BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Nana Liu (Shanghai Jiao Tong University\, China)
DTSTART;VALUE=DATE-TIME:20200522T000000Z
DTEND;VALUE=DATE-TIME:20200522T010000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/1
DESCRIPTION:Title: Introducing Adversarial Quantum Learning: Security and
machine learning on the quantum internet\nby Nana Liu (Shanghai Jiao Tong
University\, China) as part of Centre for Quantum Software and Information
Seminar Series\n\nAbstract: TBA\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Nathan Wiebe (Pacific Northwest National Labs\, University of Wash
ington)
DTSTART;VALUE=DATE-TIME:20200529T000000Z
DTEND;VALUE=DATE-TIME:20200529T010000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/2
DESCRIPTION:Title: Training fully quantum Boltzmann machines\nby Nathan Wi
ebe (Pacific Northwest National Labs\, University of Washington) as part o
f Centre for Quantum Software and Information Seminar Series\n\n\nAbstract
\nIn recent years quantum machine learning has grown by leaps and bounds b
ut a major problem still vexes the field is how to efficiently train quant
um neural networks. This is particularly challenging because of the lack
of a natural backpropagation algorithm for updating the quantum model. \nI
n this talk\, I will focus on an approach that can mitigate this problem t
hrough generative training. We will show how to construct a fully quantum
model of a Boltzmann machine and train all of the parameters of that mode
l for both the quantum and classical parameters in the model. In contrast
\, existing methods were not able to achieve this. \nIn particular\, we wi
ll show explicit query upper bounds for the cost of simulation\, provide a
formal proof for BQP-completeness for evaluating such neural networks and
also discuss remaining problems in the field and how to generalize the id
eas presented here to go beyond Boltzmann machines to allow efficient trai
ning of broad classes of quantum neural networks.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Kai-Min Chung (Institute of Information Science\, Acedemia Sinica\
, Taiwan)
DTSTART;VALUE=DATE-TIME:20200602T010000Z
DTEND;VALUE=DATE-TIME:20200602T020000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/3
DESCRIPTION:Title: How well can a classical client delegate quantum comput
ation?\nby Kai-Min Chung (Institute of Information Science\, Acedemia Sini
ca\, Taiwan) as part of Centre for Quantum Software and Information Semina
r Series\n\n\nAbstract\nIn a recent breakthrough\, Mahadev (FOCS 2018) con
structed the first classical verification of quantum computation (CVQC) pr
otocol that allows a classical client to delegate the computation of a BQP
language (i.e.\, a decision problem) to an efficient quantum server.\n\nI
n this talk\, we present several generalizations of Mahadev’s work. In p
articular\, we initiate the study of CVQC protocols for quantum *sampling*
problems and construct the first such protocol that allows a classical cl
ient to verifiably obtain a sample drawn from a quantum computation from a
quantum server. We also construct the first protocol with efficient verif
ication\, i.e.\, the client’s runtime can be sublinear in the quantum ti
me complexity of the delegated computation. Finally\, we present a generic
compiler that compiles any CVQC protocol to achieve blindness\, i.e.\, th
e server learns nothing about the client’s input\, which leads to the fi
rst constant-round blind CVQC protocol.\n\nBased on joint works with Nai-H
ui Chia\, Takashi Yamakawa\, Yi Lee\, Han-Husan Lin\, and Xiaodi Wu\n\nHos
ted by Prof Zhengfeng Ji\, UTS Centre for Quantum Software and Information
.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Maria Schuld (Xanadu\, Toronto\, Canada)
DTSTART;VALUE=DATE-TIME:20200605T040000Z
DTEND;VALUE=DATE-TIME:20200605T050000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/4
DESCRIPTION:Title: Encoding Classical Data into Quantum States for Machine
Learning\nby Maria Schuld (Xanadu\, Toronto\, Canada) as part of Centre f
or Quantum Software and Information Seminar Series\n\n\nAbstract\nWhen qua
ntum computers are used to process classical data - a setting investigated
in the emerging field of quantum machine learning - the first step is to
encode data into quantum states. In fact\, this is the most important step
: the way we encode classical data determines almost entirely the potentia
l power of a quantum machine learning algorithm.\nThis talk sheds light on
different aspects of this data encoding\, from claims of exponential spee
dups to quantum feature maps and quantum kernel methods.\nIn particular\,
it will present the framework of quantum embeddings in which a data encodi
ng can be adaptively learnt from data\, while the circuit for optimal clas
sification follows from well-known results in quantum information theory.\
n\nHosted by A/Prof Chris Ferrie\, UTS Centre for Quantum Software and Inf
ormation.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Robin Blume-Kohout (Sandia National Laboratories\, Albuquerque\, N
ew Mexico)
DTSTART;VALUE=DATE-TIME:20200612T000000Z
DTEND;VALUE=DATE-TIME:20200612T003000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/5
DESCRIPTION:Title: Understanding crosstalk in quantum processors\nby Robin
Blume-Kohout (Sandia National Laboratories\, Albuquerque\, New Mexico) as
part of Centre for Quantum Software and Information Seminar Series\n\n\nA
bstract\nModel-based quantum tomography protocols like gate set tomography
optimize a noise model with some number of parameters in order to fit exp
erimental data. As the number of qubits increases\, two issues emerge: 1)
the number of model parameters grows\, and 2) the cost of propagating qua
ntum states (density matrices) increases exponentially. The first issue
can be addressed by considering reduced models that limit errors to being
low-weight and geometrically local. \n\nIn this talk\, we focus on the sec
ond issue and present a method for performing approximate density matrix p
ropagation based on perturbative expansions of error generators. The meth
od is tailored to the likelihood optimization problem faced by model-based
tomography protocols. We will discuss the advantages and drawbacks of us
ing this method when characterizing the errors in up to 8-qubit systems.\n
\nHosted by A/Prof Chris Ferrie\, UTS Centre for Quantum Software and Info
rmation. \n\nPlease note\, Erik Nielsen's seminar will follow directly af
ter Robin Blume-Kohout's seminar.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Lieven Vandersypen (QuTech\, Delft University of Technology\, Neth
erlands)
DTSTART;VALUE=DATE-TIME:20200625T060000Z
DTEND;VALUE=DATE-TIME:20200625T070000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/6
DESCRIPTION:Title: Silicon spin qubits gain traction for large-scale quant
um computation and simulation.\nby Lieven Vandersypen (QuTech\, Delft Univ
ersity of Technology\, Netherlands) as part of Centre for Quantum Software
and Information Seminar Series\n\n\nAbstract\nExcellent control of over p
hysical 50 qubits has been achieved\, but can we also realize 50 fault-tol
erant qubits? Here quantum bits encoded in the spin state of individual el
ectrons in silicon quantum dot arrays have emerged as a highly promising a
venue. In this talk\, I will present our vision of a large-scale spin-base
d quantum processor\, and our ongoing work to realize this vision. I will
also show how the same platform offers a powerful platform for analog quan
tum simulation of Fermi-Hubbard physics and quantum magnetism.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Erik Nielsen (Sandia National Laboratories\, Albuquerque\, New Mex
ico)
DTSTART;VALUE=DATE-TIME:20200612T003000Z
DTEND;VALUE=DATE-TIME:20200612T010000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/7
DESCRIPTION:Title: Hold the onion: using fewer circuits to characterize yo
ur quoits\nby Erik Nielsen (Sandia National Laboratories\, Albuquerque\, N
ew Mexico) as part of Centre for Quantum Software and Information Seminar
Series\n\n\nAbstract\nModel-based quantum tomography protocols like gate s
et tomography optimize a noise model with some number of parameters in ord
er to fit experimental data. As the number of qubits increases\, two issu
es emerge: 1) the number of model parameters grows\, and 2) the cost of pr
opagating quantum states (density matrices) increases exponentially. The
first issue can be addressed by considering reduced models that limit err
ors to being low-weight and geometrically local. \n\nIn this talk\, we foc
us on the second issue and present a method for performing approximate den
sity matrix propagation based on perturbative expansions of error generato
rs. The method is tailored to the likelihood optimization problem faced b
y model-based tomography protocols. We will discuss the advantages and dr
awbacks of using this method when characterizing the errors in up to 8-qub
it systems.\n\nHosted by A/Prof Chris Ferrie\, UTS Centre for Quantum Soft
ware and Information. \n\nPlease note\, the starting time is only an esti
mate as Erik Nielsen's seminar will follow directly after the ~30min semin
ar of Robin Blume-Kohout.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Marissa Giustina (Google AI Quantum\, Google Research)
DTSTART;VALUE=DATE-TIME:20200609T010000Z
DTEND;VALUE=DATE-TIME:20200609T020000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/8
DESCRIPTION:Title: Building Google’s quantum computer\nby Marissa Giusti
na (Google AI Quantum\, Google Research) as part of Centre for Quantum Sof
tware and Information Seminar Series\n\n\nAbstract\nThe Google AI Quantum
team develops chip-based circuitry that one can interact with (control and
read out) and which behaves reliably according to a simple quantum model.
Such quantum hardware holds promise as a platform for tackling problems i
ntractable to classical computing hardware. While the demonstration of a u
niversal\, fault-tolerant\, quantum computer remains a goal for the future
\, it has informed the design of a prototype with which we have recently c
ontrolled a quantum system of unprecedented scale. \n\nThis talk introduce
s Google’s quantum computing effort from both hardware and quantum-infor
mation perspectives\, including an overview of recent technological develo
pments and some recent results.\n\nHosted by: A/Prof Nathan Langford\, UTS
Centre for Quantum Software and Information\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Daniel Grier (University of Waterloo\, Canada)
DTSTART;VALUE=DATE-TIME:20200616T010000Z
DTEND;VALUE=DATE-TIME:20200616T020000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/9
DESCRIPTION:Title: Interactive Shallow Clifford Circuits: Quantum advantag
e against NC1 and beyond\nby Daniel Grier (University of Waterloo\, Canada
) as part of Centre for Quantum Software and Information Seminar Series\n\
n\nAbstract\nRecent work of Bravyi et al. and follow-up work by Bene Watts
et al. demonstrates a quantum advantage with shallow circuits: constant-d
epth quantum circuits can perform a task which constant-depth classical (i
.e.\, AC^0) circuits cannot. Their results have the advantage that the qua
ntum circuit is fairly practical\, and their proofs are free of hardness a
ssumptions. In this talk\, I'll present a follow-up result\, which attemp
ts to hold on to these advantages\, while increasing the power of the clas
sical simulator.\n\nThe main result is a two-round interactive task which
is solved by a constant-depth quantum circuit (using only Clifford gates\,
between neighboring qubits of a 2D grid\, with Pauli measurements)\, but
such that any classical machine/circuit for the task would need to solve p
arity-L-hard problems. I'll focus on proving a slightly weaker result (NC
^1-hardness)\, but the techniques generalize to parity-L.\n\nJoint work wi
th Luke Schaeffer.\n\nHosted by Michael Bremner\, UTS Centre for Quantum S
oftware and Information\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Guillaume Verdon (X (formerly Google X)\, CA\, USA)
DTSTART;VALUE=DATE-TIME:20200619T000000Z
DTEND;VALUE=DATE-TIME:20200619T010000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/10
DESCRIPTION:Title: Quantum-probabilistic Generative Models and Variational
Quantum Thermalization\nby Guillaume Verdon (X (formerly Google X)\, CA\,
USA) as part of Centre for Quantum Software and Information Seminar Serie
s\n\n\nAbstract\nWe introduce a new class of generative quantum-neural-net
work-based models called Quantum Hamiltonian-Based Models (QHBMs). In doin
g so\, we establish a paradigmatic approach for quantum-probabilistic hybr
id variational learning of quantum mixed states\, where we efficiently dec
ompose the tasks of learning classical and quantum correlations in a way w
hich maximizes the utility of both classical and quantum processors. In ad
dition\, we introduce the Variational Quantum Thermalizer (VQT) algorithm
for generating the thermal state of a given Hamiltonian and target tempera
ture\, a task for which QHBMs are naturally well-suited. The VQT can be se
en as a generalization of the Variational Quantum Eigensolver (VQE) to the
rmal states: we show that the VQT converges to the VQE in the zero tempera
ture limit. We provide numerical results demonstrating the efficacy of the
se techniques in several illustrative examples. In addition to the introdu
ction to the theory and applications behind these models\, we will briefly
walk through their numerical implementation in TensorFlow Quantum.\n\nHos
ted by Chris Ferrie\, UTS Centre for Quantum Software and Information\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Lana Mineh (QET Labs\, University of Bristol)
DTSTART;VALUE=DATE-TIME:20200623T070000Z
DTEND;VALUE=DATE-TIME:20200623T080000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/11
DESCRIPTION:Title: Strategies for solving the Fermi-Hubbard model on near-
term quantum computers\nby Lana Mineh (QET Labs\, University of Bristol) a
s part of Centre for Quantum Software and Information Seminar Series\n\n\n
Abstract\nThe Fermi-Hubbard model is of fundamental importance in condense
d-matter physics\, yet is extremely challenging to solve numerically. Find
ing the ground state of the Hubbard model using variational methods has be
en predicted to be one of the first applications of near-term quantum comp
uters. Here we carry out a detailed analysis and optimisation of the compl
exity of variational quantum algorithms for finding the ground state of th
e Hubbard model\, including costs associated with mapping to a real-world
hardware platform. The depth complexities we find are substantially lower
than previous work. We performed extensive numerical experiments for syste
ms with up to 12 sites. The results suggest that the variational ansätze
we used -- an efficient variant of the Hamiltonian Variational ansatz and
a novel generalisation thereof -- will be able to find the ground state of
the Hubbard model with high fidelity in relatively low quantum circuit de
pth. Our experiments include the effect of realistic measurements and depo
larising noise. If our numerical results on small lattice sizes are repres
entative of the somewhat larger lattices accessible to near-term quantum h
ardware\, they suggest that optimising over quantum circuits with a gate d
epth less than a thousand could be sufficient to solve instances of the Hu
bbard model beyond the capacity of classical exact diagonalisation.\n\nHos
ted by Michael Bremner\, UTS Centre for Quantum Software and Information.
Email cqsiadmin@uts.edu.au to request interactive zoom link.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Ramis Movassagh (MIT-IBM Watson AI Lab)
DTSTART;VALUE=DATE-TIME:20200702T233000Z
DTEND;VALUE=DATE-TIME:20200703T003000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/12
DESCRIPTION:Title: Cayley Path and Quantum Supremacy\nby Ramis Movassagh (
MIT-IBM Watson AI Lab) as part of Centre for Quantum Software and Informat
ion Seminar Series\n\n\nAbstract\nGiven the large push by academia and ind
ustry (e.g.\, IBM and Google)\, quantum computers with hundred(s) of qubit
s are at the brink of existence with the promise of outperforming any clas
sical computer. Demonstration of computational advantages of noisy near-te
rm quantum computers over classical computers is an imperative near-term g
oal. The foremost candidate task for showing this is Random Circuit Sampli
ng (RCS)\, which is the task of sampling from the output distribution of a
random circuit. This is exactly the task that recently Google experimenta
lly performed on 53-qubits.\n\nStockmeyer's theorem implies that efficient
sampling allows for estimation of probability amplitudes. Therefore\, har
dness of probability estimation implies hardness of sampling. We prove tha
t estimating probabilities to within small errors is #P-hard on average (i
.e. for random circuits)\, and put the results in the context of previous
works.\n\nSome ingredients that are developed to make this proof possible
are construction of the Cayley path as a rational function valued unitary
path that interpolate between two arbitrary unitaries\, an extension of Be
rlekamp-Welch algorithm that efficiently and exactly interpolates rational
functions\, and construction of probability distributions over unitaries
that are arbitrarily close to the Haar measure.\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Rodney Van Meter (Keio University)
DTSTART;VALUE=DATE-TIME:20200630T020000Z
DTEND;VALUE=DATE-TIME:20200630T030000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/13
DESCRIPTION:Title: Engineering the Quantum Internet\nby Rodney Van Meter (
Keio University) as part of Centre for Quantum Software and Information Se
minar Series\n\n\nAbstract\nExperimental progress toward a general-purpose
Quantum Internet is advancing rapidly\, but the challenges in building a
Quantum Internet extend far beyond having a physical layer that can create
entanglement across a distance. Quantum Internet nodes must share manage
ment of distributed tomography\, errors\, entanglement swapping\, multiple
xing of resources\, selection of routes\, and more to support application-
requested actions for distributed cryptographic functions\, quantum sensor
networks\, and distributed quantum computation. I will introduce our Rule
Set-based Quantum Internet architecture and the simulation tools that are
enabling us to develop working protocols\, and discuss the need for multi-
disciplinary organizations to address the broad range of problems.\n\nRodn
ey Van Meter received a B.S. in engineering and applied science from the C
alifornia Institute of Technology in 1986\, an M.S. in computer engineerin
g from the University of Southern California in 1991\, and a Ph.D. in comp
uter science from Keio University in 2006. His current research centers on
quantum computer architecture and quantum networking. Other research int
erests include storage systems\, networking\, and post-Moore's Law compute
r architecture. He is now a Professor of Environment and Information Stud
ies at Keio University's Shonan Fujisawa Campus. He is the Vice Center Ch
air of Keio's Quantum Computing Center. Dr. Van Meter is a member of AAAS
\, ACM and IEEE.\n\nHosted by Simon Devitt\, UTS Centre for Quantum Softwa
re and Information\n
END:VEVENT
BEGIN:VEVENT
SUMMARY:Adrian Chapman (University of Sydney)
DTSTART;VALUE=DATE-TIME:20200707T010000Z
DTEND;VALUE=DATE-TIME:20200707T020000Z
DTSTAMP;VALUE=DATE-TIME:20200812T032730Z
UID:UTSQSI/14
DESCRIPTION:Title: Characterization of free-fermion-solvable spin models v
ia graph invariants\nby Adrian Chapman (University of Sydney) as part of C
entre for Quantum Software and Information Seminar Series\n\n\nAbstract\nF
inding exact solutions to spin models is a fundamental problem of many-bod
y physics. A workhorse technique for exact solution methods is mapping to
an effective description by noninteracting fermions. The paradigmatic exam
ple of this is the Jordan-Wigner transformation for finding an exact solut
ion to the one-dimensional XY model. Another important example is the exac
t free-fermion solution to the two-dimensional Kitaev honeycomb model. \n\
nI will describe a framework for recognizing general models which can be s
olved this way by utilizing the tools of graph theory. Our construction re
lies on a connection to the graph-theoretic problem of recognizing line gr
aphs\, which has been solved optimally. A corollary of this result is a co
mplete set of constant-sized frustration structures which obstruct a free-
fermion solution. We classify the kinds of Pauli symmetries which can be p
resent in models for which a free-fermion solution exists\, and we find th
at they correspond to either: (i) gauge qubits\, (ii) cycles on the free-f
ermion hopping graph\, or (iii) the fermion parity. Clifford symmetries\,
except in finitely-many cases\, must be symmetries of the free-fermion Ham
iltonian itself. We expect our characterization to motivate a renewed expl
oration of free-fermion-solvable models\, and I will close with an elabora
te discussion of how we expect to generalize our framework beyond generato
r-to-generator mappings.\n
END:VEVENT
END:VCALENDAR