Altenkirch Slides

Functional Quantum
Programming
Thorsten Altenkirch
University of Nottingham
based on joint work with Jonathan Grattage
supported by EPSRC grant GR/S30818/01
Tallinn Feb 06 – p.1/44

Background

Background
Simulation of quantum systems is expensive:
PSPACE complexity for polynomial circuits.

Background
Feynman: Can we exploit this fact to perform
computations more efficiently?

Background
Shor: Factorisation in quantum polynomial time.

Background
Grover: Blind search in


Background

Can we build a quantum computer?

Background

yes We can run quantum algorithms.

Background

yes We can run quantum algorithms.
no Nature is classical after all!

The quantum software crisis

Quantum algorithms are usually presented

using the circuit model.


Nielsen and Chuang, p.7, Coming up with
good quantum algorithms is hard.


Nielsen and Chuang, p.7, Coming up with
good quantum algorithms is hard.
Richard Josza, QPL 2004: We need to
develop quantum thinking!

QML

QML
QML: a functional language for quantum computations
on finite types.

QML
on finite types.
Quantum control and quantum data.

QML
on finite types.
Design guided by semantics

QML
on finite types.
Analogy with classical computation
Finite classical computations

Finite quantum computations


QML
on finite types.


Important issue: control of decoherence

QML
on finite types.


Important issue: control of decoherence

Compiler under construction (Jonathan)

Example: Hadamard operation

Matrix


QML
Matrix

,+

!" "!
#$
#
')(&
"
%
*
'() &
"
*

] /.0- /.0-
K CDB
3241 3241
E
H FG
I />5 765
F b
J :98
@ @ ?
A
C
E KL 9 8
DaB ]C \ M
C
OGN
O
O Q:P < ;
=
Q P UTSR
XZ W
H SU R H U
Deutsch algorithm
O V
C C XYZ W
E KL E KL
DaB DaB > U
C ]C \ C ]C \
?
[
O N
G N N O N
G
^
_ ^
_
Q P Q P
SU R ` ` SU R
O O
V V
Q:P Q:P
^
_ ^
_
SUTR TSUR
` `
O O V O V O
XZ W XZ W XZ W XZ W
U U U U
[ [ [ [
Overview
1. Finite classical computation

2. Finite quantum computation
3. QML basics
4. Compiling QML
5. Conclusions and further work

1. Semantics

3. QML basics
4. Compiling QML

Something we know well . . .

Start with classical computations on finite types.

Quantum mechanics is time-reversible. . .

. . . hence quantum computation is based on reversible
operations.

operations.
However: Newtonian mechanics, Maxwellian
electrodynamics are also time-reversible. . .

operations.
However: Newtonian mechanics, Maxwellian
electrodynamics are also time-reversible. . .
. . . hence classical computation should be based on
reversible operations.

Classical computation ( )

Given finite sets (input) and (output):
d
e

h
f

d
e

h
f
g
a finite set of initial heaps ,

d
e

h
f
g
an initial heap ,
kji

d
e

h
f
g
an initial heap ,
kji
a finite set of garbage states ,

d
e

h
f
g
an initial heap ,
kji
a finite set of garbage states ,

a bijection ,
j
l
m

Composing classical computations

o
n
>> 88
n
n
>> 88
>> 88
>>
8
o
o
o

n

o
n
>> 88
n
n
>> 88
>> 88
>>
8
o
o
o

n
Theorem:
U U U
$
$
%
p $
%
pq
q
Extensional equality

A classical computation
$
%
i
q
r
r
r
induces a function U by
j
q
/
l
s
O
{zy
x w
|}
vut

/
U

A classical computation
$
%
i
q
r
r
r
induces a function U by
j
q
/
l
s
O
{zy
x w
|}
vut

/
U
n
We say that two computations are
extensionally equivalent, if they give rise to
the same function.

Extensional equality . . .
Theorem:
U U U
$
%
p $
%
pq
q
Theorem:
U U U
$
%
p $
%
pq
q
Hence, classical computations upto
extensional equality give rise to the category
.

Theorem:
U U U
$
%
p $
%
pq
q
.
Theorem: Any function on finite
j
sets can be realized by a computation.
r

Theorem:
U U U
$
%
p $
%
pq
q
.
Theorem: Any function on finite
j
sets can be realized by a computation.
r
Translation for Category Theoreticians:

U is full and faithful.

Example :

function
%

j
~
r
$
%

~

Example :

function
%

j
~
r
$
%

~
r
computation

e

Example :
function
%
j

r
$
%

r

function
j
$ $ j
computation
$

r r $

% % r $
Example :

r
%
$ $
% r
$
%
r r

% r
% %
'!"#&%$



3. QML basics
4. Compiling QML

Linear algebra revision

Given a finite set (the base)

is a Hilbert space.



is a Hilbert space.

Linear operators:
induces
j
.

j
we write
j


is a Hilbert space.

Linear operators:
induces
j
.

j
we write
j

Norm of a vector:
,

%$
j %



is a Hilbert space.

Linear operators:
induces
j
.

j
we write
j

Norm of a vector:
,

%$
j %

Unitary operators:
A unitary operator is a linear iso-
j
unitary
morphism that preserves the norm.

Basics of quantum computation

A pure state over is a vector with

j

unit norm .


A pure state over is a vector with

j

unit norm .

A reversible computation is given by a
unitary operator .

unitary

Quantum computations ( )

d
e

h
f

d
e

h
f
g
a finite set , the base of the space of initial
heaps,

d
e

h
f
g
heaps,
a heap initialisation vector ,

ji

d
e

h
f
g
heaps,

ji
a finite set , the base of the space of
garbage states,

d
e

h
f
g
heaps,

ji
a finite set , the base of the space of
garbage states,
a unitary operator .
j

unitary
Composing quantum computations

Composing quantum computations
o
n
>> 88
n
n
>> 88
>> 88
>>
8
o
o
o

n

Semantics of quantum computations. .

. . . is a bit more subtle.


There is no (sensible) operator on vector
spaces, replacing .
l
~


There is no (sensible) operator on vector
spaces, replacing .
l
~
Indeed: Forgetting part of a pure state
results in a mixed state.

Density matrizes

Density matrizes
Mixed states can be represented by density

matrizes .
j

Density matrizes

matrizes .
j

Eigenvalues represent probabilities

System is in state with prob.


Density matrizes

matrizes .
j

Eigenvalues represent probabilities

System is in state with prob.

Eigenvalues have to be positive and their sum

(the trace) is .


Example: forgetting a qbit

EPR is represented by
:
j

<
<

<
<

EPR is represented by
:
j

<
<

<
<

$

%

Example: forgetting a qbit . . .
After measuring one qbit we obtain

:
j

<

<

Example: forgetting a qbit . . .
After measuring one qbit we obtain

:
j

<

<

#

#

#

#


Superoperators

Superoperators
Morphisms on density matrizes are called

superoperators, these are linear maps, which
are
completely positive, and
trace preserving

Superoperators
Morphisms on density matrizes are called

superoperators, these are linear maps, which
are
completely positive, and
trace preserving
Every unitary operator
¡
gives rise to a
superoperator .

Superoperators. . .
There is an operator
'&
¢j

super

z
called partial trace.

Superoperators. . .
There is an operator
'&
¢j

super

z
called partial trace.
E.g. is
¤£('&
¥ j

super
¥ £

z
represented by a matrix.
§
¦

Semantics

Semantics
Every quantum computation gives rise to a
q
superoperator U
j
q

super
s ¨
O
|
« ª©
¬
x

/
U

Semantics
Every quantum computation gives rise to a
q
superoperator U
j
q

super
s ¨
O
|
« ª©
¬
x

/
U
n
Theorem: Every superoperator

super
(on finite Hilbert spaces) comes from a quantum
computation.

Classical vs quantum

classical ( ) quantum ( )


finite sets


finite sets finite dimensional Hilbert space


cartesian product ( )


cartesian product ( ) tensor product ( )
¯
®


¯
®
bijections


¯
®
bijections unitary operators


¯
®
functions


¯
®
functions superoperators


finite sets finite dimensional Hilbert spaces
¯
®
injective functions ( M )



¯
®
injective functions ( M ) isometries ( )
M


¯
®
M

projections


¯
®
M

projections partial trace

Decoherence

Decoherence

'!"#&%$

e
° e
±

Decoherence

'!"#&%$

e
° e
±
Classically
²
p
~

Decoherence

'!"#&%$

e
° e
±
Classically
²
p
~

Quantum

Decoherence

'!"#&%$

e
° e
±
Classically
²
p
~

Quantum

input:

#

#

*


Decoherence

'!"#&%$

e
° e
±
Classically
²
p
~

Quantum

input:

#

#

*

output:
#

*
#

*


3. QML basics

3. QML basics
4. Compiling QML

QML basics

QML basics
QML is a first order functional languages, i.e.

programs are well-typed expressions.

QML basics

QML types are
r ´

r

QML basics

QML types are
r ´

r
Qbytes
.
µ


QML basics . . .
A QML program is an expression in a context

of typed variables, e.g.
¹
¶·¸

¹
¶·¸

!"

, +
'() &

"


QML basics . . .
A QML program is an expression in a context

of typed variables, e.g.
¹
¶·¸

¹
¶·¸

!"

, +
'() &

"

We can compile QML programs into quantum
computations (i.e. quantum circuits).

QML basics . . .
Forgetting variables has to be explicit.

QML basics . . .

E.g.
»¹
¶º

¹$
%
¶º

r
is illegal,

QML basics . . .

E.g.
»¹
¶º

¹$
%
¶º

r
is illegal,
but
»¹
¶º

¹$
*
¶º

r
is ok.

QML basics . . .
There are two different if-then-else constructs.

QML basics . . .

¼

¼

')(&
"

, +

!"

is just the identity,

QML basics . . .

¼

¼

')(&
"

, +

!"

is just the identity, but
½¾
º

½¾
º

')(&
"

,+

!"

introduces a measurement (end hence

decoherence).
QML basics . . .
Using is only allowed, if the branches are

orthogonal, i.e. observable different.

QML basics . . .

º
¿
ÀÁ

$

º
¿
ÀÁ
¿

r
$
%

r
$
%
, +

r
is illegal,

QML basics . . .

º
¿
ÀÁ

$

º
¿
ÀÁ
¿

r
$
%

r
$
%
, +

r
is illegal, but
%
º
¿
ÀÁ

$

º
¿
ÀÁ
¿

r
$
%
%

'() &
"

r
r
$
%
%

, +

!"

r
is ok.
QML basics . . .
We can introduce superpositions, e.g.

#$
%
*

!"
')(&
"

#
*

,+
!"
')(&

"

QML basics . . .
We can introduce superpositions, e.g.

#$
%
*

!"
')(&
"

#
*

,+
!"
')(&

"
However, the terms in the superposition have
to be orthogonal.

4. Compiling QML

3. QML basics
4. Compiling QML

Compilation

Compilation
Correct QML programs are defined by typing

rules, e.g.
Â
ÃÄ

ÆÃ

Å
r ³

r
ÉÈÇ
"
Í$
%
ËÊÃ
Â
Ä
ÏÐÎ
r
Å
Æ
Ì

Compilation
Correct QML programs are defined by typing

rules, e.g.
Â
ÃÄ

ÆÃ

Å
r ³

r
ÉÈÇ
"
Í$
%
ËÊÃ
Â
Ä
ÏÐÎ
r
Å
Æ
Ì

For each rule we can construct a quantum
computation, i.e. a circuit.

Â
-elim
Â
ÃÊ r
ÃÄ

Ì
Í$
r ³ ³
Å
r
Å ´
% ´
Ä ÆÃ

ÏÎ
Æ

"
È Ç
-elim
Â
ÃÄ

ÆÃ

Å
r ³

r
È Ç
"
Í$
%
Â
Ä
ÏÎ
ÃÊ
r
Å
Æ
Ì

Â
Ø
×
;;
;;
Ô ÓÒ Ñ

Ö
Õ
;
Ø
Û
Ú
Ý
z
ß
Þ
Ü
;; ;;
Ü
Ü
;; ;;
;; ;;

Ù
Ù
Compiler

Compiler
A compiler is currently being implemented by

my student Jonathan Grattage (in Haskell).

Compiler

The output of the compiler are quantum
circuits which can be simulated by a quantum
circuit simulator.

Compiler

The output of the compiler are quantum
circuits which can be simulated by a quantum
circuit simulator.
Amr Sabry and Juliana Vizotti (Indiana
University) embarked on an independent
implementation of QML based on our paper.

5. Conclusions
1. Semantics of finite classical and quantum

computation
2. QML basics
3. Compiling QML

Conclusions

Conclusions
Our semantic ideas proved useful when designing a
quantum programming language, analogous concepts
are modelled by the same syntactic constructs.

Conclusions
Our analysis also highlights the differences between
classical and quantum programming.

Conclusions
We have developed an algebra of quantum programs
which for pure programs is complete wrt the semantics
and a normalisation algorithm.

Conclusions
We have developed an algebra of quantum programs
which for pure programs is complete wrt the semantics
and a normalisation algorithm.
Quantum programming introduces the problem of
control of decoherence, which we address by making
forgetting variables explicit and by having different
if-then-else constructs.
Further work

Further work
We have to analyze more quantum programs to
evaluate the practical usefulness of our approach.

Further work
We should be able to extend our algebra and
normalisation to the full language (including
measurements).

Further work
measurements).
Are we able to come up with completely new
algorithms using QML?

Further work
measurements).
How to deal with higher order programs?

Further work
measurements).
How to deal with infinite datatypes?

Further work
measurements).
How to deal with infinite datatypes?
Investigate the similarities/differences between FCC
and FQC from a categorical point of view.
The end
Thank you for your attention.

Papers, available from
//www.cs.nott.ac.uk/˜txa/publ/
A functional quantum programming language LICS 2005
with J.Grattage
Structuring Quantum Effects: Superoperators as Arrows
MFCS 2006
with J.Vizzotto and A.Sabry
An Algebra of Pure Quantum Programming QPL 2005
with J.Grattage, J.Vizzotto and A.Sabry Tallinn Feb 06 – p.44/44

Altenkirch Slides

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Altenkirch Slides

Uploaded by

Copyright:

Available Formats

Functional Quantum

Tallinn Feb 06 – p.1/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.2/44

Tallinn Feb 06 – p.3/44

Quantum algorithms are usually presented

Tallinn Feb 06 – p.3/44

Quantum algorithms are usually presented

Tallinn Feb 06 – p.3/44

Quantum algorithms are usually presented

Tallinn Feb 06 – p.3/44

Tallinn Feb 06 – p.4/44

Tallinn Feb 06 – p.4/44

Tallinn Feb 06 – p.4/44

Tallinn Feb 06 – p.4/44

Finite quantum computations

Tallinn Feb 06 – p.4/44

Finite quantum computations

Important issue: control of decoherence

Tallinn Feb 06 – p.4/44

Finite quantum computations

Important issue: control of decoherence

Tallinn Feb 06 – p.4/44

Tallinn Feb 06 – p.5/44

Tallinn Feb 06 – p.5/44

Tallinn Feb 06 – p.5/44

1. Finite classical computation

Tallinn Feb 06 – p.7/44

1. Finite classical computation

Tallinn Feb 06 – p.8/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.9/44

Tallinn Feb 06 – p.10/44

Given finite sets (input) and (output):

Tallinn Feb 06 – p.10/44

Given finite sets (input) and (output):

Tallinn Feb 06 – p.10/44

Given finite sets (input) and (output):

Tallinn Feb 06 – p.10/44

Given finite sets (input) and (output):

a finite set of garbage states ,

Tallinn Feb 06 – p.10/44

Given finite sets (input) and (output):

a finite set of garbage states ,

Tallinn Feb 06 – p.10/44

Tallinn Feb 06 – p.11/44

Tallinn Feb 06 – p.11/44

Tallinn Feb 06 – p.12/44

Tallinn Feb 06 – p.12/44

Tallinn Feb 06 – p.12/44

Tallinn Feb 06 – p.13/44

Tallinn Feb 06 – p.13/44

Translation for Category Theoreticians:

Tallinn Feb 06 – p.13/44