hey everybody welcome to another episode of EE stalk tech my name is Mike Hoffman
I'm Daniel Bogdanoff and we're back with Lee Barford you maybe heard last time we
talked about what is quantum computing and we started to get into some of the
applications of quantum computing about you know storing superpositions of data
in registers and you know how that is applicable at the very end we started
talking bout how that is applicable to security so I want to pick that up again
you talked about the quantum computing is you know one of the applications is
factoring large numbers can you kind of pick up yeah I'm really excited about
this part because I think you're you're gonna teach us how to crack RSA is that
correct okay well actually a famous guy by the name of Peter Shor taught the world
potentially had a cracked RSA well I want to back up and take just a little
bit about the history of computing I think as far as I've been as far as I
know the first person to propose it was actually the Nobel Prize winning
physicist Richard Feynman in the mid-60s and there was a small flurry of interest
but it sort of died out however in the about about thirty years later in the
mid nineties this doctor this doctor sure this published and published a
paper where he pointed out that if you had a if you could build a quantum
computer with this defined a certain defined a certain way in a certain
abstract discussion of what sorts of operations that quantum computer would
have to be able to do which are now thought to be reasonable and is the
primary model the people are working toward if you could build a quantum
computer with certain properties then he presented an algorithm for factoring a
number that had it had
finding one factor of enough of a number of no matter what it no Matt it's what
its length and doing that in a reasonable number of these of these
quantum machine code instructions that he'll out that he outlined in the paper
okay getting there he also did a number of other interesting things including
creating one of the subroutines he needs in that algorithm is a quantum fast for
a it was a quantum fast Fourier transform and so we got at least one
thing that electrical engineers are familiar with even in that very first paper now that
paper you know much of the common security that we use every day is based
not only in the RSA public key algorithm but also in an in a method for key
exchange called the Diffie Hellman key exchange algorithm and and that a
version of that for example is what's used to establish a secure connection in
HTTPs okay and so little background on on security if you don't mind so RSA I
think stands for a really secure algorithm that correct actually I may be
pronouncing those names slightly incorrectly we may need to edit that
maybe rerecord that got it yeah but it's it's it's the as far as I recall it's
the remaining public key cryptosystem that hasn't been broken in other words
your system where you can you can publish you every person who's
participating in sending receiving messages as to ease a public one that
they can that they can publish and a private one that they hold and if you
want to send a message just to let's say Alice wants to send a message to Bob
Alice looks up Bob's public key encrypts your message with Bob's public key sends
it to Bob and then only Bob who has the public key
can the private key privateer has the private key can decrypt the message okay
Diffie Hellman solves a different problem which is two people Alice and Bob are
not known to each other and yet they want to communicate right so that's
securely so that is the that's the method that can be used to establish an
HTTPS connection right you want to establish a secure connection to a
website that you've never communicated with before right um you can't make
everybody in the world who uses the web create an RSA private public key pair
that we're sort of owner s so this algorithm called Diffie Hellman is used
which also uses the difficulty if it be presumed difficulty of factoring large
numbers that have just the smart that just f2 prime factors to this to to send
a private key between the two people who want to communicate so it's a good way
to communicate very expensively but briefly send a private okay they can be
used in a much more computationally equipped computationally efficient
crypto system for the main part of the community for the main part of the of
the session and you basically take two large prime numbers or two or more and
multiply them together and you end up with this number that you know someone
assumes is a factor it has only prime numbers as a factor but can't actually
you know the the processing power to figure out what those factors are is
yeah and there's no known out it's not actually protected hard even with a
classical computer buddy there's a series with me right because if people
have been working on on the better way to fact better ways of factoring numbers
since it's like Euclid so 2500 years nobody's found a really fast way or in
the computer science link Oh nobody's found a polynomial time algorithm for
factoring large numbers the presumption is of all these smart
people going back 2500 years haven't found one it's not like me that
somebody's going to find one soup that makes sense um but it sure is sure's
algorithm um would would would be a a random but polynomial time or meaning
you know computer science talk for its fast enough algorithms to refract for
factoring okay now to bring this back to quantum computing as the is the
necessarily necessarily a fear but is it believe that quantum computers will be
able to make these kind of factoring calculations and a you know not
thousands of years yes if you could if you could build a quantum computer with
enough quantum bits and that can operate with a with a machine language cycle
time that's reasonable say microseconds or even millisecond cycle times then it
should be possible then it would be possible to do factor factor say
thousand bit numbers with with Shor's algorithm the thing that interests me
though is I talk to famous professors at famous at well-known well-regarded top
ten in the world universities and I get a huge disparity of opinion as to when
when a quantum computer of that size could be built I've had I've had a
couple tell me five years ten years no problem I've had others tell me fifty
years and so there's a big disagreement even among the really top
experts in the field as to how long how long it's going to take and what what a
quantum computer look like so I mean not it depends it's easier to describe
architecture and then the physics okay the algorithms
although people talk about quantum computers as a separate thing from a
classical computer containing you know current modern digital hardware it isn't
all of the approaches of which I'm aware the quantum computer is really kind of a
coprocessor that has to exist with with with current forms of digital
electronics and if you look at algorithms like chores Shor's algorithm
you'll see a mixture of operations that are just traditional computer operations
you know if statements store count for for loops that are okay done in you know
with them with a with a current kind of single core computer or an FPGA or some
other form of you know or an ASIC
just conventional digital electronics and then other things that have to be
done in the quantum world with these quantum bits and quantum I'll call them
assembly language operations or quantum gates the field calls them quantum gates
I don't like the term because they happen in time not in space so I being a
computer scientist I think of them as assembly language operations and they
are fundamentally different thing for setting up the kind of super positions
or Schrodinger's cat states in the register that we talked about earlier
that takes a completely different kind of assembly language operation than what
we're used to
so I want some computers I'm sorry I was gonna ask what kind of quantum computers
have been built so far today or they just do radical still or have some been
built yet no they've some some have been built but with only a few with only a
few quantum bits I believe the current claims are twenty one twenty one quantum
bits Oh we don't know what kind of it's not
clear what's how how noisy those are there are various kinds of errors that
that can creep in we don't know what their lifetimes are whether whether you
can maintain the say a superposition of States for microseconds or set were or
were seconds is is is not so clear in the press reports have a set up and hold
time there's a bit of a set up at whole time challenges what you're saying there
well it's it's not just that it says once you once you create once um now
it's time to get into some physics if you remember sure if we continue this
analogy of Schrodinger's cat as I'm as they admittedly oversimplified model for
our for quantum bit now in in the simple hot sort of high school version of
Schrodinger's cat right it's it's isolated in the box and it's means both
alive and dead until you do the measurement until you open the box or
otherwise measure whether the cat is alive or dead and then it becomes one or
the other definitively once it's measured when we in reality if you have
a little if you have a piece of physics that you've managed to put into one of
these um superimposed quantum states there's a what it would be any
disturbance pride any disturbance of it in other words every impact of a photon
of a wrong energy of an unfortunate energy will disturb the system and cause
it to go into one of those states or even possibly some other state that was
not intended in the design and so these physical manifestations of these quantum
bits have to be highly isolated from the rest
of the universe and that's done by either putting them in extreme vacuums
or in extreme cold meaning way under one degree Kelvin
oh um which is why it's become there are other there are other
approaches that are solid that a better solid state they have other different
problems okay um and it's not clear how far along those are the the famous the
company the companies and research groups that are making big claims about
we have nineteen quantum bits we have twenty one quantum bits
those aren't those in my recollection those are not the solid-state ones but
either way you have to have a high degree of a high degree of this very
high degree of isolation that isolation can't be perfect and so you have to
there is a limited lifetime that you could for the computation before you
have the before the probability gets too high of an external disturbance ruining
the calculation so forgive me if I'm backtracking here and no no please yeah
well why does it matter when the the super position bits collapse like if
we're gonna collapse it at some point you know ah
well simply because if it collapses at the wrong time you'll get a wrong answer
no okay and something like Shor's algorithm it's easy to multiply and I
think this is how I got to talking about Shor's algorithm in the first place sure
is algorithm if you multiply it's easy to check whether or not you got a right
answer if you got a factor if you've got a number that is claimed to be a factor
of your input you simply
do long division sure then you either write and you neither you get a
remainder of zero and on a classical computer and that's fast and that's
reliable highly reliable and you either get zero as the remainder or you don't
so even though the computer line quantum computation only has to be reliable
enough that you're likely enough to get the right answer that you can do a loop
of checking to see whether you got the right answer and that will convert it
quickly enough oh that will you'll find it you'll find a right answer quickly
enough and so the algorithm is resilient to errors in the qubits this sort of
premature collapsing and even can be probabilistic in the sense that it was
only supposed to get a certain probability of getting the right answer
but if the probabilities of getting the right answer are high enough then it
doesn't matter that you that you you can afford to get the wrong answer some
number of times because in the end in the end you you'll get yo you can
verifiably get the correct answer and if the probabilities are high enough then
you'll get it much faster still than a classic computer ok question for you
without making my head explode because this stuff is definitely complicated but
we're talking about the the bits they have certain probabilities of states so
I'm assuming those probabilities can be variable right it's not like 5050 so how
how does the computer itself or maybe I'm assuming there's some sort of
physical thing that's going on how are the the probabilities of those bits
actually changed like well how do you what are you doing
yep injecting energy into the bed or something to do with the spin of the
electrons or what it depends on it depends on the physical system and the
the this is where my knowledge starts to really run out however in the approaches
where keysight is making a biggest contribution
injecting energy into the system and you're doing that as a very small number
of photons at them typically in a microwave energy with a careful
controlled as a pulse with a carefully controlled timing and phase okay and so
sort of work we're at the moment where keysight ends is it is at the generation
and and measurement of of those pulses and so that sort of explains why
keysight right we're good at microwaves we're good at generating them and and
measuring them at at at metro with metrological
levels of precision and then i mentioned that you have to do though you have to
do that it very carefully control time preferably sub sub nanosecond time
accuracy and if you're gonna have multiple quantum bits you have to have
that sort of level of time synchronization across across all the
bits because if I'm going to manipulate all the bits right now to have a multi
bit data then I need to be able to do those X I I need to be able to generate
those pulses across all the channels for for all all the bits with that sort of
third sorts of levels of timing accuracy and that's something in particular that
esight brings to the party that the that the physicists who do understand that
the physics beyond that level in the in the little tiny highly isolated thing
which is going to be the quantum which is gonna be a quantum bit what that's
that's what they gain from us and that's that's that's that's what we're helping
now what is actually out of time again one more question go ahead yeah what is
the quantum bit is it like an atom or atoms are two common con two common
kinds is are are ions trapped at trapped in a vacuum with with
laser trapping so you have interacting laser beams and the these ions can't
move can't move because they're held in place by us by standing waves of
Louisville laser beams and so that's their isolation that's their
low-temperature right the that the vacuum vessel can be it at room
temperature but the ions have very low temperature because they can't move sure
okay another kind are done with done with Josephson junctions so they're
nonlinear nonlinear circuit elements that operate at extremely low
temperatures and what's built is basically tank circuits so you have a
coil capacitor and a and this joseph´s of junction and those together produce
produce oscillations at microwave frequencies at different modes
however at sufficiently under the right physical conditions those can be
designed to behave just like just like an ideal to stay quantum an abstract
two-state quantum system with two states you just designate can designate zero
and one and behave as an ideal as an idealized idealized a bit that you can
manipulate the properly the probabilities of that bit being zero
zero and one actually what the quantum called amplitude another level of
complication where there's complex math I'm gonna try to avoid that and we'll
just call them probabilities thank you on the comments below the real
physicists will it can explain why that's completely wrong
you actually mean what they call complex amplitudes for the current purposes
let's just think of it as the probabilities of being 0 at 0 & 1 so the
it it's so so to try to summarize out all that the physicists have an ideal
oscillate non non non linear oscillator that has the right mathematical
properties that corresponds well to the assembly language instructions that you
want to write that manipulates these probabilities across single or multiple
quantum bits and that turns it that that nonlinear oscillator you can build an
extreme cold with one of these junctions thanks okay
and I think keysight had I've been meeting trade a blog post on the
Josephson Junction because it's a fascinating device that's where we get
our you know if you think back to one of our early podcasts we talked about
different standards the Josephson junction is the root of the standard for
what voltages and I think keysight had the first commercially available justice
ignition up in her metrology lab in Loveland we're out of time for today
again the time has just been flying by we're gonna definitely have Lee we're
definitely gonna have you back on for another episode maybe we'll talk about
parallel computing or some other topic well we'll talk about definitely I'm not
sure if I feel smarter or dumber after these conversations it's crazy
stuff well and the thing is every every time I go for a visit
folks who are really doing this work I come away feeling a lot dumber - yeah
it's you know in and I'm I I consider myself very much a beginner in this
despite you know this despite what I've that's what they've all sort of what
I've said I you know III am very I I'm very much beginners myself geez what
does that make us so we uh so we like to have a session a section
at the end of our podcast that we recently added and it's we like to ask
our guests what we call a stupid question I have one for you I'm gonna
steal my thunder and then we're gonna yeah you don't even get ones all right
done I'll do that do it next time if you had Schrodinger's cat in a box would you
look or not and the answer so I really had Schrodinger's cat in a box I would
go ahead and look because the cat's at room temperature and so in fact the
experiment is already over before I look so freeze the cat first basically okay
yeah the cat's wavefunction really collapsed way back from some random
infrared photons floating around the box just happened to happen to do the raw
happen to happen and and it's so it's already alive or dead and the fact that
I look in or not doesn't matter I was looking at this is a glass half full
glass half empty kind of question but that was a whole new level so thank you
least so much for coming on again it's been a pleasure having you will have you
back for another one make sure you subscribe to the keysight oscilloscopes
youtube channel check us out on your favorite podcast engine, EEs Talk Tech electrical engineering pdocast
and we will see you next time Cheers
For more infomation >> Sonic Adventure 2 - City Espace 16-bits remix + vocal - Duration: 3:03. 
For more infomation >> Audi A6 Avant 3.0 TDi BiT Quattro Pro Line Plus VOL - Duration: 0:56. 
Không có nhận xét nào:
Đăng nhận xét