winterterm 202312024

Introduction to Quantum Computing (IN2381)

Christian B Mende
1. Introduction
18.10.2023

see corresponding slides

2. Basic concepts

2.1 Quantum bits (qubits)

Nielsen and Chuang section 1.2

classical bit O 1
superposition of 0 1
Quantum bit qubit
described as
quantum state 14
is
a
a 107 Pll a Bea
14
with la t Ipl 1 I normalization

Ket notation 4 motivation from inner products

4 ed with
Mathematical description
10
113 19
m 14 Y
Different from classical bits i cannot in general
observe measure a qubit the amplitudes a and B
directly
standard measurement will result in
Instead
0 with probability al
with probability 1,112
n
the qubit wavefunction
also changes
The measurement collapse
O the qubit will be 147 10
If measuring the measurement and likewise
after
directly
qubit will be 14 113
if measuring 1
can estimate the probabilities lol ad PR
I practice
in experiments by repeating the same experiment
times via the outcome statistics
many
repetitions are called trials er shots
The
Circuit notation
14 MT classical information

realizations
what is a qubit physically Many possible e.g
are
10 and 11
of a photon e.g left right circular
two different polarizations
nuclear or electronicspin p I
alignment of a
excited state of an atom
ground state or

counterclockwise loop current states in a

clockwise or

Josephson function superconducting qubit

A useful graphical depiction of
a qubit is the

Bloch sphere represention

to be real valued then can find
If a and p happen

angle I e IR such
a cos I B sin f Tik s
lat't list cost t si er
I
3 general represent

a it cos E
p ei rt'd si E
so called phase angles y for a
using
and 8 4 for P

Then 14 eir costa o eiffel sin E 11

Tires
eir costello teil sin Y 11
be ignored here
by two angles I and p
Thus 14 is characterized
defined as
these specify the poi t

I
on the surface of a sphere

Bloch sphere Felix Bloch

Exerise why are 10 and n

at the poles
 25.10.2023
2.2 Single qubit gates

Nielsen and Chuang serious 1.3.1 2.1.8 4.2

Principle of time evolution

the quantum state 14
at current time point t transitions to a new quantum
state 14 at a later time t t
Transition desivibed by a complex unitary matrix U
14 U 14
matrix vector multiplication

Circuit notation pygtygate

ay
time flog

Notes circuit is read from left to right

beater U 14
but matrix times from right to left
a U preserves normalization

Examples
of the classical NOT gate 11
quantum analogue o

107 11 Pauli t gate

flip
go
y
on
check X 10 9 91 11
x 11 107
 Pauli Y gate
YE 62 É
Pauli Z gate
Z 63 j Y
t leaves 10 unchanged but flips the sign
of the coefficient of 11
Recall the Bloch sphere representation

14 cos E lost eilsin E 1

Then 214 cos f 10 eia sin E 11

10 si Alles
ein l
gross t.gg I
new Beech sphere angles d
p d p it

lrotation by it I 180 around t axis

X Y Z gates are called Pauli matrices

The Pauli vector 8 on or o x y z

is a vector of 2 2 matrices

Hadamard gate
H

lol pie I a 1071 plot

 Phase gate
s

Tate T
l ÉÉÉ
Note T S f si se eine ait i

Pauli matrices satisfy

I identity for j 1,213
Gj
Gj Gk G Oj for all j k
Gi Gk OnGj 2i g for j kill
oj or
commutator a cyclic permutation of 11 2,3
FA 13 AB BA e.g j 2 6 3 1 7
7,21371 1213,71 3,112

matrix exponential and rotation gates

Refill Ry101 Rell
vs see handout 1 matrix_exponential.pdf
 Z Y denomposition of an arbitrary 2 2 matrix

Ue k there exist
For any unitary matrix
IR such that
real numbers a P r S
e

in o
e

U eia o

Th
2.3 Multiple qubits

Nielsen and Chuang sections 7.2.7 2.1.71

basis States 10 and 17
single qubits
superposition of
So far

For two qubits this generalizes to

1007 101 1107 1117

basis states all combinations bitstrings

as computational
of Os and Is

General two qubit state

an 110 an 117
t do 101
14 no 100
E e such that
with amplitudes aij
Int 1 normalization
loool t front t knot t
 Can identify the basis states with unit vectors
110
too 101 in E
thus
14 GI EE

qubit of two qubit state

What happens if we measure only one a

measure the first qubit obtain result

we
Say
o with probability la t no l
not't Int
1
Wavefunction directly after measurement
do 100 an 07
measured 0 141
if
Titta
Me
1 90110 on normalization
if
measured 14 Make
Mathematical formalism for constructing two qubit
states

vector spaces
tensor product of 8.11.202
V and W
Can combine two arbitrary vector spaces
W
to form the tensor product V
see handout
vector_space_tensor_product.pdf
 Generalization to n qubits 2 computational basis states

10 011 10 107 h m
Mjg all bit strings of length u

Thus general n qubit quantum state

register
also denoted as quantum given by

14 É É fi Su tax I then Tito

If a Ix

lo 2 i such that
with axe f for all e

1114711
É last I 1 normalization

hard to simulate on classical

in general
computer for large n due to this curse of dimensionality

Vector space as tensor products

the d
Iggy
2.4 Multiple qubit gates

Nielsen and Chuang sections 7.3.2 1.3.4 2.17

As for single qubits an operation on multiple qubits

matrix U
is described by a unitary
2
For n qubits he 62
 controlled NOT gate also denoted Choti
Example
two qubits control and target
control is 1
target qubit gets flipped if
i s 1007 101 I 1017 110 171 111 1107
100
control forget

can be expressed as
all be 0,1
A la atb for a
lab addition modulo 2

circuit notation

la la

Hs I la b

Matrix representation o o o

Marot

Pauli X
unitary
Alternative sircuit notation I it
Can generalize Pauli X to any unitary operator U E K
target qubit controlled U gate
acting on is

to 100 101 4 101 110 117041071 1117 11xo 411

100
outrode
target
 it u

Example controlled Z

n
z
Exertise show that centrolled Z gate is invariant
when flipping control and target qubits
target qubits
Controlled U for multiple
control

target 1
qubit and snot gates are universal they can
Note single
operation on n qubit
be used to implement an arbitrary unitary
classical NANOgate
Quantum analogue of universality of
Nielsen and Chuong section 4.5
proof in

Example of a circuit so sisti

g
o
ly of single qubit gates
and CNOTS
H
x f Rye
g
9
Z

time
Matrix Kronecker products matrix representation of
single qubit gates i parallel
anting
A
B

Operation on basis states a b e 10,77

Ala lo Bibs ADB ab

iffy
Example A I identity B Y
1007 i 101
10701441
1011 A 107 i 100
1141
110 is 117 y 1071 i 111
1111 H 17 14471 i 110

Matrix representation
I

Y 1 1 ex
Y 212
General formula Kronecker product matrix representation of
tensor products of operators
an B an B and
an B a B B
e empty
Ao B
In B am B amnB
NumPy
AB
Ent ane npKron
for A f Be
Generalize to arbitrary number of tensor factors e.g
la A
lb B E ADB C
s C BOC
B C A
a bc

000
O o 1
0 10

Basic properties
elementwise complex conjugation
a B A 13

b IADBIT Atx Bt transposition

Kl A BIT At Bt
associative property
d ADB C A B C
I A IA c 0113.0 for matrines of

3
matrix watt compatible dimensions

I
I
I IEEE
matures is Hermitian
I Kronecker product of Hermitian
matrices
9 Kronecker product of unitary is unitary
and et
follows from s
2.5 Quantum measurement 15.11.2023

Nielsen and Chuang sections 1.3.3 2.2.3 2.2.5

Review measurement of a single qubit 14 a 107 1311

with respect to computational basis 510 117

14 Teassical data
measurenet ont ion O or 1

alternative notation
attYw iEtitonioeeaps.l

ftp.tjjt
14

measurement basis
Unitary freedom of choice of
orthonormal basis tan 197
Given an
orthonormal basis
can measure with respect to this
base change before and after measurement
by performing
e e unitary
U 14 lur
measurement with respect to Huss in

1,71 51
14 1471

representing a initiatiistitiusiacan the

Ut 4 410 a la
147 Lf
measure net result I or I
will obtain
with respective probabilities an and tail
state
After measurent and applying U 14 agile being
 Abstract general definition of quantum measurements

Quantum measurements are desivibed by a collection

system
Mm of measurement operators acting on the quantum
possible measurement outcomes
with the index m labelling
the state before measurement by 14
Denoting quantum
probability
result m occurs with
11 Mm 14711
plant 141MtMm 141
measurement is
state after
Mm 14
11Mm 14711
completeness relation
The measurement operators satisfy the
MutMm I
In
such that probabilities
sum to 1
14 94147 1
E 41Mt Mm14 412114
En plat
Example measurement of qubit 147 2107 811
a

with respect to computational basis 10 11

M 10 01
114,01 1
M
117511 18 9
94114014 141
plot 41 MTM 147
1131
pas
Projective measurements
see projection operators handout

A projective measurement is described by an

Definition
observable M a Hermitian operator acting on the quantum syste
spectral 3
decomposition
M E tm
with eigenvalue dm
with Pm projection onto eigenspace
to
The possible outcomes of the measurement correspond

the eigenvalues fu
tu
Probability of getting result

pliml yl Pe 14
PutPm Pm Pm sins Pm is a projection

the measurement
state of the quantum system after
Pm14 Pm14
11Pm 4 y plant
Remarks
Projective measurements are special cases of general measurement
framework

Projective measurements co timed with unitary transformations

to
land auxiliary qubits are equivalent
general measurement framework
Nielsen and Chuang
see pages 94 95 in
 Average value of a projective measurement

E TM Itm petal Et 4112147 412 114

M
941M 147 F
if 14 isclearfromcontext

Corresponding standard deviation

11M smit
Cim
giggap
Examples
basis 107 11
Measuring a qubit v computational
rt

is actually a projective measurement Po 10701 P like

w.at orthonormal basis
measurement un Iu
In general
is a projective measurement

set Pm fun um for me 1 2

observable M by
define
4m Pm with arbitrary di d ER I t
M I
Z
Measuring Pauli
Z 1 tri
Tio so Tensai
standard measurement
agrees with
computational basis 10 11
w rat
2.6 The Heisenberg uncertainty principle 22.11.2023

Suppose C and D are two observables and 14 a quantum state

Then 144154031471
A c 110 2

experiments repeated preparation of 14

Interpretation for
measure in some cases D in the other cases
standard deviations Cl and 1101
to obtain

see handout heisenberg_untertainty principle pdf

for derivation
3. Entanglement and its applications

A nqubit state 14 n 2 is called entangled if it

cannot be written as tensor product of single qubit states i e

14 19.7 19mn Ee
19mn70 for any 140

Einstein
also denoted EPR states
PodolskyRosen
Example Bell states

11007 111 b
β a

Pon 101 1107

1pm 100 111

1pm 1101 110

Bell states
Quantum circuit to create

1 7
1 1Pa for x.ge 011

3.1 Quantum teleportation

Nielsen and Chuang section 1.3.7

Scenario two experimental physicists Alice and Bob
other
are far away from each

when visiting each other a long time ago they generated

the EPR pair Poo each keeping one qubit of the pair
Alice's task is to send another unknown qubit 14 to Bob
Note measurement is not an option
 Quantum circuit for teleporting 14

14
Alice

my 1 11 14
141 14

Input 14 B 14 Po a 0 1005 111 piÉ Ém

C NOT
α 1 PM
after
11171 β 1 110 101
14 210 100

after Hadamard

107 11 1007 111 β 10 11 110 101

α
4
a 1000 a 1011 α 1100 1111 p 1010 β 001 p 110 p101

100 α 10 111 101 a 1 D10

α 10 β1 111 911 β 10
1107
Now Alice measures her qubits w.at computational basis

projective measurement
with P 1007500 I R 1017407 I
P 7107410 I Py 1117471 I

If Alice measures 00 then 14 will lollapse to

1007 a 10 β 100 14 Bob's
qubit at place

similarly 00 α 10 1 PH
01 a11 β10
10 a 10 β11
11 a 11 P 10
Aline transmits her measement results to Bob classical information
Bob then applies Pauli and or Pauli if necessary to recover 14
is instantaneous
Even though wavefunction collapse
no faster than information transfer possible due to required
light
randomness of measurement outcome
classical communication and
3.2 EPR and the Bell inequality

1Nielsen and Chuang section 2.6

EPR Einstein Podolsky Rosen

of physical reality
EPR paper fan quantum mechanical description

be considered complete 19351

mechanics is incomplete
The authors argue that quantum
elements of reality
since it lacks certain
property can be predicted with certainty

Scenario Alice and Bob are far from each other

state
but share the entangled two qubit spin singlet
Ben 101 110
and Bob measure the observable i KX a Y V Z
Alice
qubit
with ie IR 1011 7 on their respective
is is Hermitian and unitary and has eigenvalues 1
Recall
her measurement immediately before Bob
Alice performs
Example 1 107 01 1 11 1 571
observable Z
0,011
Istandard measurement
measures eigenvalue
if Alice
wave function collapses
to 101
1
1 to 110
as Bob will always obtain the opposite measurement result
 29.11.202
1,010 observable eigenstates 107 11
corresponding eigenvalues 1
w r t 17 1 basis
measurement
wavefunction as
can represent the
1 1 1 7
1pm
107 117 107 11
namely 1 7 1
10 117 107 11
1101 1105 Pm

Alice measures eigenvalue 1 wavefunction will

If
collapse to 1 7 Bob's qubit is in state 1
so he will certainly measure eigenvalue 1
conversely if Alice measures 1

general observable i general unit eitor P R

denote the orthogonal eigenstates of E by la b
then there exists complex members Pitis such that
10 α 1a β b
11 8 a S b
Inserted into Pen see also Exercise 8.1cal
101 110 8 Pr lab Iba
Twwith U C
U is a base change matrix between orthonormal 10 les
and a b basis a unitary Idetful 1
Exelis
Can represent def U en de R 1.2 e

in summary 1017 110 e I

fÉ
as before Bob will obtain opposite measurement result as Alice
Therefore Alice can predict Bob's measurement result
However there is no possibility that Alice could influence
her measurement
Bob's measurement after performing

since they are far apart speed of light too slow

property i of a qubit is an
EPR argument
element of reality
however quantum mechanic does not a priori specify this
property for all possible but only probabilities
and is thus an incomplete description of reality

Instead hidden variable theory there must be

additional variables hidden in a qubit which
Bob's measurement of 8 for all possible PER
determine
experimental test whish can invalidate
Bell's inequality
theories Bell 1964
local hidden variable
faster than light communication possible
local no

otherwise one could send information backwards in time

alcording to special relativity

many repetitions to
collect statistics
Experimental schematic
of the following setup

prepares two particles

sends one to Alice and one to Bob
 binary property values QE 11 RE It Se 113 Te In

Aline decides randomly whether to measure property Q or R

S or T
Bob
Bob their measurements almost l simultaneously
Alice and perform
information about the result canbetransmitted in between
such that no

this protocol Aline and Bob meet

After completing
measurement data
to analyte their
Consider the quantify
s T 2
RT Q T
QS RS 191 IEEE
Denote by p grist the probability that the system before
measurements is in state Q 9 R r S S T t then
RT QT t
ETOS RS
t 2
IEEE
49 2

at the following
By linearity of E arrive

Bell inequality
ETQS ETRS ETRT ETQT 2

Each term can be experimentally evaluated

e g I QS Alice and Bob average over cases
where Alice measured Q and Bob measured S
 Compare with quantum realization of the experiment
Charlie prepares the two qubit singlet state
14 101 1107
and sends the first qubit to Alice and the second to Bob
Observables
5 21 the secondqubit
Q E
2hrfirstqubit
R X T 2ft
Measurement averages c f Exercise 8.1
Qs 4100514 RS
RT QT

QS SRS RT QT 2 E 42
violates Bell's inequality

Actual laboratory experiments using photons agree with

predictionsby quantum mechanics thus not all limplinity
assumptions leading to Bell's inequality can be satisfied
realism physical properties Q R S T have definite
values independent of observation measurement

locality Alice performing her measurement cannot

influence Bob's measurement and vice versa

Nature is not locally realistic

most common viewpoint realism does not hold

Practical lesson use entanglement as resource

 6.12.2023
4. Quantum search algorithms

Nielsen and Chuang section 6

elements OCN
Classical search through N unordered
Quantum Grover's algorithm O r given certain preconditions

4.1 Quantum oracles

elements labelled 0,1 N 1

search space of N 2
with 1 ME N
Assume there are M solutions
0 N1 10,1 by
Define sorresponding indicator function f
if element is not a solution
f 11 1 is

Quantum version of f
Uf defined for romputational basis states a
quantum oracle

m i n
Ix un 1 1 10
Uf
197 1907117
oraclequbit
9 50113 911 2
states to basis states and
Note Up maps basis
satisfies Up I q fix fix q
basis states and is i particular unitary
thus Up permutes
 Initialize oracle qubit in superposition 1 10 11 then

101 Ut if fix o
1 30 41 13 f 130 if 711 7

Uf 10 11
In summary
1
11 1 MAY

EEEE.EE die
it uncaged

Ix 1 11th
effective astion of oraile Up

Oracle marks solution by a phase flip

sunk an oracle without knowing solution alread
tow could one construct
factorization of a large integer me N
Example
Finding prime factors of difficult on a classical computer
m is
runtime
no known algorithm with polynomial
in the bit length of m

a given
E N divides m is simple
but testing whether
digital
can perform arithmetic operations for trial division on a

well oracle which reignites a solution

quantum computer as
better quantum algorithm integer fastorition Shor's algorithm
for
Remark
4.2 Grover's algorithm

search space with N 2 elements M solutions

Overall circuit diagram for Grover's algorithm

10 0 i
G I G
I qubits
Groveroperator
 Initial Hadamard transform H 1
Note H10 110 115
EY H1 10 11

for 50,1
HH 1 11 12

Applied to
several qubits

HmKE E.ni IEI

III IE bitstring
H 10 0 12 14 equalsuperposition state
In particular

Definition of Grover P 1 2105401

17 117
mi µ p
g n

G Uf
orate

FfÉH
211911 I
4
2 4 541 I

In summary G 2147541 I Up

12 4 41 I Iaux Uf
 Geometric interpretation

Define
Ias in

1ps a
angle I defined by sink
such that
14 os a si IP
Note by definition Uf
a K Up B 1B
is a reflection about k
Up and B
spanned by k
within subspace
is a reflection about 14
Likewise 2147541 I
and β
subspace spanned by k
Sinie IT is part of
invariant
G leaves subspace
produit of two
reflections
Thus G is a

a notation by angle I
G is
cos p k sin 191 β
10 sin 19 4 β
G 10 cos y d a 13.12.20

For k applications of G
G 10 cos ptk.nl a sinlqtk0l1p
initial state 10 14 p f
for
kt 2 0 a sin 1kt 9 113
G 4 cos

Goal rotate to 117 ie K E 0 I

sine sine if t.ua
2Er
 Thus need FY rotations
Ktf k I I
K OCT for N
Final step standard measurement
state with high probability
will collapse quantum
P i.e a solution
to a basis state forming
4.3 Optimality of the search algorithm

6
Nielsen and Chuang section 6
show that any quantum search algorithm
needs R N
Goal
oracle sales N is already optimal

For simplicity single solution

that oracle flips sign of solutions
Recall
denoted Up in previous sention
Ox I 21 76 1
Most general form of algorithm oracle talls interleaved with
unitary operations Us Uh

State after k steps

14 Uk Or Uka
initialstate
We also define
U 14 omit Oracle calls
14k Uk Ukn
strategy of proof upper bound of
11 145 1414112
DK
Insisttittion

Die grows as Olk but must be R N to distinguish

between N alternatives
First show that Dk 4 k by induction
4 0 Do 0
11Man 1 11711
K 1 1
a I E 110 1 7 14 1
E
EE.IE Ei i Eiiinn
I HI 24111

KEEN
Ok 4 511145 14114 51414712 4

De t 4 The 4
i
E't 4k 8 k 4 4 4 1

Second part of proof Dr must be R N

To find solution want that 1T 17

suppose 14 14 712 for all x

probability of sulless at least 50

w log x 4H 14 141 I can multiply byphase factor

11147 1 711 24 14 7
14H
2 2 2 E
Therefore E 11145 1 711 2 KIN
F 111 7 1417111 then
Define
2 Re 14k
F 1 1 1,412
Retz 171 for a f E

11114kt 15 H 14k
Ok I

E 2
IEEE
F
1 Eax txt salt Hall 1164
2
E F 2 FE FE FF E

Iran MAY
12

N FE 2T
as Nto

N
in summary

fgf
number of oracle
evaluations
 20.12.2023

Thought experiment If it was possible to search using Olloglin

oracle calls then a quantum computer could solve NP complete

problems efficiently just search through 2 witnesses

using
a.si ii.i
5. The density operator

So far state vector 14 describing a quantum state

convenient alternative formulation for quantum systems
about which we only have partial information

density operator also called density matrix

5.1 Ensembles of quantum states

Nielsen and Chuang section 2.4.1

consider a quantum system whish is in

one of several states 14
ensemble of quantum states
with probability p pi 14
The density operators of the
ensemble pi 14 is defined as
outer product
5 I P 4 54,1
Quantum mechanics in terms of density operators

a unitary transformation U maps

unitary operations
U 14 and the ensemble to pi U 14
14 I

density operator is transformed as

Thus the
1 ut usut
s EP His a
t.EE
y S
 measurements measurement operators Mm
system in in state ti then probability for result
if m

is triABC to BCA
given i
i 4 MutMm14 trtmntmn.lt 44,13
plm
Thus overall probability for result m is

i MatMm 4 54.1 pi
plan plm pi Itr
trtmmtmw.IE
4i17 trtMIMns
Density operator Sm after obtaining result m

state i collapses to 14 I
14

Thus Mm 4 4 Mmt
Sm
I philm 14 5471 pci Mm 14 planti
Mt
x̅ I
Pi
Mu m9i 17m Is
PCAIBl PlAnB
Him Him Baye's the.ve PB
PCBAI PCA
Terms PIB
Note that gun is now expressed in of g
and the measurement operators
without explicit reference to the
ensemble pi 147
5.2 General properties of the density operator

I Nielsen and Chang section 2.4.2

Chavarterization of density operators An operator f is the density mater
associated to some ensemble
pi 14 if and only if
1 trig 1 trace condition

positivity condition
2 g is a positive operator

called positive operator if it is Hermitian

Remark s is a

and all its eigenvalues are 0 equivalently if pls p 0

for all vectors 19

Proof
Suppose s Pi 14144 then
1
fr Ts pi to514 54,1 Epi54142
and for any state 14
491519 I p 49 4 4 19 Epi 1491471 0

i e a Hermitian matrix by spectral theorem

g is an operator

Xj and orthonormal
there exist eigenvalues corresponding

eigenvectors pj such that

1957591
5
Itcondition
Since 5 satisfies the trace
1 trts di fr 5195791 fi Its
due to positivity of S tj 0 for all j
Thus can interpret eigenvalues Xj as probabilities

Xj 145 is an ensemble which gives rise to g

From now on we define a density operator as positive operator g
with trip 1

Language regarding density operators

mixedstate
pure state
state 147 s describing quantum setup
Quantum system in a

corresponding density operator cannot be written as g 147541

147641 intuition in the ensemble

g
such that representation pi 14 of s
to all the probabilities are
trts 1474445441
strictly smaller than 1
4414 1
Then tuTs pi 1

In general Let S be a density operator Then fr Ts 1

and fr Ts if g describes a pure quantum state
1 if and only

Proof Denote the eigenvalues of s by ti then O Xi 1

since g is positive and 1 fits It

1 with 1 precisely if
Moreover fr Ts It and the others 0
one of the eigenvalues is 1 are

Ensemble representation is not unique

Example
07401 4117411 I a salt b Sb
g

with la F 10 Fy11
16 107 5411
But note that 10 11 are the unique eigenvectors of g
and all to
 ensemble 14,7 10.1.2024
For the following given an
pi
Tpi 14 such that s
set 145 I 1457541

Ensemble 514T generates the density operator g

To relate an ensemble 147in m to another lyjkj.a.in

ensembles with zero vectors
in case mtn we pad one of the
such that without loss of generality m n

ensemble of density matrices

Unitary freedom in the
and 1145 generate the same density matrix
The set 145

if and only if
Uij 195
147
unitary matrix His
for some

Sketch of proof
Insert definitions
use the spectral
decomposition of the density matrix

In al with tutte See

s
set x̅ F Hk express 14T Vik x̅

for some complex coefficients E

Then
E axil 1474451 ErinYt A SEE
I s

x̅
the IX
This equation can only be satisfied since are

orthogonal and this Xi Tel enicardy independent if

 See Vik Vii VT Utr
VtV I V unitary
in other words if V is a unitary mat.it

By the arguments 145 I wjk X

for a unitary matrix Wjk
Thus II k wit145 rut 15
19 via

and V wt is as product of unitary matrices again unitary

can be generalized
The Beach sphere picture for qubits
to mixed states by the representation
I 8 8.8 MX KY MZ
g 2

ER 1151 1 the Beoch vector of g see sheet 11

with
147541
coincides with hitherto definition in g
case

5.3 The reduced density operator

Nielsen and Chuang section 2.4.3

Let n E N
Definition partial trace n

defined in terms of the

The partial traie operations are

conventional matrix trace by

fr try MoMa t.IM M
conventionaltrace

to E 4 traTM Me ATM M
E C and Mr E E together with
for all M

tu.to MoMa β Mom a t.tn 4 It In

trim
If trim trtr.nl
 Consider a composite quantum system
consisting of subsystems A and B A
for example A m qubits Bin qubits β
he 2m n 2h
Let the quantum system be described by
a density operator pAB

Define the reduced density operator for system A by

9A try SAB
pay stÉB
and analogously

gB try Is
Examples
Ian Ian A and lb lb EB
For any quantum states
147441 tarsal.fr iadcat
to I laical
fff ft
Given a density matrix s for subsystem A
B
suppose that the overall
density initix is
SAB 586
Then to is 6 s.fr d s
0 a
trats a

with
trff 100 111 Bellstate
SAB 147441 14
Expand s leads to
111 001 6111
SAB 100

10074001 1007411 111 4001 11174171

SA to TSAB 103401.4011 1 45 11815 15st
sep
107401 11 511 I
Note composite system is in the pure state 14
whereas the subsystem is described by the mixed state I
Indeed a mixed state trt E trTI I 1

Motivation justification for partial trace

Let it be any observable on subsystem A then we want that
A yields the same statistics for measuring M as
g
SAB for measuring MQ
on B
In particular
fr TM SA tr Mot SAB MPI
M
on A on AB

for all density operators g The partial trace operation

SAB is the unique operation withthis propert
for computing s from
Nielsen and Chuang Box 2 6
17.1.2024
does
Application to quantum teleportation Why quantum

teleportation not allow for faster than light communication

via the instantaneous wavefunction collapse

Recall the corresponding quantum circuit

Alice

Bob
At 143 Alice has completed her measurements

her qubits have collapsed but Bob does not know

her measurement
results yet
lecture
Intermediate state 14 see previous

141 100 107 PID 101 a 1 β10

βD 111 α 11 Plo
110 la 10

Thus directly after Alice's measurement system is in state

Bob's perspective who does not know the meanie't

from t

191 100 α 10 β 11 with probability

19 101 1 117 810

19 110 la 10 PID
111 α 117 1310
194

Corresponding density matrix of ensemble 198

j 1 4

SAB 1974451 14 611

Roowost la.to tp
101740110 x 1 Plo a 471 β 401
11074101 α 10 PM 01 1 411

Affubits i
Reduced density
1 a S Ya.cn as 9Yitfa
try s aneso gB 1 10
a 11 t 1310
111 β 411
2

12 411 8 401
901

2 401 P 11
a 10 P111

a 117 8107 1
411 Prot

211 11 101501 21 11 91
I
coefficient ᵗ
and p are 0

107501 11 11 E
independent of 14 α and it coefficients

since GB E any
measurement by Bob cannot reveal
Alice cannot transmit
any information about 14
i e

information encoded in a B via the instantaneous

wavefunction collapse to Bob

6. Quantum operations

Nielsonand Chuang section 8.2

6.1 Motivation and overview

In general shanges of quantum states affented by

unitary time evolution or wavefunction collapse during measurements

Quantum operations also called quantum channels

are a mathematical generalization and unification of

these contepts
matrix
Abstractly eggdensity
g
quantumoperation

Special cases
Ug Ut
unitary time evolution s

measurement with measurement operators Mm

recall that plan to TMI Mms

Frotability for outcome m

State after obtaining result m

Mms Mmt Mug ME

Sm
trTMᵗnMms Plm
channel without renormalization
Corresponding quantum

Em 5 Mms Mmt

plan to Emts is the probability that outcome m occurs

 Consider the scenario of performing a measurement
but not recording the outcome
is weighted sum
density matrix after this process
over

all possible outcomes

plan Sm Mms Mmt
Els
Fratorsum Kraus representation
of
Different but equivalent perspectives on quantum operations

system coupled to environment 1Stinespring dilation

operator sum Kraus representation
axioms
physically motivated
Choi matrix representation

6.2 Environments and quantum operations

be regarded as interaction between a

Open quantum system can
system initially is state g
principal quantum
environment initially in state genu

The principal system interacts with the environment

i e a time evolution of the overall system
described by some unitary U

circuit diagram mathematical representation

S
U U s genu Ut
Seur
 principalsystem is the reduced density matrix
Output on

of the principalsystem
5 Els
s treur Ulsos.nu Ut a
face Seur
t out environment
street

Example Usnot

s Els

105101 0

s 107401 son103 11 to 117401 en174

Represent s

then
MCNOT S D 10H01
U Not
Katnot
son 0074701 So170 00 for 7074101
Ucnot s 10075001
Soo 1007500 for 00 11 So 1174001 Sin 1117411

trenut So 1054015012 s.nl

fd siolkk4tsmtkly
107401 son 117511
s

PsP P SP with Po 107501 P 117411

set to zero
off diagonal entries of s are
6.3 Operator-sum (Kraus) representation 24.1.2024

Let lek be an orthonormal basis of the environment

quantum system assume w 1.0.9 Senu e Seol
12 is in a mixed state
see futorial if environment
state
then ran equivalently work with a pure eni matter
lake I on environment
1st trear U g e e 1 at
U s le.ke.tl Ut Io lek seal
I treno
trTA f.CIAlj
e ed Ut lea
see U s
I
inner product w.tt environment

Eks Eat with

matrix with entries Ek em f e k m e

E a complex
Ketose U mole
The Er's are called operation elements
or Kraus operators of E

Is Ens Ext s

Tell lek
Revisit example
USNOT EK Ext
s 1s

105101
I
0

Elem 41,014 1 m 85 E I p ay
m
E em e 11 Unot m E p
Completeness relation of Kuans operators

If E is trace preserving then for any density matrix g

1 trtest trt Eas Ent trters

in
IEtE s
trT EEntEn s
with trig 1
should hold for arbitrary s
Ent E I
channels with Eat I
we allow for quantum

A B if B A is positive semidefinite psd

is called positive semidefinite
A E E
if A is Hermitian and v A v7 0 for all been
Falligualues of A are 20

Enter stems requirement that to 151 1

I from

Physical interpretation as measurement performed on environmen

Remark
with respect to lek basis ie measurement operators lek Teal

For outcome k state of principal system is

415 leoke.tl at E seat
α
treat
5k
Ife Leffel
ie Ek's play the role of the measurement operators
on principal system
System environment model of knows representation
Given trace preserving s Ers Et
is there a corresponding system environment representation

Define model environment as vector space of dimension n

Yes
lek
with orthonormal basis that
assume environment
k o n y
starts in leo state

U 14 14 lek
Define unitary A via e
I
and use matrix extension to a unitary operator a
on the combined system
is possible since for any principal quantum
states 14 p
This
yle Utu 1976 4 EntE 9
4k
41 EntE 9 419
U preserves orthogonality
FEE
U has the desired property since
e 1 Ut treu Ees EE e seal
trenu U so e

Es Et Ers Et
6.4 Axiomatic approach to quantum operations

Alternative view point physically motivated axioms

which a quantum operation E must obey
A1 trie 151 is probability that the process occurs

thus 0 trteigl 1

for all density matrices s

A2 E is convex linear
probability vector
Pi Si pi si for any p
and density matrices si

A E is a completely positive map Counterexample

EIA must be positive semidefinite pied Effffffi

for any ps d matrix A

Moreover when enlarging the principal quantum system Q

by another quantum system R then

A must be paid matrix A
I E for any psid
Q on the combined system RQ
on
31.1.202

Theorem The map satisfies A1 AZ A if andonly if

s E S Et
makines Ek with Ente I
for some set of complex
Sketch of proof
enlarged system Ra
to verify A Let A be a p.s.cl matrix on

then for any veitor 14 on RQ

441 1404141147 441 I En ACI E

In matrix
II I t.its.ee

II
whish acts on denoted Q dimension n
Principal system
Introduce another quantum system labelled R
with same dimension as Q R

1 orthonormal basis of a
Let ja 5 n an

5 1 n i R
jp
state cf exercise 12.1
Define the maximally entangled for a single qubit
RQ 100 111
1a I lip Ia f

and
I E 1 741 p.s.d.by A ie

density matrix on combined system

Choi matrix

turns out to completely specify

For any state 14 45 1a on Q set

147 45 lip on R
 4416157 2441 lip Girl lia Jal 147

on
I 4,4 labial IE E4 liakial
1147441

decomposition 197491
spectral
If L
eigenvalues
absorbed to 15k

Xk 4 44kt
1511 sk.gg Ir i'a 15k A 14k
Can represent
coefficients

For each k define a linear map E Q Q by

Ea Ia M sk.jp j'a and linear extension

Then El 14 ME 45 sn.jp i'a

said
Effy's
a 4in R
podunt on R
En 4 41Ent n 4515 Ssn F
F 451614 1475411
n.SI 1E E1
Holds for arbitrary 14 It
1s Ens Ent for any density matrix s
 6.5 Choi-Jamiołkowski isomorphism

One to one correspondence between

density matrices on
Quantum channel
E Q Q largersystem RQ
dimension as Q
with R of same
Choi matrix

via
For given the corresponding Choi matrix is defined

6 IOE a al with a Ir Iia

maximally entangled state

conversely for a given Choi matrix 6 can recover

using the above steps

k E
Is EK S It e.AM
1041144
Q b G
6
4
7.2.202
R

Note Kuans operators can be identified as ensemble for o

Is Snl with 151 in Ir Ek sa

6.6 Unitary freedom in the Kraus operator representation

Nielsen and Shuang section 8.2.4

and F Fu Kraus operators

Theorem Suppose Em Em are

giving rise to quantum operations E and F respectively

Is EE ERSEat and FIs Frs Ft

shorter list we may assume that m
to the
By appending zero operations
F and only if there exists a unitary matrix
Then E if
such
U laiilijn.in that
Ei hii Fj
consider the corresponding Choi matrices a and
Sketch of proof
Ensemble representation
sit 1515 A Iir E Is
Ge E E
1st F
of É 1st SFI E lip F Ija
ensemble for density matrices
By unitary freedom of the
see section 5 2

for some unitary U ai

i 1 n
Ei Ui Fj
6.7 Examples of quantum operations

Nielsen and Chuang section 8 3

channel flips 10 11 with probability 1 p p tap

Bit flip
E F X
E IF I
Epels EnSET ps
1 pitex

a r

It
density matrix g
Recall Bloch vector of a

1 118112 see Exercise 11.2

trig

E
It
Earls with
i
analogously
Phase flip
E FI E Tp Z

Epp f EhsEnt
1
p.pt
plZg r
For p by unitary freedom of Kraus representation

E KE II It f 107601
P
E KE I 12 89 117411
P
PosP P sp equivalent to
Epp s
standard basis measurement

Depolarizing channel
with probability p
replaces by completely mixed state I
11 pls
Epp s p I
Bloch sphere representation uniform contraction

depolarizing channel
Circuit representation of
c controlledswap
S

11 p107101 1177511
interpretation control is active with probability p
Kuans form of depolarizing channel
a normalized density
matrix s trig7 1
trick for
Stts YSY Zez
E 4
Eppls 1 ftp.t s XSX YsY zsz
E FII E F t E Y E E Z
 see Nielsen at Chuang seation 8.3.5
Amplitude damping
Exercise 13.1

eg photon in cavity
10 no photon
11 single photon

E o
E F
E 1 F 10
Interpretation 117 10 with probability p

Circuit representation

s Ads sin
with g
10 Ryle
see Exercise 13.1

Phase damping
Nielsen and Chung section 8.3.6
phase information is lost
e It f
model apply a phase kick Rzle
with random angle
FE.atistituti
a 10 bin
initial state 4 a
matrix after random phase kick
density
did
s In Raid 4 44 Rent e

off diagonal entries

4 8 are damped by e 8
 Kraus form E 1 E

e
r

unitary freedom as can show that channel is

equivalent to phase flip channel

E F 9 E Fa Z with a
1tfH
Application to Schrodinger's cat

14 dead 10 alive 11
It Itom ItItom

dead 10 dad 01 Idead 10 saliveka

g 4 441 alive 1 dead 01 alive 7 alive 411

off diagonalentries 0
due to decoherence

dead 10 dead 401

classical ensemble
ladive 1 Lative 471