Dan: Welcome back to The Quantum Divide.

This week, we're speaking to
Nathan from the Unitary Fund.

Unitary Fund is a unique, open
source organization, in quantum.

And.

They've got so much interesting
stuff going on that we're going to

try and get, get through it all.

but also do a bit of a.

, dive into some of the
projects that are going on.

Inside the Unitary Fund.

And how you can get involved
if this is something which is

pushing the buttons for you.

Nathan, thank you very
much for joining us today.

I wanted to say before you started,
actually one of the first organisations

I came across when I was researching,
what Organisations what what people

are out there looking at Quantum
technology, Unitary Fund was one of the

first things I came across actually, and
it's quite unique in that case as well.

I'm so, so glad to have
you here to talk about it.

So let me hand over to you.

Give us a bit of a view on your background
how you got into Quantum, how you

got in, ended up at the Unitary Fund.

Nathan: Thanks so much, Dan.

Yeah, it's a pleasure to be here.

My name is Nathan Shammah.

I am a Chief Technology
Officer at the Unitary Fund.

My background is in physics.

I'm a theoretical physicist.

I started studying foundations of
quantum mechanics and information

quantum information processing, but
moved on to the simulation of many

body quantum systems in solid-state
devices for applications such as

terahertz technology, but also more
fundamental questions about correlated

systems quantum optics and phase
transitions, quantum phase transitions

and systems out of equilibrium.

And when I was doing these
studies I started using a free

and open source tool called QuTip.

Which if you are a early stage
researcher in quantum what now are

called like quantum technologies
can really save you up a lot of time

and give you also a lot of insight.

And I ended up actually doing my postdoc
in the group in Japan at Riken, which

is a Japan national lab where the
software was first developed and also

harbored and nurtured for many years.

So that's when actually with one
student who was interning at the

time, Shanoz Ahmed I did this sort
of switch from using open source

software to starting to think, to
contribute to the open source community.

So we had developed an algorithm to
simulate, many body quantum systems

in a noisy environment under very
realistic conditions that usually

if you don't take advantage of some
symmetries are just unfeasible to

simulate on on non-quantum computers.

Because, with 20 qubits when you add the
noise, if you don't make approximation,

you're really out of a game.

And we really thought, oh, this could,
this is a tool that can be used not

only for super radians, which we were
studying, but for a bunch of other stuff.

And Shahnawaz taught me how to do
open source software contributions.

He was previously a student in Google
Summer of Code, which is a great program

by Google helping open source projects.

And then progressively, I basically
became one of the coordinators and helpers

in the community of a QuTip project.

That's why when when Will pinged me
checking if I was interested in in joining

the Unitary Fund I was really elated
because it's a nonprofit foundation

who, whose aim is really to make quantum
technologies as open as possible so

that they can benefit the most people.

And one, one of the focuses really on
open source software because one way I see

it is that open source software enables
to some extent to do research at scale.

Research is something that inherently
is uh, is done in um, in a custom way.

It's almost like craft work.

You go, you learn from a
master and from a mentor.

But open source tools really
allow to accelerate discovery

and avoid reinventing the wheel.

So for this reason Unitary Fund is
a unique entity in the quantum space

because it promotes an open community,
which also means promoting open source

tools and some things we do by enabling
others to achieve their goals with grants.

Mentorship may, maybe we can speak a
bit about that later, but some things

we're uniquely positioned to, to tackle
long-term projects with a, from a

position that is different from standard
academia or corporate or startup players

which have, they have other drive and
they should have other focus of course.

Dan: Yeah.

Well, Listen let's, let's go
into the Unitary Fund now.

I know the three tenets are
micro-grant, research and community.

Give us a quick overview of those and
how they interlink and perhaps how

they support the community as a whole.

Nathan: Yes.

The first thing that we do, as
you mentioned, is giving these

microgrants, they are non strings
attached grants for explorers in

quantum technology and quantum science.

They can be students,
researchers, hobbyist.

Who want to have some validation
of a project of their own.

Sometimes it's individual performers,
sometimes it's a team of folks,

and these are $4,000 grants, no
strings attached that are given.

Usually it's project that
lasts about six months or less.

We have mid-budget team and some and a
group of volunteers that help not only

assign the grant, but also give some
mentorship throughout the projects.

So far, we've given over 90 grants
worldwide in over 26 countries.

So that's the first thing
that Unitary Fund does.

And then, as you rightly mentioned,
also, the Discord server.

Around which many of
our activities revolve.

We have a series of community
calls from open source projects.

Some that are developed mainly and
maintained by unitary fund technical

stuff, some by other communities.

And it's a place to discover more
about what's happening in quantum

software, but also quantum hardware.

We have some programs that allow people
to get more exposure to, to quantum and

get, get started or start tinkering on
open problems that are can be addressed.

So one of them is a, a special kind of
hackathon called Unitary Hack, which

puts in touch maintainers of existing
quantum projects such as Qiskit.

Or PyZX or many others.

Last year we had uh, 30, over 30
projects participating, and we put

them con in contact with a community
of, Hackers that want to help and

have a opportunity to already focus
on specific open issues that have been

listed by maintainers of these projects.

And Unitary Fund creates this platform
so that you have a sort of like

a bounty-based mechanism so that
when the issue is is closed by a

given hacker they can get a reward.

If it is accepted, the
solution is accepted.

And so it's, it's a way
to get started in Quantum.

We also do other other things
that are such as the open source.

Survey in Quantum.

It's inspired by the Stack Overflow
uh, developer survey and provides

quite unique information from an
independent source on the state of

quantum software serving especially
those developers who develop quantum

software and use quantum software today.

And finally, there is a
research that we perform.

It's a special kind of
research that we do.

We do or we do in, in the open,
developing open source projects.

that can enable others to advance the
same mechanism I was mentioning earlier.

So we have currently, I would say
two main projects and three or four

lines of interest from research side.

So the two projects are one on
benchmarking, which is called Metriq.

Which you can find that Metriq
with a q.info, which is a platform,

a web platform to really get to
know what is the state of the art

of quantum computing benchmarks.

And everyone can submit their own uh, or,
or results that have in the literature.

They can tie a data
point to an arXiv paper.

And it's uh, it's a framework that
it's inspired from other projects,

successful projects in fields such
as AI with uh, papers, with code.

So happy to speak more about
Metriq, but to get to your the

final points of a research site.

We're also active in error mitigation.

So error mitigation is
a a set of techniques.

Which allowed to improve results
of quantum programs that run

on current noise-intermediate
scale quantum computers.

And we've developed mitiq again, Mitiq-Q
with a Q, the final Q, which is a Python

toolkit to perform error mitigation
on multiple front-end and providers.

And basically any backend as a
hardware or simulated backend.

So these are the two main projects
we have, but as I was mentioning,

we also have some other focus on
simulation of quantum computers

with by classical means and also on
application-oriented solutions for using

quantum simulators and quantum computers.

Steve: Yeah, it's a broad
scope of activities.

It's very interesting and there's a
lot of very important topics covered,

especially for bringing community out
and helping people get into quantum,

especially, the rate at which Quantum
is advancing is the rate of which the

community advancing at the same, so
Unitary Fund is helping to keep the works

of the workforce alive and supplied.

But I wanted to ask a little
bit more about technical

aspect of error mitigation.

It's a bit of a.

I know interest of mine to know error
mitigation is a classical technique that

does post-processing, quantum computing.

But I thought, is there something
else you can do maybe to the

algorithm before you execute it?

Is error mitigation?

Just the classical part.

Is there more to it that I'm,
that I might not know about?

Nathan: Oh yeah, that,
that's a great question.

Thanks Steven.

So yes, that you mentioned is one kind of
error mitigation and it's probably like

one that was or it's a set of techniques
that was applied extensively in the past,

which is this one of post processing.

As you mentioned, there's more to it.

So it's not completely classical.

There is a post-processing part, which
is classical, but you can really get

more out of a out of a correction.

You can, you're able to make, if
you do a sort of quantum sampling

at the time of compilation.

So instead of having one quantum
program that you run and you get a

noisy result, which is like the current
problem of quantum computers what you

can do is that and this depends on the
technique, but generally speaking the

protocols are divided in two phases.

The first phase you take your quantum
program and depending on the technique

that you are applying and the parameters,
that quantum program is duplicated in the,

in, in multiple quantum programs, which
are all different based on compilation,

rules that are given by the technique.

They can be informed by
noise models or tomography.

And then they're run
of a quantum computer.

Then you're left with a distribution, a
statistical distribution of noisy results.

But they're all noisy in different ways.

And this difference allows you to learn
more about how noise has an impact with

respect to, on errors and how you can
correct hope, hope to correct them, at

least partially and in some regimes.

And that's why the second part,
it's fully classical and you can

actually borrow a lot of powerful
schemes from statistics, even machine

learning to do the extrapolation
to a better value and result.

Steve: Oh, okay.

That's, I didn't know that it
was that, that as deep as that.

It's very interesting.

And there's a paper on mitiq and
these features are implemented

in mitiq as well, right?

Nathan: Yes, exactly.

So we have a white paper for mitiq.

It's published in the Quantum Journal,
which uh, itself is an open source

project uh, uh, because you know,
the code is open source and it's

really a community driven journal.

And yes mitiq is a, is a live
project with now about 120 downloads

from PyPi, the Python server, and,

Steve: thousand or

Nathan: 120,000.

Yeah.

Steve: okay.

Nathan: And

Yes, and about 65 contributors.

So what's great is really a minority
of contributors come from Unitary

fund, although we maintain the
project and now we have a wider

committee of of contributors.

Steve: That's very interesting.

I didn't know.

I need to learn more about mitigation.

, especially I think later we'll talk
a little bit about networking, but

error mitigation in a network setting.

I dunno if any of, if some of these
results could could extend, perhaps.

Nathan: I think that's very interesting
and we've been thinking a bit about

it, I think it's just a little bit
of work in that, with that respect.

One of the issues there is that
generally when you have you are

sharing information over quantum
networks you really care about the

quantum state and its coherences.

And the information you get is lost
if you don't have like, if you have

like multiple shots, so to say,
you really care about the state

of that specific qubit or system.

I think it's more challenging,
but can be applied.

Steve: Yeah.

I think one of the very important
aspects, especially networking,

is the communication resources.

Computing, you have computing resources.

We're on time with the algorithm, memory.

How much memory the program takes,
how many qubits that are allocated,

but in communication setting, you
have to minimize the number of

messages between systems as well.

It slows things down a lot, so you have
to think about another optimization to

make in top of the computing aspect.

So it's a whole other axis
of optimization to perform.

It's, but I think it's, yeah, it's
a, it's gonna be important because if

you think about how quantum computing
might evolve, it's probably gonna need

multiple processors working together.

And so this communication problem,
perhaps it's going to emerge quite soon.

The communication optimization
problem, I should say.

Nathan: Yeah,

Dan: So in mitiq, there's a
calibration module that helps

you find the right optimisation.

Error correction approach.

Error mitigation approach.

Is that right?

Is that the pre or post.

Nathan: Yeah, you're correct.

And it applies to both
phases of error mitigation.

So what the calibrator does it it runs
on uh, on a set of standard benchmarks.

Some strategies, so some quantum
error mitigation techniques

with specific parameters.

And given that we have some benchmarks
that are provided by quantum programs,

specific quantum programs, quantum
circuits such as randomized benchmarking

quantum circuits, or quantum volume
quantum circuits we know the exact result.

So based on how the error mitigation
techniques are performing on your target

device, then you can pick the one that
is performing best and apply to your

problem of interest, which is not like
the benchmark with the no results.

But it'll be a new quantum circuit.

Dan: So Steve, I think what you were
saying there was a potential future

state of something like this would also
take into account the number of QPUs

involved and the topology between them.

The things like the entanglement
success rate and stuff like that to

optimise the resources that it's got.

Steve: As soon as you have to doctor
connection between devices, it's a whole

other pile of problems to deal with.

Dan: Oh, yes.

Okay.

Nathan, let's move on.

That's brilliant.

Yeah, I know, you've
mentioned mitiq and Metriq.

I was gonna say, do you have any
projects which don't have a Q in, but

maybe you could, you can answer that,
but at the same time, I'd like to know,

what would you say are the top three
contributions that the Unitary Fund has

made to quantum computing as a whole?

From the whole community?

Is it, does it include those
two or are there some other

ones you'd like to call out?

Nathan: I think it it
would include those two.

It definitely, I would say creating a
team of folks that are not in a proper

academic setting, but can focus on.

Research and technical problems in
ecosystem and hence also advance

the state of science and research
is a great value of its own.

It's a bit of a different way of
doing research that I think it's in

between industrial research and as
mentioned earlier, academic research.

So to some extent it's faster than
academic research, but, and it also

has different kind of of drive and
I would say metrics for success.

I make an example.

I think that right now in academia
there's a strong drive and focus

for, you know, publication metrics.

And at Unitary Fund sometime we find
that maybe the best way to share some

findings is to write a piece of software.

That, for example, avoids a painful
duplication of of work due to

transpilation that now you need
to make due to the, the state of,

of a current software stack where
each framework has its own rule.

But in the end, you are you just
want to run your quantum program on

multiple devices, or the best way to
present the results is in a blog post.

And you add a notebook, a Jupyter
notebook, and people can very quickly

take the data and then they can
change the data, change the technique.

And so you really have this much
faster inquiry with with ideas.

So I think this is a really
fantastic thing that we're

doing, and to some extent I hope.

That some of the work we're
doing is also influencing what

is happening at the policy level.

We're seeing more focus for support
of open source scientific software in

quantum and beyond for frameworks that
allow scientists and researchers to work

on shared data sets work on models that
are open source and can be customized

but also studied and understood.

And yeah, I hope that part of
the work we're doing is is making

this impact also the policy level.

And we're really doing quite a lot.

Things I mentioned Unitary, Hack as a
really fantastic hackathon to get started.

We have a micro grant program.

One thing that I would say brings all
this together is how many folks get

into Quantum thanks to the Unitary Fund.

And they maybe start and we see this
pattern continuously, like someone who

drops in, into our Discord calls that
are, weekly and several events on our

Discord server, which is for those who
don't know how Discord works, is, it's

a bit like Zoom, but you can, everyone
can access the call or Yeah, it's a bit

like an open teams Microsoft teams and
they, start, they start by listening in.

Then maybe they take an issue that
they can tackle and they solve it.

Maybe later on they, maybe they
participate in Unitary hack and they

contribute to several open-source projects
and then they could move on with their

own project that gets a micro grant.

And then even later on, maybe they
could get a job in the industry.

That's what they want to do.

And so I think this is really rewarding.

So one, one thing we try to to keep track
of is how many folks and which folks

really got their, where things started in
quantum thanks to Unitary Fund, and then

moved on, and our part of this ecosystem.

Steve: Yeah, it's been
a real big game changer.

I think especially, I think I
started following the unitary fund

in 2018 when I started my PhD.

That's when I first
noticed it was 2018, 2019.

No, it must have been earlier than that.

Something like that.

I don't remember.

But it was early on, right when
it was the beginning stages.

And I remember nothing like that.

There's, open source quantum
software helping the community.

It was really something different.

So what I wanted to ask is, I've
been looking at unitary fund from

the outside from a long time, but
from the inside, have you noticed a

trend, a change changing in trends?

How.

What people are interested in, what
projects have been applied for funding?

For example, I think probably two,
three years ago, a lot of the interest

was in quantum machine learning, but
now maybe it's in error mitigation

or maybe it's in quantum control.

Have you noticed a change in interest
and having noticed, I know, how

has Unitary Fund itself evolved
with over the last couple years?

Nathan: Yeah, so we, I see
a couple of trends here.

One of them is, we mentioned
these microgrants and many have

been awarded to new projects.

Or projects that were like, at, not
really just at the inception phase,

maybe it was already like a GitHub
repository, but maybe it was just a single

contributor or main developer and someone
else helping in very close contacts.

And the push was to make these projects.

And the challenge was to make this
projects be used by a wider community.

And we've seen this happen, and I
can mention a couple of projects.

I guess QNetSim is one of them
that you you designed and you were

maintaining, but there's many others.

One that really became very successful
is PyZX, which is a quantum circuit

compiler based on diagrammatic semantics.

ZX calculus for some really
exoteric math abstractions,

which are monoidal categories.

But basically what it does, it's
another way of compiling quantum

circuits to lower the overhead of
some two qubit gates, for example.

And it does this porting that the
quantum circuit to a given abstraction,

which you can think of as basically
as a graph or sort of a graph, and

then playing on this new mathematical
model to simplify this model and then

switch it back to the quantum circuit.

And this is really like we started
tracking, papers in impactful journals

and impactful papers using PyZX to advance
our the state of the art of quantum error.

Correction that's just one example.

We've seen communities like Qworld
start as a really small project and

then expanding into tens of of countries
with more than a hundred workshops

and students getting degrees also
in collaboration with universities

such as the University of Latvia.

So these are some trends that we see.

And I could also summarize it by
saying that what we've started to

steer toward to now that we have
a what we call this quantum jungle

that is more has more projects and
and tools to avoid duplication.

So not so you, you want to
avoid this atomization of, many

projects doing similar things.

So we don't need the, just
another software value kit

for writing quantum circuit.

We have many options now.

And actually you want to
simplify, unify things.

So sometimes we've gained grants to
folks just developing specific projects

that help other existing projects or
even just by opening a pull request.

And I think this has
been really fantastic.

One example comes up with QuTip.

There was like a special method to solve a
master equation based on onri of subspace.

It's a way you can decompose
matrices and it was like a

quantum quantum control problem.

And basically this became in the end,
in QuTip itself, it became an option.

In the very well-known MESolver function,
which is a master equation solver.

And it's used by millions of there's
million of downloads of QuTip, and

mostly people use this this function.

So this one example recently we've
given grants for example, to one was

really great is in the Qiskit ecosystem.

There was a project that we've given to
Harshit Gupta in India for a timeline

debugger for the Qiskit transpiler.

So you see the level of
detail is higher and higher.

But the cool thing is that now
you basically create a tool that

is like a, a command line uh, a
CLI tool that helps you debug.

The transpilation of
quantum circuits in Qiskit.

So it's super useful and it doesn't
need to be a product of its own that

doesn't speak to other projects.

So that's one trend and another trend
that we're thinking, we think looking

at, it's more at the level of really
like maybe content as you were saying.

Oh, there's like a wave of this and
wave of that is like open hardware.

Something I'm really excited about.

What does it mean to have open hardware in
quantum computing and quantum technology.

And examples of open hardware projects
historically are, Raspberry, Pi,

the Arduino project and also like
Blueprints or schemes that also

like, get close to hardware, but
maybe they're just software, like the

RISC project and so on and so forth.

So now you, you'd be surprised
to see how many projects.

Exist in quantum that have to do with
hardware and having open hardware.

And so we're seeing really
this uptick in interest.

So we recently gave two
microgrants to two projects that

have to do with open hardware.

One, it's called the Open
Quantum by Max Shokawa-Alto.

And the grant was really to develop these
CAD files, so it's like for software.

But the CAD files are used to develop
electronic schematics control firmware

that then is used to develop atom-based
quantum device that can be used for

computing or for other experiments.

And another recent project
was a grant was given to Fred.

Sorry for the pronunciation to develop lab
script QC, which then evolved in another

project called the Sqewler with a Q.

And that's something that we see a lot.

It's projects evolving into other
projects or changing course.

And this is a SDK that allow
developers to provide cloud access

to their code in a secure fashion.

So it started at something really
specific to cold atoms and then it

evolved into something more general.

So this is something we see a lot.

Steve: Yeah, this is such a wide variety.

And then I guess what the
natural extension would be like,

what do you think is missing?

What would be, something that if someone
applied tomorrow, it'd be like, yes,

this is finally an answer to my call.

Is it, is there something there that you
think would make a big impact right now?

Or a piece of the pie that's missing?

Nathan: That's a good question.

So one thing where definitely
we don't have too much software

or tools or projects, maybe even
hardware beyond quantum computing.

So if you go to quantum
sensing to quantum networks.

It's limited what you have at hand.

And there's some great project
like QNetSim and InterlinQ for

which you were an ORD or mentor.

are great, but I think there's room
for more open hardware is something I

mentioned that I think it's still there's
a lot of room for expansion, I think

in the quantum error correction field
to develop algorithms and also at the

interface of, quantum error correction.

Not yet in the full tolerance regime,
but, deploying these algorithms to current

devices, I think where there's space to
develop tools and software platforms.

Recently we, I.

We've awarded now it's already a
couple of years ago, but we awarded

a grant to a team of of students and
researchers for lattice surgery based

quantum error correction compiler, which
then led to a, this was in Canada and

this led to a very cool project that
you can look up at lattisurgery.com,

which now has grown a lot in scope.

But yeah, I think there is a lot
to do in quantum error correction.

Definitely we need benchmarks.

So these are, I will, I will
draw attention to Metriq.

We can probably increase
the APIs we have for Metriq.

We have an API for an automated
pipeline to run some standard

algorithm to see how far we are.

From applications with quantum computers.

So let's say that you want to
solve the max-cut problem with

a variational quantum algorithm.

So the there folks from the Quantum
Economic Development, Consortium

QDC, who have very set of quantum
algorithms, and we created an automated

pipeline to, to run these benchmarks
on hardware backend and upload the

results on the Metriq platform.

So that may show up as data
points in these charts.

I think there's more need for these
application-oriented benchmarks.

Dan: Nathan, whilst yor there?

You're talking about Metriq.

I wonder if you could explain that in a
bit more detail to a layman like myself.

So I understand it's a
benchmarking platform of some

kind, but does that mean that.

It's connected to some quantum computers
or is it to store records of benchmarking?

In which case, how do you, how
can you be sure that they are

fair among the different systems?

Nathan: Yeah so it is not directly
connected to a hardware platform,

but you have the possibility
to connect it through its API.

So what we did, for example, for the QEDC
benchmarks was to add to the Metriq API

the capability to integrate with this
set of uh, of, uh, of algorithms that

then can be run on the specific machines.

So Metriq is uh, uh, independent on,
on the backend, but I think it would

be a great, to expand the Metriq,
to be more fully integrated, even

maybe at the CI level, continuous
integration level with some backends.

So there's partner, potential
partners listening in, if you have

a quantum computer that you would
like to connect to Metriq, reach out.

Talking about the results, this
is a question we get a lot.

Oh, how can we trust the
result that we have there?

Because right now anyone who registers
on Metriq can can submit a new result.

And two things here.

One, we have comments on each page.

So there is a level of control
from the community itself.

You can report a problem to the Metriq
team and we do moderate the results.

So they are.

Accepted and reviewed
all what we want to do.

We want to make the mechanism even more
Trustworthy, and we actually have an

open call for the Open Quantum Benchmark
Committee ending February Twenty-third.

So I don't know if these Podcasts
will be out by then, but we already

got some amazing application from
candidates, and we really want to create

a lightweight committee that really
helps in part review Contributions,

but also helps with the taxonomy
and the presentation side of things.

There's a lot of data visualization
choices that you need to make

and we're adding some of these
options to to the website.

So one chart that I think is really
insightful is a progress chart, which

we also call what we need and what we
have, which shows you on a logarithmic

scale vertically qubits and horizontally
logical operation and shows results

or estimated calculations from the
literature for various application

of quantum computing algorithms.

So they go from quantum dynamics
to factoring to derivative pricing

in in finance to quantum chemistry.

And that really helps uh, uh, yeah,
to some extent demystify the field

because you start to understand what's
the pace of progress and what are the

gaps for respect to specific, use cases
and applications in the real world.

Dan: That's fascinating.

I'm even having a look now and
seeing all the different modalities

that are on here and all the
different tests which have been run.

There's a huge amount of effort and
input that's been put into this, which

can only be achieved through a community
really, I think by the looks of it.

Nathan: Yeah.

And we have every Friday we have the
Metriq jam session as we do for many

other projects that we host on our server.

And everyone can tune in on Friday
at 3:30 PM central European time.

Dan: I dunno if we've covered
it already, in, in my list I

wanted to talk about multi-QPU.

By that I mean applications or
algorithms which have been developed

to run on multiple quantum computing,
either systems or processes.

Now, okay, this is at the moment,
unless you work for IBM, it's

probably a simulation thing for you.

But are you seeing any of the
contributions or the projects.

Coming through Unitary fund or in the
ecosystem, which are looking at that.

We mentioned earlier on optimising
algorithms to run the right type of

error correction, error mitigation
and the potential of building in the

topology into those calculations as well.

What are your thoughts on that?

Or is it too early?

Nathan: I've seen mostly
research on, on, on this topic.

So tangentially is also something
that we we investigated.

There is a paper co-authored by
Misty Wahl is the first author.

It's with other folks at the Unitary Fund,
Ander, Amari William Zeng, myself, and

collaborator from university of Chicago,
Gokul Ravi at the time of the University

of Chicago now University of Michigan.

And in that work we, we are porting
error mitigation, a specific

error mitigation technique to the
Quantum error Correction framework.

And the way you do it is I need to
give a bit of context to say what

we did with due respect to this
multiple cores to some extent.

So what we did is you have an
overhead in the qubit number and the

quality of these qubits in order to
run Quantum error Correction codes.

But you can use error mitigation
to sort of like run approximated

or smaller scale codes.

Quantum error correction codes and
then extrapolate what would be the

result if it run on a larger error
correction code involving more qubits.

And you can really think of a
quantum processor as a 2D grid,

which at each point you have a qubit.

And the way these, that's the way you
can also think of quantum codes such as

the surface code, which has an extension
on these quantum processing unit.

What we did there is to some extent it's
not really like a dual core or multi

core like QPU, but it's introducing
this idea of virtual cores such that

with the same quantum processing unit.

That is, I don't know say 20 by 20 matrix.

What you can do, you can reuse
qubits by running, let's say,

multiple multiple patches on your
grid that refer to different for

example, surface co surface code.

So instead of involving 20 by 20
surface code, it is something like,

five by five and you have many of
them that can run at the same time.

And then you can post-process the results
to then extrapolate what will happen

for a 10 by 10 and so on and so forth.

So this helps especially with
respect to the overhead that

you have in the number of shots.

'cause these quantum error mitigation
techniques I mentioned, may allow

you to use just, a reduced number of
qubits, of the same number of qubits

that you have at, at, at your disposal.

But then they have an overhead in how many
repetitions you need to run your code.

And they're also called shots.

And so if, if you can do this, you
can drastically reduce the number

of codes at least quadratically.

So that's one thing where I've seen and
we did internally this this experiment

or propose this really framework for
quantum error correction hybridized

with quantum error mitigation.

I cannot think of specifically
other projects that have

this multi-QPU approach.

I know in the community
there is these approaches.

I think that especially
some experimental platforms.

Oxford, for example, are
investigating how, so to say smaller

scale QPUs connected by quantum
networks can work experimentally.

But is that something I've
seen much more closely.

Dan: I think it really varies
platform by platform, right?

If you think about the Rydberg arrays
that the neutral atom computers out there

that are, they're pushing the technology
there by creating these zones of qubits

that they're moving around inside the
lattice to interact with each other.

And that's almost like, um, I dunno, it
is not like a separate core, but it's

like you're re-tuning the positioning
of all of the qubits in order to

create some interactions at some point
which is very different to something

like the fixed grids that you're
talking about, that you get in an ion

trap or a superconducting platform..

Nathan: Yeah.

Now, you have even shuttling
of ions these on some devices.

So even thinking of

Dan: shuttling.

That's right.

Nathan: yeah, thinking of of a quantum
processing unit at something static,

it's it's really not uh, you need to
be more flexible than that, in some

systems it is so to say, hard coded.

And it's a monolithic piece in
some uh, uh, in some material.

Could be a device with a.

Superconducting elements or something
that is the most compatible.

As for spin qubits, in some other
cases, as you were mentioning,

like atoms that are trapped by by
lasers you have the flexibility.

Every time you run a quantum program,
you can potentially rearrange your

grid in a way that is not even
like a fixed 3D grid, but it can

be like a different kind of mesh.

So of course at Pasqal we did the Tour
Eiffel 3D rendering, because they're not,

yeah, it's pretty amazing with the level
of control they get on the single atoms.

And and then of course, there are
challenges on how you can you can

implement two qubit gates, but those are
like fantastic quantum simulators where

yeah you could, for example, when you
were speaking, I was thinking about how

when you have in some quantum materials.

You have different zones that
start to interact with each other.

These zones are really like even in
physical space, but you can look at

them in some other like complementary
phase space and you can really study

some dynamics that are self-emerging and
are also due to some of your driving.

But then the kick-in of process
such as chaos or correlated

mechanisms such as super radials.

So yeah, I think it's really
interesting to think in this time.

And then, then we were mentioning
that the ions were one of approaches

to try to enable these long distance
entangling with electromagnetic fields.

But another approach is to
always use this electromagnetic

fields, but shuttle every time.

The two ions that need to be entangled.

And so you have also like the mechanical
degree of freedom that is controlled by,

by some electrical degree of freedom.

And so it's really fascinating to
see the the challenges that we have

at engineering level to develop
these uh, quantum processing units,

but a lot of creativity there.

Dan: Yeah.

The innovation that's happening with all
these different modalities is uh, yeah.

The investment going in to that
kind of innovation's really

pretty amazing at the moment.

Anyway, coming back to algorithms,
coming back to stuff in software.

I wanted to ask you this one, and
maybe this is this is my naive

um, noob question, but from what
I understand, shor's algorithm.

Grover's algorithm,
they're the two algorithms.

They're the flagship quantum
algorithms that we know.

If they can be run, have a a, an
advantage over being, trying to

run them on a classical system.

The more I hear about these hardware
improvements that we're talking

about, and the more I hear about the
ability to provide logical qubits

with a much higher fidelity and less
decoherence in the system, the more

I think about the algorithms being
the next point of the next thing,

the, where there's a gap, right?

As the hardware gets better and better,
ultimately the hardware's gonna be, some

of the hardware platforms are gonna be
ready to take on more advanced algorithms.

What are you are seeing inside the
Unitary fund around tackling this gap?

Is that algorithm development
happening in any of the projects?

I appreciate that quite a lot.

A lot of the projects you've
mentioned today, and thanks for that

already are things which algorithm
developers could use as part of

their set system of tools and things.

But yeah, is there an overlap
there between the Unitary fund

and algorithm development?

Nathan: Yes.

I'm not sure.

I don't know if how can I relate
it to your the first part of your

comment about a Shor and Grover.

But I would say you also meant
like we need more algorithms

and we need someone to come

Dan: That's basically it.

Yeah.

There's a community communities
in the Unitary fund.

Is there a, an algorithm community
inside it that's working on algorithms?

Nathan: Yeah.

So we, I would say we are developing
algorithms is quite challenging.

And you need uh, yeah, I mean after
so many years we have just, a couple

of frameworks at least, right?

You have the VQA many flavors,
but they all relate to QAOA

or VQE, a lot of acronyms, but
these are relational approaches.

And then as you said, there is
Shor, there is, there is there is

Grover and there's a couple more.

And actually, one thing that we did we do
support is actually not only benchmarking

and a taxonomy of benchmark, but even
a taxonomy of existing algorithms.

There is a project that is every
year lately in that Unitary Hack,

which is called the Error Correction
Zoo, which is developed by some

researchers at Caltech and other places.

Victor Albert is one of the maintainers.

And there you see, really see like a
taxonomy of these error correction codes.

Dan: I've gotta say that.

Is that, does that come from
the Algorithm Zoo, which I think

was the original online zoo.

And I also, earlier on you also
mentioned the Quantum Jungle of

different things in the Unitary Fund.

Oh, I'm sensing a trend
here, a theme, some kind.

Nathan: Yeah.

Yeah.

So it's actually a different
project from that one.

But it's it's really this one, it's
really specific to all these different

taxonomy of quantum error correction
codes, so LDPC codes and other codes.

So it's it's, and it has
now a community of its own.

I can maybe receive some material.

We can share the link, but it's really
fascinating respect to respect to

what we're doing at Unitary Fund.

I would say we actually proposed,
you can call them algorithms.

You can call them more like maybe
techniques for error mitigation.

We're not really like the algorithm in
the sense of obtaining a calculation

that has a an application per se, but
it's more like a technique that allows

you to perform other calculations.

Yeah I need to think about
specific projects where

algorithms are at the center.

We, that we've added some, but
generally it's some approaches

to quantum compilation.

A recent one is approach an approach
to quantum compilation using genetic

uh, genetic approaches, genetic
algorithms that come from like,

you know, the classical field or
Casper is a recent project given

to Michael Edmund in Singapore.

Quantum representations of standard
financial notation and classical assets.

So yeah, would have to think about it.

Dan: Another thought maybe
to you as well, Steve.

Who are the types of individuals
that develop these algorithms?

Are they mathematicians or
computer scientists, or is it

normally a combination of the two?

Steve: Yeah.

From my perspective it's deep theory.

It's, you have to understand the
linear algebra very deeply to

invent a new quantum algorithm.

And then it's about the error
mitigation, all the other techniques

come after, but just knowing in pen
and paper how to solve a problem

with the quantum algorithm, I think

Dan: It's a conceptual thing, isn't it?

Steve: yeah, you have to understand the
mathematics very deeply, but to have,

especially to have an advantage, you
can do things potentially, you could

just encode everything classically
and then use classical manipulation,

but if you want to extract the power
of the quantum computer using the

properties of quantum mechanics, it's,
it takes quite some skill, I would say.

I don't think I, I would even be capable
of doing it through all my education

in academia and all this stuff.

I still think it's inventing a new
quantum algorithm from scratch to

do something particularly special.

Yeah.

That's why most of these
algorithms have names on them.

Shor's algorithm, because, there's
very unique HHL as an acronym for the

people who invented the algorithms.

There's the, their names on them
Grover because they're so spectacular.

Dan: Yeah, Shors is interesting
'cause it's got a whole bunch of

different features in it, like
the quantum Fourier transform

and some Eigen-solving and so on.

And did he, didn't he actually
come across it by accident?

He was trying to solve some other problem
and then ended up factoring primes.

He was, I can't remember what the original
thing was that he was solving, but, um,

I mean For so somebody of that level
of brilliance to having to accidentally

come across it through some of the work.

It's not just the
brilliance that's needed.

It's also a bit of luck

Steve: Yeah.

So like small mistakes.

Yeah.

I think finding applications for your
work, it just, you attend a conference

or something, someone says something,
it triggers something in your head

and the stuff you're working on,
you say, I can apply this to this

problem with the small modification.

Dan: So Nathan.

Let's go back to the Unitary Fund.

Thanks for all of that insight.

What's on the roadmap for you?

What's coming up?

Maybe you know what's in store for 2024.

Nathan: Yeah, thanks.

Thanks for your question.

We have some of the
programs we run nearly.

Of course, we want to execute
also this year and, grow in,

in, in impact and activities.

So definitely look out this spring
for Unitary Hack 2024 and later on

in the fall we want to run even far
reaching, further reaching open source

survey, quantum open source survey.

Mitiq is really at the level, I
mean it has now a sizable community

of contributors and users.

And what's next?

I think it's two things.

It's integrating more with noise
characterization tools and capabilities

to empower even better performing
error mitigation techniques.

And at the same, and to do this also
to do this, but I would say even more

in general for, to have a better flow
from a user perspective is to integrate

Mitiq closer to the metal, so to say.

So next to hardware.

So again, it was like,
quantum hardware, pro provider

listening, tuning in reach out.

And we're actually performing as part
of a National Science Foundation grant

for the growth of a Mitiq ecosystem.

The team just marked, about the first 50
interviews with the community in the of

users, but also researchers in the space.

And we're getting like
some great insights.

So we really look forward to put,
to work these these findings.

On the partnership level.

A couple of years ago there weren't many
non-profit or collaborative initiatives.

You had like academic research when
industrial research centers from

May, maybe some big corporations
that have been investing for long

in quantum computing research.

And then startups starting to pop up.

Now we're seeing like a more complex
and elaborate ecosystem, which I think

it's beneficial for tech transfer.

So we're looking forward to
work more and more with these

sort of like local environments.

There's been some that have been
really successful in development.

One is in France where
recently we opened up a shop.

So we have inaugurated last
year, Unitary Fund France.

And that's really like a thriving
and an ecosystem that is scaling

up with some major startups, but
a lot of research and new emerging

startups and government involved.

And so various actors o other nonprofits
like Le lab Quantique LA French tech.

So it's really important to have these
sort of open ecosystems nowadays.

The Netherlands is another
environment where it's been the

quantum delta innovation hub.

So to say that it's really helping
start-ups and spin-offs from Delft and

Amsterdam um, you know, work with each
other, understand what are like the their

strengths and how they can integrate.

Some merging of start-ups and numbers
that women in quantum development,

non-profit that emerge in that ecosystem.

So we're looking forward to work
with many of these in in, in the uk.

Where's large large government support
the center doctoral training centres.

So we're looking forward to, to partner
with realities there, the University

of Southampton and other new doctoral
training centers emerging there in Italy.

We are helping establish Quantum
Italia, which is an initiative from

one of our sponsors, scientific
Adventure Capital, and really

helping create an open ecosystem.

So this like a venture capital fund
really focused on, on helping scientists

bring their research out of a lab and
start to think more startup oriented way.

But very, we're a bit unique because they
provide fellowships to data scientists

and scientists they give awards to thesis.

They have open calls.

So this this environment
and also, these last year.

November, 2023, we hosted our first
in-person meeting UnitaryCon in Rome.

Thanks for our support.

So we look forward to helping them
more give member understanding that you

also need to fund things that don't go
directly into a startup, that you need

an ecosystem of doers and developers.

I did mention now UnitaryCon, and
this was our first in-person meeting.

And it was just amazing.

I mean, We got really um, fantastic
reviews from the community of developers

and Micrograntee awards that joined.

It was a sort of like closely-knit
atmosphere, about 40 people, but the

level of interaction and, finally meeting
in person, that and that maintainer or

developer, it was just fantastic together
with, some of our advisory board members,

which I should really thank because they
really helped Unitary Fund thrive and

review the application of the Micrograds.

Yes.

You could see like, you know, uh, many of
these, of these folks interact with each

other and that, that was just fantastic.

There was Patricia Agarwal from from
India who started this quantum machine

learning textbook interacting with Nathan
Killoran, who is a software CTO of Xanadu

Startup and Spencer Churchill, who wrote
these quantum tales, which are these

tales through which he's explaining.

A very specific technical, I mean,
it's very technical, but you know,

it's, it's, it's a tale that you
can read specific algorithm or, or

settings in uh, the quantum world.

And who now has got a is now
working at IonQ, one of the

prominent startups in the field.

So it's really an exciting environment.

And next year, this year we, we
look forward to, to have, again

this event and grow it further.

So I think these are the main
things that are on my radar.

Dan: The event, the face-to-face
event sounds like a real milestone

and I'm already imagining.

Something much bigger.

Think five in five years time.

Think about the open source community
or convening on a conference center

somewhere, all the sponsorship
and all the, people involved.

Could be fantastic.

It's I'm looking into my crystal
ball and that's what I see Nathan.

Nathan: Yeah, we just had this
discussion with the board.

Should we go into the direction of
making it start to working right

now to make it bigger and bigger?

Should we just first, nurture this
community and keep it a bit close knit?

So yeah, there's different
directions you can go.

But yeah, of course we call
it UnitaryCon because, the

Python ecosystem is this PyCon.

You have different levels like
PyData and EurosciPy, but this

idea of this conference, it's it's
it's really something we'd like

to see grow and in this direction.

Definitely.

Partner more and create bridges with with
existing nonprofits in the open source

space and as well as in the quantum space.

The 5th of March was gonna be the
launch of official launch of the Open

Quantum Institute managed by CERN and
hosted by the Jasda Institute in Geneva.

We look forward to partnering with
them, there's also like some some

big players in, in, in the open
source space, such as the Linux

Foundation is one that comes to mind.

NumFOCUS, I'm a mentor for
NumFOCUS for the QuTip project

and Google Summer of Code.

There is a center for open science.

There was also award of the
national Science Foundation

grant to grow their ecosystem.

So I really look forward to further
partnership and interactions at the

time at which really feel like Unitary
Fund has established itself as some call

them research-focused organizations.

So it's new kind of organization
where you can do many things that

usually could be thought of or done
only in a, in an academic environment.

So we don't do just academic stuff.

We also do more like pop fun stuff.

But yeah, there's this research
part is always at core,

Dan: Very nice.

Yeah I did read about the
Open in Quantum Institute.

That sounds like those guys are gonna
have a similar kind of vision and goal

as you in terms of working towards
open software, open architectures.

Nathan: I was just passing saying
it, but I should really emphasize it.

A lot of what the Unitary Fund does.

It, thanks, to the co
community of contributors, the

volunteers the technical staff.

And I also want to thank our members.

So we have some core members
and supporting members.

You can find them all cited at
unitary.fund 2023 members are like

IBM, Q, IBM Quantum, Scientifica,
Agnostiq, AWS, Cisco, DoraHacks,

Pasqal, Candela, Koval, and Kyber.

And we are about to
announce our 2024 members.

So I'm really excited.

I really look forward to the growth of
this cohort of members and supporters

this year in the upcoming years.

Dan: Yeah.

Hey, at the heart of it's the people.

Sure.

Um, and, And let's, let's do that.

You've called out some of the
members there and some of the

contributors advisors and so on.

What about the actual people
that are contributing to code?

Are there any key people
you wanna call out?

Any heroes of the community that
you think you'd like to mention?

And I'm sorry we've left it till the
end, but let's get it in there anyway.

I'm sure there are a few and if you forget
one or two, I'm sure they'll forgive you.

Nathan: Yeah, it, it's very hard to,
it's very hard to single out some,

really here, back over Steven, he
started as an awardee for his project

and I moved on to be a mentor.

And now and this year we worked also, he.

Working at Cisco and helping bridge
the partnership and the support

from Cisco to the Unitary fund.

And actually, we seen, and he was
also a mentor for for the Interlink

Q plugin of a QNetSim, which
was the original Quantum Network

simulator that he built while at TUM.

It's hard to name them, I
mean to, to single out one.

I really think yeah, you can you can look
at the map on the Unitary.Fund website.

I maybe they come to my, some of the
latest additions include Amir Abbas and

Sonica Iori and yeah, they've really been
fantastic in mentoring and evaluating some

of the microgrant application we've had.

And I would really like to thank
board members that have have

really helped since the beginning.

One of them is Travis Schulten
who was like involved with the

Unitary Fund even before I, I
joined, so before, previous to 2020.

So these are some folks that come to mind.

And then again, really many of our
team members, Unitary fund technical

staff that, were working full-time
on, on these community projects.

They started as
contributors and volunteers.

So there's a, Misty Wahl who started
as a Unitary fund ambassador.

First thing that she did
with Unitary Fund or.

Dan Strano, who developed the Q-Rack
Simulator, which is a highly enhanced GPU

simulator for quantum computers that was
beating the state of the art of simulators

from Google and IBM at some point.

And I dunno, now you can see who's
leading, who's not on Metriq.info, but

he's now a full-time developer for Mitiq
and engineer for many things to do.

But yeah, no, no offense to
those I didn't cite because.

Dan: No, I'm, no, it's great.

Thank you for that.

I know it takes a village
I, your organization is the

real epitome of that phrase.

It takes a village 'cause.

There's so many different
people that are contributing.

I know you've mentioned a few
there, and I'm sure it's only

a small percentage overall.

Thank you for that.

Let's wrap up.

I just wanna say thanks again
for coming to talk to us, Nathan.

It's always exciting to talk to you.

I look forward to the next one.

I'd like to take this moment to thank
you for listening to the podcast.

Quantum networking is such a broad domain
especially considering the breadth of

quantum physics and quantum computing all
as an undercurrent easily to get sucked

into So much is still in the research
realm which can make it really tough for

a curious IT guy to know where to start.

So hit subscribe or follow me on your
podcast platform and I'll do my best

to bring you more prevalent topics
in the world of quantum networking.

Spread the word.

It would really help us out.