Mutually unbiased bases explained

In quantum information theory, a set of bases in Hilbert space C^d are said to be mutually unbiased if when a system is prepared in an eigenstate of one of the bases, then all outcomes of the measurement with respect to the other basis are predicted to occur with an equal probability inexorably equal to 1/d.

Overview

The notion of mutually unbiased bases was first introduced by Julian Schwinger in 1960,^[1] and the first person to consider applications of mutually unbiased bases was I. D. Ivanovic^[2] in the problem of quantum state determination.

Mutually unbiased bases (MUBs) and their existence problem is now known to have several closely related problems and equivalent avatars in several other branches of mathematics and quantum sciences, such as SIC-POVMs, finite projective/affine planes, complex Hadamard matrices and more [see section: [[#Related problems|Related problems]]].

MUBs are important for quantum key distribution, more specifically in secure quantum key exchange.^[3] MUBs are used in many protocols since the outcome is random when a measurement is made in a basis unbiased to that in which the state was prepared. When two remote parties share two non-orthogonal quantum states, attempts by an eavesdropper to distinguish between these by measurements will affect the system and this can be detected. While many quantum cryptography protocols have relied on 1-qubit technologies, employing higher-dimensional states, such as qutrits, allows for better security against eavesdropping.^[3] This motivates the study of mutually unbiased bases in higher-dimensional spaces.

Other uses of mutually unbiased bases include quantum state reconstruction,^[4] quantum error correction codes,^[5] ^[6] detection of quantum entanglement,^[7] ^[8] and the so-called "mean king's problem".^[9] ^[10]

Definition and examples

\{|e_1\rangle,...,|e_d\rangle\}

and

\{|f_1\rangle,...,|f_d\rangle\}

in Hilbert space C^d are said to be mutually unbiased, if and only if the square of the magnitude of the inner product between any basis states

|e_j\rangle

and

|f_k\rangle

equals the inverse of the dimension d:^[11]

|\langlee_j|f_k\rangle|²=

	1
	d

, \forallj,k\in\{1,...,d\}.

These bases are unbiased in the following sense: if a system is prepared in a state belonging to one of the bases, then all outcomes of the measurement with respect to the other basis are predicted to occur with equal probability.

Example for d = 2

The three bases

M₀=\left\{|0\rangle,|1\rangle\right\}

M₁=\left\{

	\|0\rangle+\|1\rangle
	\sqrt{2

},\frac

1 \rangle

\right\}

M₂=\left\{

	\|0\rangle+i\|1\rangle
	\sqrt{2

},\frac

1 \rangle

\right\} provide the simplest example of mutually unbiased bases in C². The above bases are composed of the eigenvectors of the Pauli spin matrices

\sigma_z,\sigma_x

and their product

\sigma_x\sigma_z

, respectively.

Example for d = 4

For d = 4, an example of d + 1 = 5 mutually unbiased bases where each basis is denoted by M_j, 0 ≤ j ≤ 4, is given as follows:^[12]

M₀=\left\{(1,0,0,0),(0,1,0,0),(0,0,1,0),(0,0,0,1)\right\}

M₁=\left\{

	1	(1,1,1,1),
	2

	1	(1,1,-1,-1),
	2

	1	(1,-1,-1,1),
	2

	1
	2

(1,-1,1,-1)\right\}

M₂=\left\{

	1	(1,-1,-i,-i),
	2

	1	(1,-1,i,i),
	2

	1	(1,1,i,-i),
	2

	1
	2

(1,1,-i,i)\right\}

M₃=\left\{

	1
	2

(1,-i,-i,-1),

	1	(1,-i,i,1),
	2

	1	(1,i,i,-1),
	2

	1
	2

(1,i,-i,1)\right\}

M₄=\left\{

	1	(1,-i,-1,-i),
	2

	1	(1,-i,1,i),
	2

	1	(1,i,-1,i),
	2

	1
	2

(1,i,1,-i)\right\}

Existence problem

Let

ak{M}(d)

denote the maximum number of mutually unbiased bases in the d-dimensional Hilbert space C^d. It is an open question^[13] how many mutually unbiased bases,

ak{M}(d)

, one can find in C^d, for arbitrary d.

In general, if

	n₁
p
	1

	n₂
p
	2

…

	n_k
p
	k

is the prime-power factorization of d, where

	n₁
p
	1

	n₂
p
	2

	n_k
< … <p
	k

then the maximum number of mutually unbiased bases which can be constructed satisfies^[11]

	n₁
p
	1

+1\leak{M}(d)\led+1.

It follows that if the dimension of a Hilbert space d is an integer power of a prime number, then it is possible to find d + 1 mutually unbiased bases. This can be seen in the previous equation, as the prime number decomposition of d simply is

d=pⁿ

. Therefore,

ak{M}(pⁿ⁾=pⁿ+1.

Thus, the maximum number of mutually unbiased bases is known when d is an integer power of a prime number, but it is not known for arbitrary d.

The smallest dimension that is not an integer power of a prime is d = 6. This is also the smallest dimension for which the number of mutually unbiased bases is not known. The methods used to determine the number of mutually unbiased bases when d is an integer power of a prime number cannot be used in this case. Searches for a set of four mutually unbiased bases when d = 6, both by using Hadamard matrices^[11] and numerical methods^[14] ^[15] have been unsuccessful. The general belief is that the maximum number of mutually unbiased bases for d = 6 is

ak{M}(6)=3

.^[11]

Search methods

Weyl group method

Let

\hat{X}

and

\hat{Z}

be two unitary operators in the Hilbert space C^d such that

\hat{Z}\hat{X}=\omega\hat{X}\hat{Z}

for some phase factor

\omega

. If

\omega

is a primitive root of unity, for example

\omega\equiv

	2\pii
	d

then the eigenbases of

\hat{X}

and

\hat{Z}

are mutually unbiased.

By choosing the eigenbasis of

\hat{Z}

to be the standard basis, we can generate another basis unbiased to it using a Fourier matrix. The elements of the Fourier matrix are given by

F_ab=\omega^ab,0\lea,b\leN-1

Other bases which are unbiased to both the standard basis and the basis generated by the Fourier matrix can be generated using Weyl groups.^[11] The dimension of the Hilbert space is important when generating sets of mutually unbiased bases using Weyl groups. When d is a prime number, then the usual d + 1 mutually unbiased bases can be generated using Weyl groups. When d is not a prime number, then it is possible that the maximal number of mutually unbiased bases which can be generated using this method is 3.

Unitary operators method using finite fields

When d = p is prime, we define the unitary operators

\hat{X}

and

\hat{Z}

\hat{X}=

	d-1
\sum
	k=0

|k+1\rangle\langlek|

\hat{Z}=

	d-1
\sum
	k=0

\omega^k|k\rangle\langlek|

where

\{|k\rangle|0\lek\led-1\}

is the standard basis and

\omega=

	2\pii
	d

is a root of unity.

Then the eigenbases of the following d + 1 operators are mutually unbiased:

\hat{X},\hat{Z},\hat{X}\hat{Z},\hat{X}\hat{Z}^2,\ldots,\hat{X}\hat{Z}^d-1.

For odd d, the t-th eigenvector of the operator

\hat{X}\hat{Z}^k

is given explicitly by^[13]

	k
\|\psi
	t

\rangle=

	1
	\sqrtd

	d-1
\sum
	j=0

	tj+kj²
\omega

|j\rangle.

When

d=p^r

is a power of a prime, we make use of the finite field

F_d

to construct a maximal set of d + 1 mutually unbiased bases. We label the elements of the computational basis of C^d using the finite field:

\{|a\rangle|a\inF_d\}

We define the operators

\hat{X_a}

and

\hat{Z_b}

in the following way

\hat{X_a}=

\sum
	c\inF_d

|c+a\rangle\langlec|

\hat{Z_b}=

\sum
	c\inF_d

\chi(bc)|c\rangle\langlec|

where

\chi(\theta)=\exp\left[

	2\pii
	p

\left(\theta+\theta^p+

	p²
\theta

+ … +

	p^r-1
\theta

\right)\right],

is an additive character over the field and the addition and multiplication in the kets and

\chi( ⋅ )

is that of

F_d

Then we form d + 1 sets of commuting unitary operators:

\{\hat{Z_s}|s\inF_d\}

and

\{\hat{X_s}\hat{Z_sr

} | s \in \mathbb_d \} for each

r\inF_d

The joint eigenbases of the operators in one set are mutually unbiased to that of any other set.^[21] We thus have d + 1 mutually unbiased bases.

Hadamard matrix method

Given that one basis in a Hilbert space is the standard basis, then all bases which are unbiased with respect to this basis can be represented by the columns of a complex Hadamard matrix multiplied by a normalization factor. For d = 3 these matrices would have the form

	1
	\sqrt{d

}\begin 1 & 1 & 1 \\ e^ & e^ & e^ \\ e^ & e^ & e^\endThe problem of finding a set of k+1 mutually unbiased bases therefore corresponds to finding k mutually unbiased complex Hadamard matrices.^[11]

An example of a one parameter family of Hadamard matrices in a 4-dimensional Hilbert space is

H_4(\phi)=

	1
	2

\begin{bmatrix} 1&1&1&1\\ 1&e^i\phi&-1&-eⁱ\\ 1&-1&1&-1\\ 1&-e^i\phi&-1&e^i\phi\end{bmatrix}

Entropic uncertainty relations

There is an alternative characterization of mutually unbiased bases that considers them in terms of uncertainty relations.^[22]

Entropic uncertainty relations are analogous to the Heisenberg uncertainty principle, and Hans Maassen and J. B. M. Uffink^[23] found that for any two bases

B₁=\{|a_i

	d
\rangle
	i=1

and

B₂=\{|b_j\rangle

	d

	j=1

H
	B₁

H
	B₂

\geq-2logc.

where

c=max|\langlea_j|b_k\rangle|

and

H
	B₁

and

H
	B₂

is the respective entropy of the bases

B₁

and

B₂

, when measuring a given state.

Entropic uncertainty relations are often preferable^[24] to the Heisenberg uncertainty principle, as they are not phrased in terms of the state to be measured, but in terms of c.

In scenarios such as quantum key distribution, we aim for measurement bases such that full knowledge of a state with respect to one basis implies minimal knowledge of the state with respect to the other bases. This implies a high entropy of measurement outcomes, and thus we call these strong entropic uncertainty relations.

For two bases, the lower bound of the uncertainty relation is maximized when the measurement bases are mutually unbiased, since mutually unbiased bases are maximally incompatible: the outcome of a measurement made in a basis unbiased to that in which the state is prepared in is completely random. In fact, for a d-dimensional space, we have:^[25]

H
	B₁

H
	B₂

\geqlog(d)

for any pair of mutually unbiased bases

B₁

and

B₂

. This bound is optimal:^[26] If we measure a state from one of the bases then the outcome has entropy 0 in that basis and an entropy of

log(d)

in the other.

If the dimension of the space is a prime power, we can construct d + 1 MUBs, and then it has been found that^[27]

	d+1
\sum
	k=1

H
	B_k

\geq(d+1)log\left(

	d+1
	2

\right)

which is stronger than the relation we would get from pairing up the sets and then using the Maassen and Uffink equation. Thus we have a characterization of d + 1 mutually unbiased bases as those for which the uncertainty relations are strongest.

Although the case for two bases, and for d + 1 bases is well studied, very little is known about uncertainty relations for mutually unbiased bases in other circumstances.^[27] ^[28]

When considering more than two, and less than

d+1

bases it is known that large sets of mutually unbiased bases exist which exhibit very little uncertainty.^[29] This means merely being mutually unbiased does not lead to high uncertainty, except when considering measurements in only two bases. Yet there do exist other measurements that are very uncertain.^[27] ^[30]

Infinite dimension Hilbert spaces

	b
\|\psi
	s

\rangle

and

	b'
\|\psi
	s'

\rangle

are said to be mutually unbiased if^[31]

|\langle

	b
\psi
	s

	b'
\psi
	s'

\rangle|²=k>0,s,s'\inR

For the generalized position and momentum eigenstates

|q\rangle,q\inR

and

|p\rangle,p\inR

, the value of k is

|\langleq|p\rangle|²=

	1
	2\pi\hbar

The existence of mutually unbiased bases in a continuous Hilbert space remains open for debate, as further research in their existence is required before any conclusions can be reached.

Position states

|q\rangle

and momentum states

|p\rangle

are eigenvectors of Hermitian operators

\hat{x}

and

-i

	\partial
	\partialx

, respectively. Weigert and Wilkinson^[31] were first to notice that also a linear combination of these operators have eigenbases, which have some features typical for the mutually unbiased bases. An operator

\alpha\hat{x}-i\beta

	\partial
	\partialx

has eigenfunctions proportional to

\exp(i(ax^2+bx))

with

\alpha+2\betaa=0

and the corresponding eigenvalues

b\beta

. If we parametrize

\alpha

and

\beta

\cos\theta

and

\sin\theta

, the overlap between any eigenstate of the linear combination and any eigenstate of the position operator (both states normalized to the Dirac delta) is constant, but dependent on

\beta

|\langlex_\theta|x\rangle

2=	1
	2\pi\|\sin\theta\|

where

|x\rangle

and

|x_{\theta\rangle}

stand for eigenfunctions of

\hat{x}

and

\cos\theta\hat{x}-i\sin\theta

	\partial
	\partialx

Notes and References

Schwinger . J. . 1960 . Unitary Operator Bases, Harvard University . Proc. Natl. Acad. Sci. U.S.A.. 46. 4. 570–9. 222876 . 16590645 . 10.1073/pnas.46.4.570. 1960PNAS...46..570S . free .
Ivanovic . I. D. . 1981 . Geometrical description of quantal state determination . J. Phys. A . 14 . 12. 3241–3245 . 1981JPhA...14.3241I . 10.1088/0305-4470/14/12/019 .
M. Planat et al, A Survey of Finite Algebraic Geometrical Structures Underlying Mutually Unbiased Quantum Measurements, http://hal.ccsd.cnrs.fr/docs/00/07/99/18/PDF/MUB_FP.pdf.
Wootters . W. K. . Fields . B. D. . 1989 . Optimal State-Determination by Mutually Unbiased Measurements . Ann. Phys. . 191 . 2. 363–381 . 10.1016/0003-4916(89)90322-9. 1989AnPhy.191..363W . 10338.dmlcz/141471 . free .
Gottesman . D. . 1996 . Class of quantum error-correcting codes saturating the quantum Hamming bound . Phys. Rev. A . 54 . 3. 1862–1868 . 10.1103/physreva.54.1862. 9913672 . quant-ph/9604038 . 1996PhRvA..54.1862G . 16407184 .
Calderbank . A. R. . etal . 1997 . Quantum Error Correction and Orthogonal Geometry . Phys. Rev. Lett. . 78 . 3. 405–408 . 10.1103/physrevlett.78.405. quant-ph/9605005 . 1997PhRvL..78..405C . 15326700 .
Huang . Yichen . Entanglement criteria via concave-function uncertainty relations . Physical Review A . 29 July 2010 . 82 . 1 . 012335 . 10.1103/PhysRevA.82.012335. 2010PhRvA..82a2335H .
Spengler . C. . Huber . M. . Brierley . S. . Adaktylos . T. . Hiesmayr . B. C.. Beatrix Hiesmayr . 2012 . Entanglement detection via mutually unbiased bases . Phys. Rev. A . 86 . 2. 022311 . 10.1103/physreva.86.022311. 1202.5058 . 2012PhRvA..86b2311S . 34502667 .
Vaidman . L. . etal . 1987 . How to ascertain the values of
\sigma_x,\sigma_y,

and
\sigma_z

of a spin-1/2 particle . Phys. Rev. Lett. . 58 . 14. 1385–1387 . 1987PhRvL..58.1385V . 10.1103/PhysRevLett.58.1385 . 10034422 .
Englert . B.-G. . Aharonov . Y. . 2001 . The mean king's problem: prime degrees of freedom . Phys. Lett. A . 284 . 1. 1–5 . 10.1016/s0375-9601(01)00271-7. quant-ph/0101134 . 2001PhLA..284....1E . 14848100 .
Book: quant-ph/0610216 . 10.1063/1.2713445 . Three Ways to Look at Mutually Unbiased Bases . AIP Conference Proceedings . 2007 . Bengtsson . Ingemar . 12395501 . 889 . 40–51 .
quant-ph/0309120 . Klappenecker . Andreas . Roetteler . Martin . Constructions of Mutually Unbiased Bases . 2003 . 2003quant.ph..9120K .
Durt . T. . Englert . B.-G. . Bengtsson . I. . Życzkowski . K. . 2010 . On mutually unbiased bases . 1004.3348 . International Journal of Quantum Information . 8 . 4. 535–640 . 10.1142/s0219749910006502. 118551747 .
P. Butterley, W. Hall "Numerical evidence for the maximum number of mutually unbiased bases in dimension six, 2007, https://arxiv.org/abs/quant-ph/0701122.
Brierley . S. . Weigert . S. . 2008 . Maximal sets of mutually unbiased quantum states in dimension six . Phys. Rev. A . 78 . 4. 042312 . 10.1103/physreva.78.042312. 0808.1614 . 2008PhRvA..78d2312B . 9295036 .
quant-ph/0406032. Wootters. William K.. Quantum measurements and finite geometry. 2004.
Stacey. Blake C.. SIC-POVMs and Compatibility among Quantum States. Mathematics. 4. 2. 36. 10.3390/math4020036. 1404.3774. 2016. free.
Ingemar . Bengtsson . Kate . Blanchfield . Adán . Cabello . A Kochen–Specker inequality from a SIC . . 376 . 4 . 2012 . 374 - 376 . 10.1016/j.physleta.2011.12.011 . 1109.6514. 2012PhLA..376..374B . 55755390 .
quant-ph/0406175. Grassl. Markus. On SIC-POVMs and MUBs in Dimension 6. 2004.
Book: Geometry of quantum states : an introduction to quantum entanglement. Bengtsson. Ingemar. Życzkowski. Karol. Cambridge University Press. 2017. 9781107026254. Second. Cambridge, United Kingdom. 313 - 354. 967938939. Karol Życzkowski.
quant-ph/0103162 . Bandyopadhyay . Somshubhro . Oscar Boykin . P. . Roychowdhury . Vwani . Vatan . Farrokh . A new proof for the existence of mutually unbiased bases . 2002 . Algorithmica. 32 . 4 . 512–528 . 10.1007/s00453-002-0980-7 . 2001quant.ph..3162B . 1280557 .
Hirschman . I. I. Jr. . 1957. A note on entropy . American Journal of Mathematics . 1957 . 1. 152–156 . 10.2307/2372390. 2372390 .
H. Maassen, J.B.M. Uffink: Generalized entropic uncertainty relations: Phys. Rev. Lett. 60, 1103–1106 (1988).
quant-ph/0612014 . Damgaard . Ivan B. . Fehr . Serge . Renner . Renato . Salvail . Louis . Schaffner . Christian . A Tight High-Order Entropic Quantum Uncertainty Relation with Applications . 2006 . 2006quant.ph.12014D .
Deutsch . D. . 1982 . Uncertainty in Quantum Measurements . Physical Review Letters . 50 . 9. 631–633 . 10.1103/physrevlett.50.631. 1983PhRvL..50..631D .
0909.3720 . Ambainis . Andris . Limits on entropic uncertainty relations for 3 and more MUBs . 2009 .
S. Wehner and A. Winter, 2010 New J. Phys. 12 025009: http://iopscience.iop.org/1367-2630/12/2/025009/.
Wu . S. . Yu . S. . Mølmer . K. . 2009 . Entropic uncertainty relation for mutually unbiased bases . 0811.2298 . Phys. Rev. A . 79 . 2. 022104 . 10.1103/physreva.79.022104. 2009PhRvA..79b2104W . 119222014 .
Ballester. M.. S. Wehner. 41654752. Entropic uncertainty relations and locking: tight bounds for mutually unbiased bases . Physical Review A. 75. 1. 022319. 2007. quant-ph/0606244. 2007PhRvA..75a2319C . 10.1103/PhysRevA.75.012319 .
Wehner. S.. A. Winter. 118268095. Higher entropic uncertainty relations for anti-commuting observables. Journal of Mathematical Physics. 49. 6. 062105. 2008. 10.1063/1.2943685. 0710.1185 . 2008JMP....49f2105W .
0802.0394 . 10.1103/PhysRevA.78.020303 . Mutually unbiased bases for continuous variables . 2008 . Weigert . Stefan . Wilkinson . Michael . 67784632 . Physical Review A . 78 . 2 . 020303 . 2008PhRvA..78b0303W .

Mutually unbiased bases explained

Overview

Definition and examples

Example for d = 2

Example for d = 4

Existence problem

Related problems

Search methods

Weyl group method

Unitary operators method using finite fields

Hadamard matrix method

Entropic uncertainty relations

Infinite dimension Hilbert spaces

See also

Notes and References