This presentation features QC2, an open-source software that seamlessly integrates computational chemistry codes with quantum computing frameworks. Designed for hybrid workflows, QC2 relies on ASE calculators to efficiently offload electron integrals for various quantum computing SDKs.

- Carlos Boschen is a research software engineer at the Netherlands ESI center. In this talk, he will give you an overview about QC two, a modular open source software for quantum chemistry with quantum computers.
- Atoms and molecules are indeed true manifestation of quantum offsets. They show discrete characterized by discrete energy levels. This can be naturally described by solving the corresponding Schrodinger equation. Among the most popular strategies to perform these approximate solutions in quantum chemistry is by researching through the so called variational principle.
- QC two uses popular quantum computing libraries like Qskit, Nature and Benellan Qgan. QC two is also designed to be highly modular and features are built in algorithm package. It's also very user friendly and has a very intuitive interface.

Hi everyone, I'm really glad to be here. I'm Carlos Boschen.
I'm a research software engineer at the Netherlands ESI center.
In this talk, I will give you an overview about QC
two, a modular open source software for quantum chemistry
with quantum computers. Before going on,
I will be giving you an overview about my organization.
The Netherlands Escience center is the center of spurges
in research software in the Netherlands and we
are founded by the dutch research Consumpt and Discernment.
The Netherlands Design center is a center around two
main ambitions, and the first one, we work
with researchers collaboratively to develop open
source research software, and this is done
via open call or even external projects.
In the second ambition, we build and seek to provide
digital expertise for researchers in the
Netherlands. And for this we also provide workshops
hanging from basic to intermediate python, good practices
in software engineered and also more involved topics like
the CudA program and QC two is
a direct outcome of one of such projects,
QC for QC, which is collaborative effort between
the Netherlands Design center surf and a full Amsterdam.
The main objective of this project was to develop a
dedicated software that I'll be reviewing here as UC
two to facilitate the interoperability between
traditional quantum chemistry codes and byted
quantum computing libraries, essentially to facilitate
users to perform quantum chemistry calculations using quantum
computers. Quantum chemistry may be abused in various contexts
and show several bases as well.
So there are researchers that are interested in you study
a very big systems, for example like the folding octane.
And for this system, because they are so big, the researchers
have to rely on very push made
mathematical or physical models, for example,
relying on classical mechanics. For other medium
sized systems, we can indeed apply some
formally, physically, mathematically sound quantum
mechanical or chemistry methods like the
density functional. But it is for
these molex systems that we so far can apply
or test the most accurate state of the art
quantum chemistry and quantum mechanical models. And these are
the ones that UC two is currently dealing with
and that are focused during this presentation.
Atoms and molecules are indeed true manifestation
of quantum offsets and they show discrete
characterized by discrete energy levels that
are actually the discrete spectrum of
the corresponding molecular hamiltonian.
And this can be naturally described by solving the corresponding Schrodinger
equation that I show here in a very simplified form.
So as you can see here, this is an eigenvalue problem where
a is the molecular hamiltonian, c is the corresponding
wave function of your system that completely describe
describes your target molecule, and e is associated
discrete energy associated with this wave
function. Unfortunately, this equation can only be solved
exactly for one electron systems, for example,
like the hydrogen atom, and for all other atoms
and molecules. We must rely therefore on approximate
mathematical and physical models that seem only to
approximate the exact solution of the Schrodinger
equation. Perhaps among the most popular strategies
to perform these approximate solutions in quantum
chemistry, and perhaps also in quantum mechanics, is by researching
through the so called variational principle. And here,
the unknown wave function of your system is approximately
given as a linear combination of
known basis functions, and the corresponding coefficients
are then obtained in a variational manner
in such a way to minimize as much as possible the
energy or the ground state or most stable energy of the system.
Essentially, when they are resorting to the quantum chemistry
method based on the variational principle,
your classical computers are essentially trying
to diagonalize very large, very rude matrices
and in such a way as to obtain a diagonal matrix
containing the corresponding eigenvalue energy
values of your ground state or even excited state systems.
This process consists certainly the computational bottom
end of classical computers. So, just to give you here
some idea, one of the prime or
most accurate methodologies to perform high degron
state or even side state energies, show exponential
complexity in relation to the number of
the electrons containing your target system.
Okay, now that you have overhauled idea on how we approach quantum
chemistry using classical computer,
let's try to then approach start approaching how
we use then quantum chemistry and quantum
computing to also perform oplasmate solutions
through this Schrodinger equation. In the noisy intermediate
scale quantum theorem, quantum chemistry is traditionally
approached, and approximate solutions
to this Schrodinger equation are traditionally approached
using hybrid quantum classical algorithms
like the variational quantum eigen solver, in which
QC two and several other Python quantum computing libraries
are primarily based on.
As you may know, in this algorithm, we start with a set
of qubits in specific or
initial states, and to this we apply a
variational form that contains a series of quantum
gates that depend on certain rotational
qubit rotation parameters. If this parameterized quantum
cpu and associated appropriate form of
the molecular amygdala, we are then able to measure the
spectation values, or in simple words, the energy of
the molecular system that depends on the
circuit parameters. This is obviously
a cost function that we can be optimized.
We're sorting to, or using an external classical optimizer
that will then guess the best meters which
to evaluate the circuit. So this process continues
on and on until we obtain via convergence,
the lowest possible energy of your system. I'm not
going deep into details, but I just would like to let
you know that UC two and any VQE algorithm need
some initial information to start with
and when we are talking about quantum chemistry, this initial
information can only be done by running or by resorting
before that traditional quantum chemistry
goals. This will supply information like
those needed, for example the buildup of the reference
quantum circuit, and also some physical information
that will be later on or physical quantities that will
be later on be used to build up
the molecular amygdalin in qubit representation.
It's obvious that we need to provide a proper variational form
or sats that will be used
and in compunction chemistry it is very traditional
to use physically motivated and satsy like
those derived from dario couplet.
Just to give you some overview here of the core design principles
of QC two. QC two has been primarily designed
to leverage existing tools and are as I'll be discussing
later on the next slide, it uses very
popular quantum chemistry python libraries like atomic
simulation environment calculators and also standard
data schemas to offload the necessary
information. Initial information to build that deep start
the quantum algorithm. Also, the CGU has
decided to smoothly integrate with popular quantum
computing libraries like Qskit, Nature and Benellan
Qgan. QC two is also designed to be
highly modular and features are built in algorithm
package that is designed also to make
it easier any extension and advancements
individually. It's also very user friendly and
has a very intuitive interface. As I'll be discussing
later on, that enables users to
focus on that research, obviously with minimal technical
things. Here just give you an overview also of the workflow
of QC two, as you can see here. For example, QC two
uses Python bindings or custom
atomic simulation environment calculators that run
on the back behind the curtains your previewed
quantum gems three program. This will then
the SCE will then get the relevant information and
save this into formatted data files that will
be later on read by QC two. For example
to build up Erb zeros, quantum circuit
Er, molecular hamiltonian incubate representation,
and also instantiate algorithm classes
that will then be run using Qscape nature and penny lane
QK. Here I just give you a very minimal
minimal input of QC two to perform
a CQE calculation on water.
So as you can see here, after importing some important modules,
the key aspect of QC two here starts with
of its QC two data class that
will contain information about the target system that
we want to deal with and also provide a
file name in which will be all the output of the calculation
will be saved and might be also used to restart
a new calculation the second step here is
to specify and run an appropriate custom QC
two is ase calculation calculator for
your refuted quantum camp SQL. Here I'm
just using the one from Byscf.
Here we need to instantiate the corresponding algorithm class
that will also be run afterwards. And that's
it. Very simple and very straightforward.
Here also I show another input which now uses
or performs an ortho optimized VQE
calculation, which is extension of VQE
in which we are not only here optimizing
the circuit parameters, but also allowing the initial molecular
orbitals to be relaxing throughout
the calculation. And as you can see here,
the structure of the boot is exactly the same. The only difference
here that we are now instantiating corresponding
o VQE class. And this certainly reflects
the modularity of juice in which we can play
around with some blocks or some pieces and
maintain the same structure of the input.
Finally, I just show here some representative benchmark results
using these qc two input examples.
And here I show the ground state energy convergence
of water in its equilibrium structure as obtained
with qc two using CQE algorithm and
ortho optimized VQe or oofqe
using also qcg. Also here I show the expected
convergent energies of ground state water as
obtained using traditional quantum chemistry methods.
And as you can clearly see here, the OVQE energy
is financially well below the corresponding VQE
as expected, right? And converges quite
nicely and is nearly exactly to the traditional
energy obtained using gas SES.
So just, just a final remark. QC two
is an ever growing open source project that is constantly
being enriched by new algorithms and new features.
And as an open source software we are warmly welcome
any type of contribution, either if
it's uncommon, new ideas, new approaches,
or even folks. And here I put the link,
full link of QC two GitHub free to go and
visit and make your contribution. And thank
you very much for inviting me and I wish you all
an amazing Confi 42 quantum computing
congressional.

Priority access to all content

Video hallway track

Community chat

Exclusive promotions and giveaways