Skip to content

Naimark Dilation

Naimark dilation says that a generalized measurement can be viewed as an ordinary projective measurement after enlarging the Hilbert space. In finite dimensions, every POVM on a system can be obtained by coupling the system to an ancilla and then projectively measuring the larger system.

The slogan is:

POVM on S=PVM on a larger space, compressed back to S.\text{POVM on }S \quad=\quad \text{PVM on a larger space, compressed back to }S.

This result explains why POVMs are not an ad hoc replacement for projective measurement. They are the effective probability rules seen by a subsystem when the apparatus, ancilla, or environment is not kept as part of the final description.

Let HS\mathcal H_S be the system Hilbert space and let {Fm}\{F_m\} be a POVM:

Fm≥0,∑mFm=IS.F_m\ge0, \qquad \sum_m F_m=I_S.

A Naimark dilation consists of a larger Hilbert space K\mathcal K, an isometry

V:HS→K,V†V=IS,V:\mathcal H_S\to\mathcal K, \qquad V^\dagger V=I_S,

and a projective measurement {Πm}\{\Pi_m\} on K\mathcal K such that

Fm=V†ΠmV.F_m = V^\dagger\Pi_m V.

The projectors obey

ΠmΠn=δmnΠm,∑mΠm=IK.\Pi_m\Pi_n=\delta_{mn}\Pi_m, \qquad \sum_m\Pi_m=I_{\mathcal K}.

For any system state ρ\rho, the POVM probability is reproduced by the projective measurement on the dilated state:

Tr⁡S(Fmρ)=Tr⁡K ⁣[Πm VρV†].\operatorname{Tr}_S(F_m\rho) = \operatorname{Tr}_{\mathcal K} \!\left[ \Pi_m\,V\rho V^\dagger \right].

The theorem is about probabilities. A particular dilation can also define a state update, but a POVM alone does not uniquely determine that update.

For a finite POVM, choose one positive square root for each effect:

Mm=Fm.M_m=\sqrt{F_m}.

Let the ancilla have an orthonormal outcome basis {∣m⟩A}\{\lvert m\rangle_A\}. Define

V∣ψ⟩=∑mMm∣ψ⟩⊗∣m⟩A.V\lvert\psi\rangle = \sum_m M_m\lvert\psi\rangle \otimes \lvert m\rangle_A.

This map is an isometry because

⟨ϕ∣V†V∣ψ⟩=∑m⟨ϕ∣Mm†Mm∣ψ⟩=⟨ϕ∣(∑mFm)∣ψ⟩=⟨ϕ∣ψ⟩.\begin{aligned} \langle\phi\rvert V^\dagger V\lvert\psi\rangle &= \sum_m \langle\phi\rvert M_m^\dagger M_m\lvert\psi\rangle\\ &= \langle\phi\rvert \left(\sum_m F_m\right) \lvert\psi\rangle\\ &= \langle\phi|\psi\rangle. \end{aligned}

Equivalently,

V†V=IS.V^\dagger V=I_S.

Now measure the ancilla projectively with

Πm=IS⊗∣m⟩⟨m∣A.\Pi_m = I_S\otimes \lvert m\rangle\langle m\rvert_A.

Then

V†ΠmV=Mm†Mm=Fm.V^\dagger\Pi_m V = M_m^\dagger M_m = F_m.

Thus the POVM has been realized as a projective measurement on S⊗AS\otimes A after the isometric embedding VV.

An isometry can be embedded into a unitary on a sufficiently large space. Operationally, prepare the ancilla in a fixed ready state ∣0⟩A\lvert0\rangle_A and choose a unitary UU such that

U(∣ψ⟩S⊗∣0⟩A)=∑mMm∣ψ⟩S⊗∣m⟩AU(\lvert\psi\rangle_S\otimes\lvert0\rangle_A) = \sum_m M_m\lvert\psi\rangle_S \otimes \lvert m\rangle_A

for all ∣ψ⟩S\lvert\psi\rangle_S.

Then a projective measurement of the ancilla in the {∣m⟩A}\{\lvert m\rangle_A\} basis gives

p(m)=Tr⁡ ⁣[(IS⊗∣m⟩⟨m∣)U(ρS⊗∣0⟩⟨0∣)U†].p(m) = \operatorname{Tr} \!\left[ (I_S\otimes\lvert m\rangle\langle m\rvert) U(\rho_S\otimes\lvert0\rangle\langle0\rvert) U^\dagger \right].

The result reduces to

p(m)=Tr⁡(FmρS).p(m)=\operatorname{Tr}(F_m\rho_S).

This is the standard indirect-measurement picture: couple system to apparatus, then perform a sharp pointer readout on the apparatus.

The construction above does more than reproduce probabilities. If the ancilla outcome is mm and the ancilla is discarded afterward, the system operation is

Im(ρ)=MmρMm†.\mathcal I_m(\rho) = M_m\rho M_m^\dagger.

For the square-root choice Mm=FmM_m=\sqrt{F_m}, this is the square-root instrument:

Im(ρ)=Fm ρ Fm.\mathcal I_m(\rho) = \sqrt{F_m}\,\rho\,\sqrt{F_m}.

But the same POVM can be realized by other measurement operators KmK_m satisfying

Km†Km=Fm.K_m^\dagger K_m=F_m.

For example, Km=UmFmK_m=U_m\sqrt{F_m} has the same effect for any unitary UmU_m:

Km†Km=Fm.K_m^\dagger K_m=F_m.

It generally has different backaction. Therefore:

Naimark dilation explains POVM probabilities;the instrument specifies the state update.\text{Naimark dilation explains POVM probabilities;} \qquad \text{the instrument specifies the state update.}

See Measurement Backaction for this distinction in physical terms.

Consider the two-outcome qubit POVM

F±=12(I±ησz),0≤η≤1.F_\pm = \frac{1}{2} \left( I\pm \eta\sigma_z \right), \qquad 0\le\eta\le1.

For η=1\eta=1, the effects are the projectors onto σz\sigma_z eigenstates. For 0<η<10\lt\eta\lt1, they are positive but not projectors:

F±2≠F±.F_\pm^2\ne F_\pm.

The square-root Naimark construction uses

M±=F±M_\pm=\sqrt{F_\pm}

and an ancilla basis {∣+⟩A,∣−⟩A}\{\lvert+\rangle_A,\lvert-\rangle_A\}:

V∣ψ⟩=M+∣ψ⟩∣+⟩A+M−∣ψ⟩∣−⟩A.V\lvert\psi\rangle = M_+\lvert\psi\rangle\lvert+\rangle_A + M_-\lvert\psi\rangle\lvert-\rangle_A.

Measuring the ancilla projectively gives probabilities

p(±)=Tr⁡(F±ρ).p(\pm) = \operatorname{Tr}(F_\pm\rho).

The unsharpness is not because the final pointer measurement is fuzzy. In the dilated picture, the pointer measurement can be sharp; the unsharp system POVM arises because the system is coupled to the pointer only partially.

A qubit projective measurement on the qubit alone can have at most two nonzero rank-one orthogonal outcomes. A qubit POVM can have three or more nonzero outcomes.

The trine POVM has three effects

Fk=13(I+nk⋅σ),k=1,2,3,F_k = \frac{1}{3} \left( I+\mathbf n_k\cdot\boldsymbol\sigma \right), \qquad k=1,2,3,

where the unit vectors nk\mathbf n_k lie in the equatorial plane separated by 120∘120^\circ and satisfy

∑k=13nk=0.\sum_{k=1}^3\mathbf n_k=0.

Then

∑k=13Fk=I.\sum_{k=1}^3F_k=I.

These three effects cannot be a three-outcome PVM on a two-dimensional Hilbert space, because three nonzero mutually orthogonal projectors cannot all fit inside a qubit space. Naimark dilation says they can be represented as a projective measurement after embedding the qubit into a larger Hilbert space.

A dilation is not unique. One can always add unused ancilla dimensions or split an outcome into several projectors and classically coarse grain them later.

A minimal dilation avoids redundant dimensions in a precise operator-theoretic sense. For most physics applications, the important point is not minimality but the operational structure:

system couples to ancilla→sharp pointer measurement→effective POVM on system.\text{system couples to ancilla} \quad\to\quad \text{sharp pointer measurement} \quad\to\quad \text{effective POVM on system}.

Different dilations can implement the same POVM with different post-measurement states if the system output is kept. That is why the instrument remains the complete measurement object.

Naimark dilation and Stinespring dilation are closely related but answer different questions.

Naimark dilation starts from a POVM and asks:

How can these outcome probabilities arise from a projective measurement?\text{How can these outcome probabilities arise from a projective measurement?}

Stinespring dilation starts from a completely positive map or channel and asks:

How can this state transformation arise from unitary evolution on a larger system?\text{How can this state transformation arise from unitary evolution on a larger system?}

For measurements, both ideas often appear together. A system couples unitarily to an apparatus, a pointer is projectively measured, and the resulting system operation is described by Kraus operators. But the canonical home for POVM-as-PVM is this page; the channel-as-unitary-plus-environment result belongs to the Stinespring page.

A POVM is not a vague or approximate measurement rule. It is the exact probability rule induced on a subsystem by a larger projective measurement model.

The same POVM can be realized by different dilations and different instruments. Naimark dilation explains the statistics, not a unique post-measurement state.

The unitary-apparatus version requires a specified initial ancilla state. Changing the ready state changes the effective POVM.

Confusing system projectors with pointer projectors

Section titled “Confusing system projectors with pointer projectors”

In the dilated picture, the final projective measurement acts on the enlarged space, often on the apparatus pointer. The induced operators on the original system need not be projectors.

Let V∣ψ⟩=∑mFm∣ψ⟩⊗∣m⟩V\lvert\psi\rangle=\sum_m\sqrt{F_m}\lvert\psi\rangle\otimes\lvert m\rangle. Show that V†V=IV^\dagger V=I.

Solution

For arbitrary ∣ϕ⟩\lvert\phi\rangle and ∣ψ⟩\lvert\psi\rangle,

⟨ϕ∣V†V∣ψ⟩=∑m⟨ϕ∣Fm †Fm∣ψ⟩.\langle\phi\rvert V^\dagger V\lvert\psi\rangle = \sum_m \langle\phi\rvert \sqrt{F_m}^{\,\dagger}\sqrt{F_m} \lvert\psi\rangle.

Since FmF_m is positive, Fm †Fm=Fm\sqrt{F_m}^{\,\dagger}\sqrt{F_m}=F_m. Therefore

⟨ϕ∣V†V∣ψ⟩=⟨ϕ∣(∑mFm)∣ψ⟩=⟨ϕ∣ψ⟩.\langle\phi\rvert V^\dagger V\lvert\psi\rangle = \langle\phi\rvert \left(\sum_mF_m\right) \lvert\psi\rangle = \langle\phi|\psi\rangle.

Thus V†V=IV^\dagger V=I.

Using Πm=I⊗∣m⟩⟨m∣\Pi_m=I\otimes\lvert m\rangle\langle m\rvert, show that V†ΠmV=FmV^\dagger\Pi_mV=F_m for the square-root construction.

Solution

The projector Πm\Pi_m selects the mmth ancilla component:

ΠmV∣ψ⟩=Fm∣ψ⟩⊗∣m⟩.\Pi_mV\lvert\psi\rangle = \sqrt{F_m}\lvert\psi\rangle\otimes\lvert m\rangle.

Therefore, for arbitrary ∣ϕ⟩\lvert\phi\rangle and ∣ψ⟩\lvert\psi\rangle,

⟨ϕ∣V†ΠmV∣ψ⟩=⟨ϕ∣Fm †Fm∣ψ⟩=⟨ϕ∣Fm∣ψ⟩.\langle\phi\rvert V^\dagger\Pi_mV\lvert\psi\rangle = \langle\phi\rvert \sqrt{F_m}^{\,\dagger}\sqrt{F_m} \lvert\psi\rangle = \langle\phi\rvert F_m\lvert\psi\rangle.

Hence V†ΠmV=FmV^\dagger\Pi_mV=F_m.

For the trine POVM, assume ∑knk=0\sum_k\mathbf n_k=0. Show that ∑kFk=I\sum_kF_k=I.

Solution

Using

Fk=13(I+nk⋅σ),F_k = \frac{1}{3} \left( I+\mathbf n_k\cdot\boldsymbol\sigma \right),

we find

∑k=13Fk=13(3I+(∑k=13nk)⋅σ).\sum_{k=1}^3F_k = \frac{1}{3} \left( 3I+ \left(\sum_{k=1}^3\mathbf n_k\right) \cdot\boldsymbol\sigma \right).

Since ∑knk=0\sum_k\mathbf n_k=0,

∑k=13Fk=I.\sum_{k=1}^3F_k=I.
  • M. A. Naimark, Normed Rings, Noordhoff (1964).
  • A. S. Holevo, Probabilistic and Statistical Aspects of Quantum Theory, North-Holland (1982).
  • P. Busch, P. J. Lahti, and P. Mittelstaedt, The Quantum Theory of Measurement, Springer (1996).
  • K. Kraus, States, Effects, and Operations: Fundamental Notions of Quantum Theory, Springer (1983).
  • M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information, Cambridge University Press (2010).
  • J. Watrous, The Theory of Quantum Information, Cambridge University Press (2018).