution because you know

so there is some amount of control that being said yeah you could have there is some information on how to break RSA using heat wave so it was not by getting some to do it more that you're assuming that you know that the only way to access these computer is to register system if you can bypass the security to access the system yeah what was was was five points yeah but and again there's a traceability element because all the staff IBM tree I don't know about three four jetty because I haven't used them but the problems that you set up on IBM queue actually logged that could run some simulations but then again you know you do take time as well

of the whole system and its cloud-based so you have a whole sorts of extra data and later they talk that comes in with your simulation so you could set up some really hackneyed things but then again you know there are also supercomputers that exists and a lot of them and if you can find time are not to in Leeds and write some short factorizations right now during your legal time answer allow you to do it then I am Not sure that you could yeah but there's a point where the research has to go on as well it's very easy I think at the moment what saves us is a traceability aspect into it and that's right yeah people watching for

astronauts who how's the operation state axis like if you do have a opinion conspiracy theories online but you know the NSA is just tracked inches your factorization conducive comes to the tightness on their how much credibility do you think there is behind some of the sort of grounded in nation states might try it a couple of these things might happen how will pull the repeater they're liable to do this what do you think which is completely outside the realms of possibility with Road I don't think at the moment if they want to bend stem cells or Polikoff user the main problem is going to be the cost materials you have been you going to needs to have to pee

running it in it you need to have been so struck trenches units have the power to keep fish you need to information state actors will help yeah didn't mean the end I get I'm sure if someone who came up to the NSA is that G will apply my qualities that they can find the money not to you get paranoid about here so she mean that it hasn't already there's a point here where you know you can only deal with the enemy now you can see you can only tips the devil you know at this point we know where some of those become shooter song any because the guy running those computers have been very very false community saying

look at how shiny your machine right but I mean Microsoft is doing research on it I accused in research on it Google was doing research on youth who use how she jumps we don't care about it we knew there was something going on but we didn't know exactly how much and how many etc I mean d-wave have been on the market for like since 2007 2007 which might be wrong and all they're pulling out at five thousand cubits and that's already like two or three iterations in what there's a point at which you know there's only so much that we can do which is why we get here but I mean at the end of the day there's nothing

preventing us from somebody malicious login into idhu or login into Italy or anywhere answers to Sonny you know are I'm gonna stop running disclose or break heresy but um so what happens then well what happens is that we continue to do some research into quantum key distribution because the key is also another weaker part of the encryption so once if we can make sure that that key is unhackable then we've already step closer and post partum crypto is I think the very very underrated area of research people in Royal Holloway are very keen to tell you that there are at least two three codes that somehow came up in the 60s and that none of us know

about are now actually unhackable while computer or otherwise so probably like post quantum crypto it's going to be the one thing the one place where everybody should look especially because at the moment there's also something that all of that layer of encryption that can potentially be broken it would be good to either renew the encryption or add a layer of encryption onto it and what you can do is add a layer of both quantum crypto to it which is a minimum effort if you will because it doesn't actually force you to do anything through the encryption that's already in place it's just I don't know I think another another layer to it so that you

can be sure that even if Apollo computer or somebody has used a common computer address breaking RSA or diffie-hellman or AES or you know all of any of those schemes then you have a layer that that just can't break into it and course quantum is inherently classical it is pure maths as in pure maths but it's also very underrated I think somebody ought to give those guys in Royal Holloway of shootout so that he knows some who does actually like make it I'm side offering Holloway but they did I think submit stuff to NIST so there might be a few schemes that I currently being reviewed under the thinnest tqc competition okay right that's me

thanks for listening sorry I took the piss [Applause] yes the stationary cubit is one that is stable it adds the way you define stitchers in physics asides basically static equilibrium so it's a it's a something that isn't evolving and the flying qubit in that respect is something that is evolving so it's you need to be able to track has information on from the product from Frankie via tube its I'm seeing it as Aransas is wrong these kind of interactions but maybe that time yes you've actually the first quantum algorithms that it says it number of approaches and how much struck people in for example lattice probably as a solution the way I've heard it is a lattice

cryptography at the moment this condom secure what I mean is that it's secure against quantum attacks oh by the way one price we're actually only have mentioned is this home at the moment that's the biggest stretch it's the next big quantum algorithm I don't know we know we can defend against the one that we have which is what we're trying to defend against show quite basically increasing the length or something but some other genius could come up with something else entirely yeah with first we confusing we have obviously feature you see problems to work out roughly how difficult all the means is how hard it could possibly be do we have a similar thing for compton

commutativity easter halt this is also to limit the power these algorithms and honestly that is completely outside of my realm of expertise I couldn't tell you it's like I don't know enough complexity theory or anything to even begin giving you an answer a green screen please based on what just I get the feeling that infrastructure is what is for dinner slat beer and if the infrastructure problem is solved and this can definitely become a practical solution is that is the theory behind the expound enough life if the virtual problem is right I've been distracted at the moment the way I see it is that it

yeah the at the moment I think the technology is holding quantum computing the technology of quantum computing is holding its standard back to you again yeah the theory is now that being said that's also where post continents comes in and why I think those problems should take a bigger spot under the under the Sun basically because yeah there's only so long this acknowledge she is going to hold back basically sorry yeah yeah is that a promise just it's transcended all right um topological quantum computation right to my brain literally like is it takes my brain and puts it through at a lot the blender and it helps me so much that I'm supposed to make sense I've got

Guardians so honestly it's the same I just I don't know enough about topological give you an answer about this and like there are a lot of really smart people that will probably like be able to give you an answer and a simple answer I I'm not one of them sorry like again topology is smooth I have I understand what its Horace's but that's about as mention of the t-distribution and one of the points that is often made about that is you can't eavesdrop on it which is like because if they just don't get to the receiver didn't have to intercept it could attract redress care yeah I remember when I was lost from the main

plot that was theoretically it's great you said the single photon that would get single vote on a ruby great look you have to account the loss of the system at the time they were having a problem where they basically had to accept the early flight sent might ever get to the receiver so they've had to go like as long as we've got five percent of what we said and we think we've got it which leaves room or an eavesdropper to take like to get into that system because you know they could take any of the other 95% of photons that were sent and you wouldn't know that they were there this is that something they're still moving

towards are they still looking at getting down to the single throw taught in the or is that something we're looking at or is it you know we've given up I know I'm looking at a penny to improve them out there are always looking at improving their system and including the fidelity and including all the steps in between especially when you come through repeaters on having those and all of those the quantum part is also a lot of research being made being done in the classical bits all the reconciliation and error correction privacy amplification is a big one as well we as long as you have you have two legal parties that can talk to each

other even if anybody can hear but no one can change the information being passed around person between if they start with a launch in or they say they can't fit it here certainly so there is research being done goes into the production and passage of the signal for transmission and there's the whole error correction and reconciliation and privacy amplification on top of that so Kroenke distribution really has two bits in it and you can't do without the classical building up yeah surely I don't understand information contents of qubits at all - you could just tag it is it this achieving of the wall this is keeping episodes and you say which one with the seem to that it's

always essentially you wouldn't be able to sleep a few and so how do you get me sure it's not um now the problem is if so the several aspects to this the first thing is that it is very very difficult to send eight full-time as it is technologically it is humongously difficult to st. just one photon but if you could send photon one hold on to you filter three etc you have to do with what gets lost in the fiber because an optics fiber has caught a lot of laughs in it especially if it's a single photon traveling so somebody could expect one to use three photons and then two is missing for instance so they expect one two and

three because the detector is supposed to click at top times you have times where the detector clicks when there's nothing there because detectors are not necessarily perfect either so that's also why it's very difficult to just send one photon then a lot one then another one simply because it's just light is you know very very difficult to you write it one of your packet okay this is the network being one of your packets gets low December four hundred different packets of differ case and you get one of them the other side yam just agreed to use that while those were plastic and telling you who to say I forgot I got you a lecture number 23 that she's what since like

that as I can yes but if something if somehow these refer has got Amina some information about packet number 23 and enough information about packet numbers 23 that if you just say let's use the hundred fits that's in packet number 23 these are folks who have 15 bit set of those do you want the interpreter have any bits as opposed the form is the Yi that's why you need the classical information the classical communication after that so if we quote you sender Kosovo is said you'd send she'd said like a hundred and a hundred of those packets and then you include which packets you agree on some form of secret foods to take those targets without

having to disclose which packets a so you talk about the cube it's all sort of you see you send a cube yeah you the plumbers they're getting lost on transit you can't observe zip wet and it being lost yes so surely if you just go for if you send lots of lots of different ones of go through the world but then there's got to the other side intact and you know hasn't been all so let me just read examples oh yeah that's usually the weight have events are happening in red if elder is trying since we have anything else is a 0 and this one being the one so you have sequences of 0 and

then everything good thanks missing how a never wait doesn't really matter it just takes slightly longer flu thing okay the problem is if your error rate is too large then there are other bits of classful constitution it's not worth doing because you don't decrease the amount of errors there's a point where if you have let's say if you have 25% errors no matter how many times you run your error correction you're going to end up with 20 percenters i'm counting five percent that disappear and by the way I am completely inventing those over the point is that you have you can start with a certain amount of errors and that's fine because this noise there's

pollution there's lots there is God knows what backyards and it's it's right you wanna get free take people about it so lots of things that happen and accept that yeah if you if the number of errors is too large then it's no more to time simply because it won't reduce no matter how much effort you put into it and if you can't reduce the errors no matter how much effort you put in I don't know tighten the screw and straight again yes streets on in the road or something obviously even yourself miles long process in silico my interesting research I most of me I'm not familiar with that so the answer to question is no I couldn't comment simply because I

haven't I just haven't kept it close enough to it yeah yes believe that more or less struck out roasted so big that is so bit quiet as the research around it sorry exhales mate of mine who happens to be a researcher there and it's we have I don't know of any photography work they work they're more in how they embellish the chips and then well as well as far as I can tell it's three professors with three different methods and and they're trying to collaborate because they're getting some high precision i I've heard that they can they can have basically introducing single atom defects into the silicon crystal to make well atomic level well transistor for whatever you

call it in yeah yeah yeah and it's one of their so again Connie superconductors and I'm trapping out two of the ways that people have created qubits but they are not the only ways people have created qubits and there is research being done a missing items in diamonds for instance and quantum dots and you can get some really high precision qubits if that is the particular research my worry about it is but it is that it's not just the shape itself that you have to consider looking at the different chancel you need to be able to make a machine like you need to have the rest of you have the rest of it and sometimes the rest of it takes a lot of

space and sometimes it you that's basically how the scalability aspect of it is going to work out so it can be that there is a lot of research being done in the end designs and a sentence in diamond and all quantum dots in there etc that might be relating to the ones in New South Wales again I'm not entirely sure but at the moment the two architectures that have been the most successful certainly is the superconducting one because I think it's the one where they found it easier to just cram as many qubits as they can thank very small amount of space but yeah research is still being done because it is still expensive to read

it's like the extra first the occasion decoherence ya decoherence is uh it's one of the major pitfalls if you want to that yeah entanglement dolly it's very quickly but qubits tend to dicker here very quickly as well if there are left in the wrong environment to put it bluntly that's why superconducting has to it only happens when it's very very cold that's why ion trapping works you know I'm a very specific set of conditions because you can't trap iron out from temperature yeah you mentioned that a lot of quantum property is reducing the step count rather than necessarily reducing how long each step takes it isn't it isn't if you go ahead your question so over those guys for what

what sort of time scale are we talking are talking about for the actual measurements and the individual steps in a yeah computer you're using honestly or a magnitude is all within fireballs of magnitude inspired I probably depends specifically up there you know do you include all the time takes to sell or do you likely just private exten do the operations is to do the operations yes kinda back time all but if you've got you know ten different lasers that you want to set up as I've asked eight weeks yeah I mean the quantum I mean just to get to ultra-high thank you yeah it takes it takes a while to get to ultra-high vacuum just that and then you

have to you know set up the stuff properly there's all bad quantum quote the there's the classical coding that you have to do before to make sure that you have all of your qubits initialized the proper way so um I think the quantum bits itself the advantage is that it goes much faster and you can drastically reduce the amount of operations that you have to do so you could be talking about I don't know the day of computation if it comes to the epochal I'm hoping for a day of other honest airplane but each other a tional step is it is it but we still talking like nanosecond picosecond other thing well that's why you need to

you have to go quite fast to be decoherence and that's why you need long decoherence slide decoherence can happen in the nano scale almost none a second decoherence an anti-human sudden death to events that happen extremely quickly so if you can't control this and make hits she beat die live four seconds two minutes and as the way i see it is quite simply this is how long you have to run your code if you could can't run hasta it is going to outlive your qubits and if you don't leave your qubit you don't have the answer at the end so you need a code that basically can run within the lifetime of your qubits the actual like

time it takes to do the transformations and read the different gates quantum system isn't necessarily that number all but in these stories it's long enough you to be able to measure it there are some of these things that probabilistic measurements so you need to measure for quite a long time to see when like it's most probably going to fall rather than it just being like a statistic going on it's like okay we need a key

so yes it's more like the amount of time it would take all right I know how about a online sneezes natural faculty doing themselves well yeah I mean getting measurement of the laser it's if you can't shine the laser at a photo targeting you know retinoids and it takes hours just to get the laser to point in one direction like if I haven't done it it is surprising the amount of time that it takes to make sure that your laser is going to point exactly where you want it and you're thinking it's like it's a dog okay no moving I am very familiar with you the annoyance places but that's why also why people have moved towards superconducting but

then you do have other issues with super conductance and you could have you know if you have a dud and I mean that Sycamore paper in nature was published with one day she bit in it it literally is just dead cubic in it all they admitted there's one of the activity that didn't work for something which is you know that's that's whatever it's fine the wind it you know it's like you have some micro processors or Michael that come in with effects in it and sometimes you don't exact control exactly what comes out which is why lasers are nice because laser can be readjusted it outside cetera but I mean at the human it's it's all about checks

and balances at this point you have to balance the cost of manufacturing the cost of stuff and then if you have to I don't know make a new chip every six months because some amount of computation just dead and still the old one then that's animate that's basically how long have to run your computation and it's longer impossible and so I'm not seeing the hacking I'm not seeing hackers using quantum computers anytime soon that being said there is nothing preventing anybody from using one of those IBM QM actually does some very scoot on leads to simulate something that is going to break sure eventually so I think people should keep an eye out for that but

unfortunately it works with everything it's the same with supercomputers if you have anybody using supercomputer in the malicious way that finds a way to break some some encryption or something you know hacking on it makes managers to put it on to the radar of something then this very disorder you can do anybody's work at the end of the day yeah it's a toy it's like Glenn was saying this morning with all the codes that gets you know it's a whole idea about the proof of concept and finding the bugs and publishing the bugs and publishing the you know all of that it cuts both ways at this point yeah are the computational properties of the kid of the qubits the

same across yeah a qubit should react the same way regardless of the infrastructure dating d-wave is nicely different not necessary because the way the qubits are connected because the way I talk to qubits talk to each other with a qubit is supposed to be in front of it the same way as a classical bit is the same bit in you know a Mac or Windows all dead or anything I mean I guess my question is like you say there's a difference in pitch but was a d-wave how they are connected so so like are the different different quantum computers who they do they exhibit the same properties in the sort of asymptotic bounds on not

unpacking yet and so they said to the different Renzo when he talks about Universal Kong gates is that the content gates have to be a harder market of a bitch-fit gates is supposed to do the same thing regardless of the infrastructure that it's in so the gate itself shouldn't change it's been argued that the width is not a universal quantum computer because again you need to be able to write the code in a way that that computer can understand it and that is basically all I can say about you question because I have not entirely sure I know enough for the rest I like my field is not entirely quantum computing ammo on a t-distribution I'm

like close enough I understand quantum computing and a parity I can talk for a long time about it as well which is a big surprise I wasn't expecting to be talking at this time with it but never mind [Music] yay on this video I'm gonna go and have a coffee