A STUDY OF QMA, QCMA, AND THEIR ORACLE SEPARATION
by
MIR SHAHAB ALDIN RAZAVI HESSSABI
(Under the Direction of E. Rodney Canfield)
ABSTRACT
To compare the power of quantum and classical advice, Aharonov and Naveh introduced QMA and QCMA classes. Although QMA=QCMA is still an open problem, it is still possible to construct a relativized world in which these two classes can be compared. Aaronson and Kuperberg presented an oracle separation of these two classes. They showed that even if in addition to an m bit classical description of a marked state , a quantum black box which | recognizes is provided to any quantum algorithm, it still needs | queries before it can find the marked state . In this study a simple review of | MA and IP followed by an in depth and more intuitive description of
QMA/QCMA oracle separation is presented.
INDEX WORDS: QMA, QCMA, Oracle, Quantum Computing.
A STUDY OF QMA, QCMA, AND THEIR ORACLE SEPARATION
by
MIR SHAHAB ALDIN RAZAVI HESSABI
B.S., Shahid Beheshti University, Iran, 2006
M.S., University of Tehran, Iran, 2007
A Thesis Submitted to the Graduate Faculty of The University of Georgia in
Partial Fulfillment of the Requirements for the Degree
MASTER OF SCIENCE
ATHENS, GEORGIA
2009
© 2009
Mir Shahab Aldin Razavi Hessabi
All Rights Reserved
A STUDY OF QMA, QCMA, AND THEIR ORACLE SEPARATION
by
MIR SHAHAB ALDIN RAZAVI HESSABI
Major Professor: E. Rodney Canfield
Committee: Michael R. Geller Jeffrey Smith
Electronic Version Approved:
Maureen Grasso Dean of the Graduate School The University of Georgia December 2009 iv
ACKNOWLEDGEMENTS
I am heartily thankful to my supervisor, Rod E. Canfield, whose encouragement, guidance and support from the initial to the final level enabled me to develop an understanding of the subject.
Lastly, I offer my regards to all of those who supported me in any respect during the completion of the project.
Shahab Razavi
v
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ...... iv
LIST OF FIGURES ...... vii
CHAPTER
1 PRELIMINARIES ...... 1
1.1 INTRODUCTION ...... 1
1.2 QUBITS ...... 2
1.3 MULTIPLE QUBITS ...... 5
1.4 QUANTUM GATES ...... 6
1.5 THE DENSITY OPERATOR ...... 9
1.6 PURE AND MIXED STATES ...... 10
1.7 QUANTUM ORACLE ...... 10
1.8 QUANTUM TURING MACHINE ...... 11
2 CLASSICAL PROOF SYSTEMS ...... 12
2.1 INTERACTIVE DEFINITION OF NP ...... 12
2.2 INTERACTIVE PROOF SYSTEMS ...... 14
2.3 ARTHUR MERLIN GAME ...... 21
3 QUANTUM PROOF SYSTEMS ...... 23
3.1 QUANTUM INTERACTIVE PROOF SYSTEM ...... 23
3.2 FORMAL DEFINITION OF QIP ...... 25 vi
3.3 QUANTUM MERLIN ARTHUR ...... 27
3.4 QUANTUM CLASSICAL MERLIN ARTHUR ...... 29
4 ORACLE SEPERATION ...... 30
4.1 QUANTUM ORACLE SEPERATION OF QMA AND QCMA ...... 30
5 CONCLUSION ...... 36
5.1 GENERAL QUESTIONS ...... 36
GLOSSARY ...... 38
BIBLIOGRAPHY ...... 42
vii
LIST OF FIGURES
Page
Figure 1: Bloch Sphere ...... 4
Figure 2: Visualization of the Hadamard gate on the Bloch sphere ...... 8
Figure 3: Controlled NOT gate and its matrix representation ...... 9
Figure 4: The NP proof system ...... 13
Figure 5: An interactive pair of Turing machines ...... 14
Figure 6: Quantum circuit for a 3 message QIP ...... 27
1
CHAPTER 1
PRELIMINARIES
1.1 INTRODUCTION
It was thirty years ago that Richard Feynman raised the question, namely, how can we simulate quantum phenomena in a computer. Considering the nature of quantum mechanics such a simulation needs exponentially many bits to be able to represent a quantum system of size n. He then came up with the idea of a computational system based on quantum mechanics.
In 1985 David Deutsch [12] described a universal quantum Turing machine.
He proved that any physical system could be simulated if a two state system could be made to evolve by means of a set of simple operations. Due to the similarity between those operations and classical gates they are called quantum gates. In 1994 Peter Shor [13] found a quantum algorithm to factor large numbers in polynomial time.
A year later, Lov Grover [11] proved a tight bound of for the quantum √ search algorithm based on quantum oracles.
During the last 30 years the research on quantum computation has had its own ups and downs, but it never could stop researchers form exploiting the different aspects of this field. One of the ongoing research areas on QC is quantum complexity theory. Understanding the relations between different 2 classes and their power and weakness not only does give us a better
understanding of the field but also it helps us in developing more efficient
algorithms.
Two of these classes are Quantum Merlin Arthur and Quantum Classical
Merlin Arthur. The study of these two and their relation gives us a proof on whether the quantum advice contains more information than classical advice of
the same length. Although the equality of these two classes still remains an
open problem to be exploited, it is possible to consider these two classes in a
relativized world and compare their strength. Such a world is defined by the
use of quantum oracles. Aaronson and Kuperberg [15] have studied the
relation between these two classes and showed an oracle separation between
them. In this study we present an in depth and more intuitive version of their
proof.
The rest of this study is prepared as follows: in chapter one the essential
definitions and classes are given. In chapter 2 a brief review of classical
interactive proof system is presented. QMA and QCMA is then defined in
chapter 3. Chapter 4 contains the proof of QMA and QCMA oracle separation.
At the end in chapter 5, conclusion and open problems are given.
1.2 QUBITS
Like bits in classical computation, quantum bits or qubits for short are the
fundamental blocks of quantum computing. Qubits are actual physical objects,
but treating them as a mathematical concept with specific properties benefits 3 us in generalizing the theory of quantum computation and quantum
information in a way that our model does not rely on any specific system.
Like its classical cousin, a state is associated with each qubit. This state is
represented by “ ” notation called Dirac notation. But unlike bit which by its | definition can only be in state 0 or 1, qubit can be in a state other that or | 0 . That is, qubit can be in a superposition of states which is a linear | 1 combination of states:
| | 0 | 1 where and are complex numbers. The states and are the | 0 | 1 computational basis states which form an orthonormal basis for the state vector space.
Another difference between bit and qubit is how we measure its state. It is quite easy to determine the state of a classical bit, that is, to examine the bit and get 0 or 1 as the result. But rather remarkably it is not true for qubits. We cannot examine a qubit to determine the state of it, which is actually the value of and . As an alternative, quantum mechanics provides us with much more restricted information about the quantum state. This information is given to us by measuring the qubit and getting 0 with probability or 1 with probability | | . And since the probabilities must sum to one, we have . | | | | | | 1 Despite this restriction in determining the state of a qubit, before measuring or in another word collapsing a qubit into either 0 or 1, a qubit can exist in a continuum of states between and . These states can be manipulated and | 0 | 1 transformed in ways that result in an outcome of computation. 4
We can picture the state of a qubit with a geometric representation. Since
, we may rewrite the equation of state as | | | | 1