Errata list, Nielsen & Chuang. rrata/errata.html

Transcription

1 Errata list, Nielsen & Chuang rrata/errata.html

2 Part II, Nielsen & Chuang Quantum circuits (Ch 4) SK Quantum algorithms (Ch 5 & 6) Göran Johansson Physical realisation of quantum computers (Ch 7) Andreas Walther

3 Quantum computation and quantum information Chapter 4 Quantum circuits

4 Chapter 4 Quantum circuits Quantum circuits provide us with a language for describing quantum algorithms We can quantify the resources needed for a quantum algorithm in terms of gates, operations etc. It also provides a toolbox for algorithm design. Simulation of quantum systems

5 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

6 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

7 Why are there few quantum algorithms Algorithm design is difficult Quantum algorithms need to be better than classical algorithms to be interesting Our intuition works better for classical algorithms, making obtaining ideas about quantum algorithms still harder

8 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

9 Chapter 4, Quantum circuits 4.2 Single qubit operations We will analyse arbitrary single qubit and controlled single qubit operations

10 In quantum information data is represented by quantum bits (qubits) A qubit is a quantum mechanical systems with two states > and > that can be in any arbitrary superposition of those states +

11 Qubit representation e i cos e i sin 2 2

12 The Bloch sphere i i e cos e sin 2 2

13 Single qubit gates Single qubit gates, U, are unitary

14 .3. Single qubit gates

15 Fig 4.2, page 77

16 Single qubit gates on Bloch sphere X-gate 8 rotation around x-axis Y-gate 8 rotation around y-axis Z-gate 8 rotation around z-axis

17 The Bloch sphere i i e cos e sin 2 2

18 Fig 4.2, page 77

19 Single qubit gates on Bloch S-gate sphere 9 rotation around z-axis T-gate 45 rotation around z-axis H-gate 9 rotation around the y-axis followed by 8 rotation around the x-axis

20 Rotation an arbitrary angle around axes x, y and z

21 Rotation an arbitrary angle around axes x, y and z H-gate 9 rotation around y-axis followed by 8 rotation around x-axis Show this from Eqs 4.4 and 4.5 (does not commute)

22 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

23 CNOT is a MODULO-2 addition of the control bit to the target bit c t c t c

24 Fig 4.4, control-u c t c U c t

25 Controlled arbitrary rotation c t c U c t Lets start with inputs c> = > and t> = >, what is the output? Now we choose c> = > and t> = >, what is the output?

26 Controlled arbitrary rotation

27 Preparing for logical gates Page 76

28 Arbitrary single qubit operation in terms of z and y rotations (page 75)

29 Preparing for logical gates Page 76

30 Controlled arbitrary rotation

31 Notation Computational basis states (page 22) A quantum circuit operating on n qubits acts in a 2 n -dimensional complex Hilbert space. The computational basis state are product states x,...,x n > where x i =,. For example a two qubit state x > x 2 > in the computational basis is expressed as > = > + > + > + d > Where we have the basis states for the tensor product of the x > x 2 > basis states

32 .4.3 Deutsch s algorithm Calculates f ( ) f () Read Deutsch-Josza algorithm before Görans lectures on Thursday

33 Hadamard gates, exercise 4.6

34 Exercise 4.6, page Matrix for H-gate on x 2 Matrix for H-gate on x 2 2

35 H-gate is unitary Matrix multiplication H H=I 2 2

36 Exercise 4.2

37 Multiplying the Hadamard gates We can just extend exercise 4.6 by multiplying the two Hadamard gates 2 2 2

38 Exercise 4.2 Multiplying Hadamard and CNOT gates give Answer: a CNOT gate where the first bit is the control bit 2 2 What operation is this matrix carrying out?

39 Two control bits

40 Many control bits

41 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

42 .3.6 Bell states (EPR states) An entangled state basis

43 .3.7 Quantum teleportation Quantum teleportation

44 Principle of deferred measurement Measurements followed by classically controlled operations can always be replaced by conditional quantum operations Why is Alice not transmitting information faster than the speed of light through the measurment now when no classical operations need to be carried out by Bob?

45 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

46 Universal quantum gates (4.5) A set of gates is universal if any unitary operation can be approximated to arbitrary accuracy by quantum circuits only involving these gates

47 CNOT, Hadamard and T-gates form a universal set. An arbitrary unitary operator can be expressed as a product of unitary operators acting on two computational basis states ( subsection 4.5.). 2. An arbitrary unitary operator acting on two computational basis states can be expressed as a product of single qubit operations and CNOT gates. (page 92-93) (subsection 4.5.2, exercise 4.39)

48 Exercise 4.93, Page 93

49 CNOT, Hadamard and T-gates form a universal set. An arbitrary unitary operator can be expressed as a product of unitary operators acting on two computational basis states ( subsection 4.5.). 2. An arbitrary unitary operator acting on two computational basis states can be expressed as a product of single qubit operations and CNOT gates. (page 92-93) (subsection 4.5.2, exercise 4.39) 3. Single qubit operations may be approximated to arbitrary accuracy using Hadamard and T gates (subsection 4.5.3)

50 Single qubit operations may be approximated to arbitrary accuracy using H- and T-gates (subsection 4.5.3) From the corrected version of Eq. 4.3 page 76 we see that (exercise 4.): If m and n are non parallel vectors in three dimensions any arbitrary single qubit unitary operation, U may be written as U e i R n ( ) Rm( ) Rn ( 2) Rm( 2)... It is shown on page 96 that rotations of arbitrary angles can be carried out around two different axes using combinations of H- and T-gates

51 Approximating arbitrary unitary gates (4.5.3 & 4.5.4) In order to approximate a quantum circuit consisting of m CNOT and single qubit gates with an accuracy e, only about O[m*log(m/e)] gate operations are required (Solovay-Kitaev theorem). This does not sound too bad! The problem is that of the order 2 2n gates are required to implement an arbitrary n-qubit unitary operation (see page 9-93), thus this is a computationally hard problem.

52 Quantum Computational Complexity Relation to classical complexity classes BPP is thought to be subset of BQP, enabling QC to solve some problems more efficient than classical computers P BPP BQP PSPACE BQP is a subset of PSPACE a) Any problem consuming polynomial time can consume a maximal polynomial amount of space b) A problem inefficiently solvable regarding time might still be solved efficiently regarding space (usage of the same set of gates)

53 Complexity classes

54 The power of quantum computation NP complete problems are a subgroup of NP If any NP-complete problem has a polynomial time solution then P=NP Factoring is not known to be NP-complete Quantum computers are known to solve all problems in P efficiently but cannot solve problems outside PSPACE efficiently If quantum computers are proved to be more efficient than classical computers P=/=PSPACE Quantum computers are (supposedly) experimentally realizable, thus this is not just a mathematical curiosity but also tells us something about how the limits and power of computation in general

55 Quantum computing changes the landscape of computer science QC algorithms do not violate the Church- Turing thesis: any algorithmic process can be simulated using a Turing machine QC algorithms challenge the strong version of the Church-Turing thesis If an algorithm can be performed at any class of hardware, then there is an equivalent efficient algorithm for a Turing machine

56 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

57 The quantum circuit model of computation Properties: Quantum computer may be a hybrid of quantum and classical resources to maximize efficiency A quantum circuit operates on n qubits spanning a 2 n dimensional state space. The product states of the form x,x 2,..,x n > ;x i ={,} are the computational basis states Basic steps: Preparation Computation Measurement

58 The quantum circuit model of computation Maybe it would be better that the computational basis states are entangled Maybe the measurements should not be carried out in the computational basis It is not known whether the quantum circuit model constitute an optimum quantum computer language

59 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement 4.5 Universal quantum gates 4.6 Circuit model summary 4.7 Simulation of quantum systems

60 The quantum-mechanical computation of one molecule of methane requires 42 grid points. Assuming that at each point we have to perform only elementary operations, and the computation is performed at the extremely low temperature T=3-3 K, we would still have to use all the energy produced on Earth during the last century Page 24

61 Quantum simulation Feynman 982 Chemistry (large molecules, reactions) Biology (even larger molecules) Storing the amplitudes of 5 connected qubits may be possible on the supercomputers in the second half of the 2st century provided Moore s law continues to hold Simulation of the properties of new synthesized molecules

62 Simulation of quantum systems i d dt H With n qubits there are 2 n different differential equations that must be solved The equation above has the formal solution i Ht ( t) e ()

63 Simulation of quantum systems While quantum computers are hoped to solve general calculations efficiently, another field in which they are hoped to succeed is the simulation of specific physical systems described by a Hamiltonian. However the physical system must be approximated efficiently, to do so: y()> y ()> Approximating Approximating the the Operator state Discretizing The continuous of the function differential is discretised equations to begins arbitrary with choosing precision an using appropriate a finite set Dt. of It basis has to vectors fulfill the demands set by the maximum Exp[-iHt] error. y(t)> c(x) (x)dx ' The approximation of the differential operator is a cthree kk step process. k The set of basis vectors must be chosen such that for First, any given if possible time, separate the approximated H into a set state of has Hamiltonians, to be equal to H the i, which original act state on a within maximal a given constant error number tolerance of particles. (next neighbor interaction, etc). Secondly, write the effect of H on the system as time evolves as a product U of the effect of the individual y (t)> H i. Thirdly, the effect of H i is written in terms of a quantum circuit.

64 Simulation of quantum systems Approximating the Operator The evolution operator is of the form Exp[ iht] If By possible varying H n is we broken can obtain down a into product a sum representation of local interactions of the evolution operator within any given error H Exp (Trotter formula) H k 2 i( Hk H2) Dt Exp ihdtexp ih 2Dt O( Dt ) If the sub Hamiltonians do not commute it follows that Now that H is broken down into a product of local Hamiltonians which may be written on a quantum circuit form (Exercise 4.5). The product of the circuit corresponds to a unitary operator U. Exp[ i(h H...)t] Exp[ ih t]exp[ ih t]... U 2 2 ih t Exp ih 2 t Exp ih n We now introduce the Trotter formula Exp[ i( A ( t t Exp... ( t Dt) f ) B) jt ] U U ( t ( t) lim ) ; t n f Exp iat / nexp ibt / n jdt n

65 Exercise 4.5, page 2

66 Chapter 4, Quantum circuits 4. Quantum algorithms 4.2 Single qubit operations 4.3 Controlled operations 4.4 Measurement

67 Chapter 4, Quantum circuits 4.5 Universal quantum gates Single qubit and CNOT gates are universal A discrete set of universal operators Approximating arbitrary unitary gates is hard Quantum computational complexity 4.6 Circuit model summary 4.7 Simulation of quantum systems