BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Theory - Chemistry Research Interest Group
SUMMARY:Developing Hybrid Quantum Monte Carlo Algorithms f
or Low Quantum Overheads and Improved Noise Resili
ence - Dr Maria-Andreea Filip\, University of Camb
ridge
DTSTART;TZID=Europe/London:20230524T143000
DTEND;TZID=Europe/London:20230524T153000
UID:TALK194146AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/194146
DESCRIPTION:Quantum chemistry problems have recently become pa
rticularly interesting targets for quantum\ncomput
ing algorithms due to their exponentially scaling
Hilbert space which can be efficiently\nmapped ont
o a linear number of qubits\, promising significan
t memory improvements in quantum\nover classical a
lgorithms.\nHowever\, in the era of noisy intermed
iate-scale quantum (NISQ) devices and hybrid algor
ithms\,\nthe size of chemical systems that can be
treated is still severely limited\, with hardware
noise\nand qubit decoherence precluding useful com
putation even for modest numbers of qubits.\nAditi
onally\, most current quantum algorithms aim to us
e quantum devices to measure physical\nquantities
of interest effectively exactly. Lowering noise to
acceptable levels therefore requires\nnon-trivial
repetitions of the preparation and measurement pr
ocedure\, which is both time- and\nresource-consum
ing.\nQuantum Monte Carlo (QMC) algorithms[1\,2] h
ave proven effective at lowering the computa-\ntio
nal overhead of challenging problems\, both in cla
ssical[3] and quantum settings.[4] In this\npresen
tation\, we combine ideas from conventional QMC al
gorithms with quantum computation\nto devise less
cumbersome hybrid quantum algorithms. First\, we s
how that\, by using a quantum\nprocessor to comput
e projective Monte Carlo (PMC) residuals\, one avo
ids the issue of having\nto importance sample the
wavefunction contributions to the residuals\, whic
h is one of the\nbottlenecks of many QMC algorithm
s. Secondly\, the resulting Monte Carlo estimate o
f the\nwavefunction and its properties is resilien
t to noisy measurement of the residuals so very fe
w\nshots are necessary for the algorithm to succee
d.\nGoing further\, we use stochastic representati
ons of the wavefunction and the Hamiltonian to\nfu
rther reduce quantum overhead. While truncating th
e wavefunction parametrisation or the\nHamiltonian
would introduce a systematic error in a determini
stic approach\, this is not the case\nfor QMC\, in
which we find that biases average out over the co
urse of the calculation.\nFinally\, we explore the
expansion of these methods to excited electronic
states.\n[1] G. H. Booth\, A. J. W. Thom\, A. Alav
i\, Journal of Chemical Physics 131\, 054106 (2009
).\n[2] M.-A. Filip\, A. J. W. Thom\, Journal of C
hemical Physics 153\, 214106 (2020)\n[3] J. E. Deu
stua\, J. Shen\, P. Piecuch\, Journal of Chemical
Physics 154\, 124103 (2021)\n[4] M.-A. Filip\, N.
Fitzpatrick\, D. MuĂ±oz-Ramo\, A. J. W. Thom\, Phys
ical Review Research 4\,\n023243 (2022)
LOCATION:Unilever Lecture Theatre\, Yusuf Hamied Department
of Chemistry
CONTACT:Lisa Masters
END:VEVENT
END:VCALENDAR