Thanks.
As David said Thanks for
having me Spencer Falak I'm representing
the cost of quantum systems of vision
which is right here if you want to.
Ask them any questions after the talk and
I'll be talking about advances in AI and
trap quantum computing and it's no concern
if those words don't mean anything to
you at this point the talk is intended to
be you know soup to nuts introductory.
You know what is quantum
computing what are trap diets and
how do we work with them.
On that note that's a question I just
asked which is what is quantum computing
and how does it compare to classical
computing So for us classical computing
means by an area logic with transistors
the same type of logic that goes
on in your cell phone or
your MacBook Pro or your F.P.G.A.
and I'll talk a little about quantum
computing it's a completely separate
paradigm where we use a new type of bit
called quantum bits to solve different
types of problems that you can't solve on
your own your best classical computer.
I'll talk next about trapped which
are our choices of quantum bit
talk about why we've we use them
what are their advantages and
how do we control them to sort of change
their state in the same way that you might
change the state of a classical computer.
Next I'll talk a bit little
bit about scaling so
how do we go from our current
experiments which are a few
a few quantum bits up to two hundred so
maybe thousands if hope.
One day that will be the case just like
you know there was an exponential speed up
in classical computing and then finally
highlight one of our newest algorithm
demonstrations very briefly just to give
a feel for where the fields that but
also how far we have to go to
really do a powerful computation.
And I just like to point out how
multi-disciplinary this field is and
I know there are a lot of students in here
from different backgrounds and the talk
is sort of split up into thirds where the
first third is more computer science and
then we get into atomic physics and then
sort of what the Nano seminars are more
about electrical engineering which is
using micro fabrication as a tool to
to advance our research and
the goals of the field.
So here's a little bit of the history
of classical computing and
this is something that you might see on
day one of electrical engineering class.
But what's useful for us is to sort
of look at it as a comparison for
quantum computing and to see maybe if
we are following the same track so
classical computing starts out in the
seventeen hundreds with the introduction
of binary logic followed by Boolean
algebra and the eight hundred fifty S.
and then maybe one of the most important
inventions of the twentieth century is
the transistor at Bell Labs by Bardeen
Bradley supervised by Bill Shockley.
And then in one hundred fifty eight
with the first integrated circuit.
From nine hundred sixty five until
now we've seen incredible growth
in classical computing there's
you know companies like Intel and
Moore's Law which say we can double our
device complexity every year not here to
talk about Moore's law there are certainly
skeptics of the future of that but
I'm here to talk about a completely
different paradigm and
maybe something that's separate and
you know today we have incredible
supercomputers that you know have twenty
Petta flops of logic operations that
can do incredible things and
I'd like to compare this waiter at.
To where we're at in
the quantum computing field.
So there are some limits
on the way that we've
approached scaling we use lots of
large clusters operating in parallel
to solve some of the most complex problems
we can add graphical processors or
F.P.G.A. to supplement our cluster but
there are certain problems that aren't
really tractable with this type of scaling
and in one thousand nine hundred one
Richard Feynman propose that there are
some problems that could be solved better
with a quantum bit with
a completely different style.
And that brings us sort of to the history
of quantum computing which you can
sort of make a comparison to
what I just showed you before.
The one thousand eight hundred Max Planck
quantizes light to account for
black body right eight radiation
there's big names like zero and
showed in German Albert Einstein who sort
of formulate modern quantum mechanics and
then eighty years later Richard Feynman
publishes this quantum computing lectures
which talks about in a very abstract way
about how there could be such thing as
a quantum computer and it takes another
fifteen years before this sort of abstract
idea becomes a reality when researchers at
the National Institute of Standards and
Technology perform the first two quantum
bit logic gate and it's for this work that
Dave Wineland who's a physicist at NIST
won the Nobel Prize a few years ago and
you know in the past ten years since that
sort of the field has started to take off
but it hasn't been very quickly it's sort
of the same way that classical computing
take has taken off and we've gone from one
sort of gate and now we can sort of do
algorithms with three to five quantum
bits very small number of bits and
what we're hoping for is the same type
of rapid growth that classical computing
saw from one nine hundred sixty S.
until now where we can scale up from
five bits to thousands hopefully.
So we talked a little about quantum.
Computing in the history but
what is it actually useful for and
right now it's a bit unknown there are a
few algorithms that have been theorized
that will be much more
efficient on a quantum computer
the key algorithm is Shore's factorization
algorithm which I'll talk about a bit
about in the next slide there are some
other applications here the discrete for
a transform simulating quantum systems and
then solving linear systems but
the current currently the quantum speed up
is limited to only a few algorithms and
we expect or we hope that as the field
expands they'll be more algorithms
that sort of bootstrap along with
the development of new technology
that can also surpass their
classical counterparts.
So here's a source factorization
algorithm and this is probably the king
of the quantum algorithms because it's the
one that's most interesting to the federal
government and the reason is that R.S.A.
public private key distribution
assumes private factorizations
computationally hard are saying corruption
is ubiquitous it's used on Internet
security cellphone password things
are not source and so for much of your
bank account password things like that.
And the intelligence community has very
strong interest in an algorithm which
would either render this type of system
useless or maybe come up with a new
way to avoid other countries from
breaking our current encryption code so
here's an example of classical scaling and
quantum scaling for
factoring a prime number and
as you can see as you add more and
more digits the quantum scale is far far
better than the classical scaling it takes
far far fewer number of operations to to
factor a number Quantum with a quantum
computer than it would classically and
this is sort of why this field took off so
so quickly and here's sort of like
a an example you know what's the biggest
number that one can factor in
a reasonable amount of time.
If I sat down with a pen and paper maybe
I could do a three digit number and
your MacBook Pro could maybe do a fifteen
digit number if you wrote a nifty little
algorithm and then the best published
computer could do a two hundred
thirty one digit number and we're hoping
that you know a quantum computer could do
much much larger more complicated
number in a reasonable amount of time.
So we've talked a bit about
quantum computing and
maybe a few of the algorithms that will
surpass a classical counterpart but
what exactly does that mean what is
the quantum bit how does it work.
And there are two key components that I'll
talk about the first is superposition and
the second is entanglement and
these are two.
Key points to describing quantum bits so
what's.
So here's an example of a classical
bit that has a value of zero or
one you can make
the analogy to a switch and
it just turns on and off and quantum
mechanically we have this principle called
superposition which I'll talk
about in the next slide.
But here let's look at an example of a two
bit word which holds for value zero zero
zero one one zero one one one and then.
We can look at a quantum bit and
we describe it with this
state notation side where these are
complex coefficients A zero and a one and
what you'll notice here is that each
quantum bit of this quantum state that
we've written here can store information
about both zero and one simultaneously.
But there's sort of a weird Kaviak
which is if you measure the quantum bit
the state will collapse into zero or one
with probability given by the magnitude
of these coefficients squared and the
analogy I like to use it's not a perfect
analogy because these are quantum
mechanical States but is a weighted coin
a sort of like you have a heads or tails
and when you flip the coin it away and
either heads or tails and
you can change the weighting of that coin.
About how we do that
with quantum bits later.
And then to return to the example
from the previous slide
if we look at it two cubits state it
now requires two to two squared for
complex numbers or two to the third
real coefficients to describe So
this is more information in each quantum
state than that in a classical state and
this is sort of a powerful implication for
computing.
The second key point I'll talk
about is entanglement and
this is the ability to link to
quantum bits so that their states
cannot be described separately and
there really isn't any classical.
And that like comparison to
this in classical computing and
the analogy that I like to use before
which is if you had two coins and
flipped them they would both land heads or
both land tails together and
never opposite from each other and
this weird entanglement is something that
troubled Einstein and other researchers
in quantum mechanics for a long time but
it's it happens to be a very key component
to doing a quantum computation so
let's return back to our outline for
a second I talked a little bit about
what was quantum computing and
now it's sort of what are the key
components of quantum computing but
now all talk about trapped ions and
trapped ions are what we use a teacher or
I for quantum bits and
I'll sort of highlight how we
control the trap Dion's and how we can
generate superposition entanglement to do
interesting algorithms
with these particles.
So just before I dive into trapped
on thought talk a little bit about
the technology that people have used over
the past ten years to as this field has
grown people have tried to make quantum
bits out of polarization States and
photons circuit in superconducting loops
to the electron gases quantum dots and
then.
Energy States in neutral atoms and
here's a picture of a lattice of
neutral atoms and then here's a.
Picture of chain of trapped ions
which is what we have a D.T.I.
this is a picture from our lab and
over the past ten years it's sort of
become clear that superconductors and
trap diamonds are the most promising
candidates they have their
trade offs superconducting cubits
are incredibly fast you can do a quantum
operation in nano seconds whereas in trap
times it's on the order of microsecond but
the coherence time which is really
the ability to to do operations.
Coherently with previous operations and
read them out coherently is far longer for
trap Dion's than it is for superconductors
and sort of as an aside this you know
picture of five or six cubit
superconducting circuit at Santa Barbara
is their team has recently merged
with Google to continue this work so
it sort of gives you an idea of
the hype surrounding this field and
sort of where people think it might go.
So moving into trap Diane's.
Cubits are the species that we're using.
We have calcium terbium and
we're starting work with strontium.
I'll talk mostly about you terbium
that's the chapter on that we've done
the most interesting
algorithm work when recently.
And so
let's get in now to how do you trap and
so here's a really small
picture of an eye and trap and
I'll talk about that in the next slide and
we put the trap into a vacuum and
the reason for that is each eye obviously
doesn't weigh very much and if it
collides with the particle an area could
get knocked out of the trap it's also very
sensitive to magnetic field fluctuations
so we put magnetic field shielding and
biased magnetic fields around the vacuum.
And then what else do we need to trap and
we need laser cooling to cool it down and
then of course we need a source of
neutrally terbium or calcium and
photo ionisation laser kicks
an electron off the neutral source.
To turn it into an island.
So here's a picture in of the trap that I
showed you before this is the classic for
a trap that's pretty popular among groups.
And what's interesting about it is this
radio frequency all RAF electrode pair and
you think why do I need to put radio
frequency to to trap in Iran and
the answer is you can't actually make
a three dimensional potential little so
that's how we trap the UN in
a electric potential mill.
With static electric fields by the you
know some of the fundamental questions
of electrostatics so Wolfgang Pauli
to determine that you could use
radio frequency voltages to create
a ponder motive force which is sort of
a net inward force that's created
by this R F A lecturer and
here's a cool little video of an eye and
bouncing around in.
You know the profile view of that trap and
it's oscillating at
the two different frequencies I'm only
showing the slow frequency here and
then you there you can see sort of
a cartoon picture of laser cooling and
its extent of its oscillations will
reduce over time the mechanical analogy
that Wolfgang Paul uses in his Nobel
paper is that of a ball on a saddle and
the idea that if you put a ball on
the saddle and spin the saddle fast enough
the ball will not slide down
any of the sides of the saddle.
So we just looked at this trap here and
now are looking along the axis of the for
a trap and there are certainly some
disadvantages which when it comes to
scaling a trap like that you can see here
there's only three segmented electrodes
we use these electrodes to move and move
on trap which moves the irons back and
forth left and right there's also some
difficulty in the lining these electrodes
you know they're macro structures maybe
the size of a couple millimeters and
in the mid two thousand
researchers at NIST
determine that you could
actually fold the structure as.
Into a plane or structure and the word
planner probably means something you know
a lot in this building where people spend
a lot of time working on silicon chips and
it is enabled us to take some
of the technologies from.
Silicon chip manufacturing and
utilize them to our benefit so
how do we do that we've this is a cartoon
drawing of one of our chips and
what the planners ation does is
it allows us to make many many
electrodes these short segment electrodes
which we can use to move the on back and
forth and there you see the R.F.
potential that I discussed before and
here you can see sort of the layering
build up of the of the silicon ion trap
chip Here's a colored in picture of one
of the chips that we fabricated here
at the nanotech research center there's
hardly Hayden who's our chief bad guy
who made this trap and here's a zoomed
out picture of it it's one of our
older traps the Gen three trap and
you can see the slot in the center
you can't really see
the small lecture it's here.
In here a lot of capacitors which we use
to filter out noise pick up on these
on these electrodes and
wire bonds out to the external world and
I'll talk a little later about
how we've modified this design
over the years to remove some of these
capacitors or make them less bulky and
then move the wire bombs away so
we have more access for lasers and
other techniques that we've
used to improve this design so
here is a video from two thousand and
eleven two videos which show you
sort of what you can do with these traps
and it's pretty awesome this is a chain
of islands individual islands sort of
getting merge and split back and forth and
therefore wrestling with the cooling laser
light and you know someone who goes and
sees this every single day it kind of
loses its alert maybe a little bit but
if you it's pretty wild to think
about that you're looking at
an individual particle on your screen.
In here getting pushed back and forth and
controlled it's location being controlled
really well and here on the bottom
you can see as we load it chain up
the ions are coming up from the neutral
oven source we talked about for
the ionized and then getting transported
over there to build the chain.
OK so now we have ions in there in
the trap but this talk is about quantum
computation so how do we make
superposition how do we make entanglement
with these individual and so here's
an energy level diagram of you terbium
and there are two relevant transitions
I'll talk a bit about today there's some
added complications as well but the ones
I'll focus on are this transition at three
hundred sixty nine point five nanometers
and this one at twelve point six you get
hurt in the microwave band and here we
have two energy levels of very terbium
mine which correspond towards zero and
one quantum states of our quantum bit.
And this top transition is our detection
transition it's how we flip the coin or
measure what state our quantum bit is in
and then this is our gate transition and
if we shine resonant radiation here we
can switch the state of our quantum
bit from zero to one.
So how do we figure out whether our
quantum bits in the zero one what
we do is we shine through in a sixty
nine then a meter light onto our own and
if it's the atom is projected into the one
state it will scatter photons which we
can collect on a P.M.T.
for the multiplier two and
we'll get twenty photon counts histogram
like that if we repeat it over and
over again on the contrary if
the atoms in the zero state and
we shine the same light this light is no
longer resonant with this transition and
we won't get any photon counts so you can
see how we can easily sort of distinguish
between these two states by saying if
we have more than eight photons or so
on our P.M.T. the Ion was one and
if we had less than a photons and zero.
OK so now we know how to
read out our quantum bit.
So how do we actually.
Change the state of our quantum
bit we use resonant microwaves to
drive transitions so this is a microwave
horn which is a cartoon drawing
at twelve point six gigahertz to drive
the state of our quantum bit back and
forth Alternatively we can use
two laser beams that are detuned
by this twelve point six you get Hertz to
drive the same transition and what will
get here is sort of a signature of quantum
mechanics this is called a Robbie flopping
curve where the the state of the eye and
goes from zero to one and back and forth.
Just like the cartoon.
So what's illustrate a little bit
about superposition which is what we
talked about before so
we here at this point on the curve we know
what are our history to look like it'll
it'll be we'll get almost no counts
because the items in the zero state.
And at the top will get a bunch
of twenty photon counts or
so because I was in the one state but what
about here on the raw beef up in curve and
this is sort of where the power of quantum
computation comes in is half the time
we're going to get of a bright histogram
and half the time going to get a dark
histogram and that's the state
described here where each.
State is each level is described
by this probability coefficient So
that's something that you
really do not get classically.
So let's return to our
trapping out of apparatus
I've replaced the old classic
trap with one of our G.'s here
I had micro fabricated traps
now we're strapping in terbium.
But we've got to add a bunch of
different things now we've got to add
radio frequency electrode for confinement
we've got to add one hundred diddle
digital analog conversions to move ions
back and forth with the outside electrodes
we've got to add a big lens stack
to collect light for detection
to measure what state our quantum bit was
in and we've got to add microwaves or gate
lasers to perform quantum logic so this
picture has gotten a little bit complex a.
And what does it actually look
like in a lab it's pretty wild.
You got optical fibers going
everywhere this is one half of one
room that's filled with tons of equipment
and here if you zoom in on one of our
traps you can see the magnetic bias field
you can see the lens stack to our P.M.T.
you can see the digital analog converters
coming into our vacuum chamber and
the radio frequency electrodes.
So that sort of brings me to the next
point which is scaling and the purpose
of scaling isn't necessarily to shrink
the size of the equipment in the room but
it's to get more quantum bits that we
can control with higher fidelity and
detect faster and
with better detection fidelity and
the technique that we've adopted
at is to build integrated service
a lecture traps which is to take some of
the components I just talked about and
smush them into our service electro trap
and use micro fabrication as a tool for
example we can replace the microwave horn
with integrated waveguides into our micro
into our our trap so we don't need
a horn radiating onto our vacuum instead
we just send microwave lines
into our trap and can do.
Quantum logic with the built in waveguides
and this is something that we did in
two thousand and thirteen so
now highlight three of our most recent
integrated trap chip trap designs some
that we've done in-house some that we've
done in partnership with
Honeywell International and others which
really try to solve some of the scaling
problems and make life easier for
the experimentalists So the first thing
that we've worked on is trying to improve
the detection issue and the the picture I
just showed you you know in real life was
a big one stack going into a mirror in
a P.M.T. or onto a camera which is how we
saw the ions moving back and forth before
so in two thousand and eleven we said Well
we're we're not really collecting very
much light from the atom when we get.
Twenty photon counts in two hundred or
four hundred microseconds it's it takes
quite a while to figure out what state are
I in is in so what if we put a mirror onto
our trap surface we can collect
more light from from the eye and
as we detect it and here's a cool
video of an eye in transporting
over the mirror that we put onto this trap
and you can see as it surpasses the mirror
we start to collect more light that's
actually reflecting off the surface so
this is one idea that we had to use
integration of microfiber cation to maybe
reduce the load and
speed up this process so the integrated.
Trap was it worked well but
there's these funky electro designs and
it's sort of difficult to manufacture so
recently with partners at
the friend Hoffer Institute as well as
a group at the university we decided
to replace the mirror with the fact of
optics on the surface of the track.
And here's a picture of the chap that
I believe it was fabricated here and
the mirrors and the colonnade and
other diffract of optics were patterned in
Germany was yeah so
here's a picture of the track it's
got the key component here the collimated
which collimate some of the light from
the eye and so that we can collect more
this is really the name of the game and
researchers at Griffith University
using the strap were able to actually
adjust the height of the eye and to get
a great diffraction limited limited image
of the eye and so which isn't really
particularly useful but if you but
you could add a couple to the light
from the eye into a fiber but
it's also very interesting
to see that you could.
Collect more light for faster detection.
So the second idea we had was to take.
The hundred digital analog conversion
converters which are really bulky and
this is sort of a picture
of how we wire up wire.
The one hundred channels from that
card onto our trap chip used to
move ions back and forth and this I mean
there's so many wires here trying to get
that all into a vacuum while keeping very
high vacuum qualities a difficult task so
what we decided to do was
to put off the shelf.
Digital analog converters integrated
with the chip integrated with
the ion trap chip into the vacuum and
this way we could reduce the overhead
of the the number of wires here and just
replace it with the serial interface and
you know add two channels you get forty
more digital analog converters and instead
of using a ninety six vacuum few things
are only using nine vacuum Peters now.
So here's a picture of some of the work
done in Honeywell International where they
took some off the shelf
digital analog converters and
actually DI capsulated them with acids to.
Put onto a Rogers P.C.B. board and this is
one side of the P.C.B. board on the other
side is our trap chip where you can
see in a nice compact package here and
this is sort of another way that we've
used this is necessary microfiber cation
But you know try to scale up and
reduce the overhead
on the experimentalist and
make things much more easy to work with.
In the final trap that I'll talk a bit
about today is called a ball grid or
a trap and the purpose of this is to
reduce the area of the trap itself.
Instead of using those planar capacitors
which took up a whole large portion of
the trap here we can replace them instead
with through substrate vias and capacitors
there so now we move to wire bonds for
a laser axis way down to interpose or
in the trap chip area takes up a far
smaller amount than it did before and
this allows for tight laser focusing so
you can hit one at a time do Gates on one
at a time I've actually got a bum
version of the strap here we can pass it
around it's got some additional wire bond
scratches on it but you can see what it.
It's like in real life.
And here's a picture of the trap here you
can see the radio frequency electrode
coming in here and here is a picture
of some of the trench capacitors and
the through substrate vias that we've used
and the key takeaway here is we've taken
this whole surface and reduced it
down to a very small area here so
that we can focus lasers more tightly and
address Adams individually.
So what we want to do with all
this I mean this isn't just for
the sake of what it is for
the sake of making life easier but
it's also about making
a complex quantum system and
the idea here is to smush all of these
things together that we've built so
we would add the gates region where we
have tightly focused laser beams that I
just discussed we have a detection reason
we read out our quantum bits with diffract
of optics we could have junctions which is
a trap that we built in two thousand and
thirteen here at Georgia Tech where
you could rearrange your eye and
chain and move them back and
forth sort of like a register.
And then if you really wanted you
could integrate optics you know
into the trap surface to do gates and
we haven't really worked with that yet.
OK So back to the outline.
What we talked about thus far we talked
about a little bit about quantum computing
we talked about trap Dianne's and
we talked about scaling but
now I'd like to give you guys a and
feel for where we're at in the field.
You know it's going to take
a large number of quantum
bits to do an interesting algorithm.
Unfortunately we're not quite there yet
but
we're we've been pushing the frontier
doing cutting edge research
on basic quantum algorithms and
that sort of where we're at right now and
here's sort of a path to where
we've come from two thousand and
eight all the way to now we started in two
thousand and eight doing trap development
here's a picture of the same junction
it was one of our older ones yeah.
Then we started doing single
cubit control in two thousand and
twelve two thousand and
thirteen and then we started.
Making scalable traps with
integrated features and
simultaneously working out a bit control
three four five quantum bits and
then now we're working on simple
algorithms with two or three trapped ions.
So the algorithm we've been working on
lately is called the Bernstein Vasser Ani
algorithm and
how it works is sort of a guessing game or
trying to figure out
an unknown bit string ass and
there's a black box Oracle and
the Oracle gives you.
Two So you were X. is the string that you
input so you say to the Oracle I here's
my bit string X.
will you give me this this answer out X.
two and the goal is to figure out
what the hidden string is here so
for example you could Querrey your
Oracle with one zero zero and
the if asked for a one one zero you
would get out one and to determine S.
precisely you need to run this three
times once with one zero zero and
then zero one zero zero zero one So
it takes three runs of the algorithm
to figure out a three bit
string as a hidden string.
Following this quantum circuit which I'm
not really going to describe you can
determine S. in a single query so
it's a simple algorithm not really
factoring large numbers but what it's
doing is trying to figure out a show
a quantum speed up on a small scale system
one that we can hopefully expand later.
So here's a picture in a paper
that we published recently
showing all of the gate laser pulses that
is required to do this algorithm and
each one of these purple bars represents
a single cubit rotations seventeen of them
this is an untangling
operation in orange and
it takes six hundred fifty gate laser
pulses nine point eight milliseconds and
nineteen transport operations
to do this simple algorithm so
what's my point here this is it we're
working with three bits here but
it's it's really difficult research
there's lots of control that has to go in.
There's lots of you know stabilization
keeping the environment outside of our
quantum bits and classically from the from
this algorithm you gain one point
one bit of information every time you run
it quantum mechanically we showed that we
can get one point one bits of information
so a slight improvement over the classical
or the right what's my point here is
that this research is cutting edge but
we still have a lot a lot of work to do
to do really powerful calculations and
that's something that we hope
to be at the forefront of.
So in summary.
We've got quantum computers which can
supplement classical computers to solve
certain key algorithms the main one to
take away one is factoring which is
important to the intelligence community
entanglement in superposition are the key
components of a quantum computer
which enable this type of speed up.
A G.'s your I we use trapped.
Their strong choice of cubit but
scaling is difficult we've
just started to work with short
chains of three to five ons and
to sort of scale up our focus has been on
building integrated chip traps using micro
fabrication to our advantage to bring
things into the vacuum system itself and
finally basic quantum algorithms with a
few quantum bits are within our reach now.
The speed up is nothing to write home
about but hopefully within the next few
years we can start expanding our systems
and do really cool calculations.
This is our team here I'd like to thank
particularly Jason to me Adam Meyer and
to Merrill who made some of the slides
in this talk and thank you.
Two mature superconductors in the.
Simple energy level structures that
resemble hydrogen when you take one
electron away is sort of a criteria for
us and what a good two level
system isolated to level system that can
represent a quantum bit well as your own
one.
Where you put the neutral
the terbium are cast and.
Yet it's unlike a metal of unsourced
which we heat up with current to
spew off as as a gas.
It's not on the chip No it's below
the chip and it comes through a hole.
That gas comes through
a hole into the yeah.
Yeah.
I didn't talk a bit about that it's
a little bit more complicated there's
a procedure called a moment Sorenson to
keep the gate which we like to use but
we actually use the motion
of the atoms as a bus to
to connect their internal logic states.
And I can talk a little bit more
about that later if you're interested
in.
Yeah.
They can decode here there's
heating which occurs on the motional say
such a heat up if they're not cooled and
they can actually leave the trap it's
also just hard to control you know and
laser beams that are directed at all
the different islands have to have very
precise frequency very precise phase
to do quantum logic gates as have many
other things you want to add to the list
goes on and on to be ousted him.
Yeah.
Right so that's the sort of the the
question it's not really a question it's
the trade off in and doing these types
of computation is that you can actually
propagate your input which represents
many states at once so you can so
you can run many calculations in parallel
but at the end it's going to collapse into
one state and the trick to
creating smart quantum algorithms
that few people have determined is how to
trade off these two things so that you can
propagate through the calculation in
parallel and still learn something about
interesting about your your system
with the collapsed measurement to some
extent.
Yeah for us decoherence is
actually not a big problem.
It's we have coherence times
on the order of a second or
two the superconducting cubits are the
ones that have the coherence problems they
do go here in there so
our problem is more about
getting lasers that can do
coherent interactions and
maintaining a static environment
without magnetic field fluctuations.
Yes' or anything else because well yeah.
Sure.
Yeah they superconductors tend to be much
faster they can do their gates in seconds
whereas it takes us microseconds or a few
microseconds However they're decoherence
is much more quick but if you look at the
ratio of the number of coherent operations
you can do to the coherence time it's
similar for the two technologies.