advanced32 min studyLesson 368 in Physics

Generalized measurements and POVMs

Predict outcomes using positive effects that sum to identity, including noisy two-outcome and nonorthogonal multi-outcome measurements.

Before you begin

On this page

The question for today

What you’ll understand

Verify positivity and completeness of a discrete POVM.
Compute generalized Born probabilities for a specified density operator.
Explain why measurement outcome effects do not by themselves specify postmeasurement states.

A detector can have more outcomes than the dimension of the measured system, and its outcomes need not correspond to orthogonal eigenstates of one sharp observable. Positive-operator-valued measurements describe this wider range of probability assignments. Their effects identify which state features influence each outcome. Separating those probability rules from a specific physical state update makes room for noisy readout, auxiliary systems, and different apparatus designs that share the same outcome statistics.

Effects are positive probability functionals

A discrete POVM is a list of positive operators whose sum is identity. Each effect has eigenvalues between zero and one because its complement is the sum of the other positive effects. The trace of a density operator times an effect is therefore a nonnegative probability, and summing over the complete outcome list gives one. Unlike projectors, effects need not square to themselves, be mutually orthogonal, or commute with one another.

The outcome labels are classical records assigned by the experiment, not automatically eigenvalues of the effect matrices. A projective measurement is a special POVM whose effects are orthogonal spectral projectors. A measurement with more effects than dimension is not contradictory: the effects need not identify mutually orthogonal states. Continuous outcomes require an operator-valued measure and the corresponding integration, rather than applying a finite sum formula to a point density without its measure.

E_a\ge0,\qquad \sum_aE_a=I,\qquad p(a|\rho)=\operatorname{tr}(\rho E_a),\qquad 0\le E_a\le I

ρ and all effects are dimensionless operators on the same Hilbert space. The formula describes outcome probabilities for a specified measurement on a normalized state. It does not prescribe a postmeasurement state or require every effect to be a rank-one projector.↔ Scroll wide equations horizontally.

An unsharp spin measurement interpolates between limits

For a qubit, effects proportional to identity plus or minus a chosen Pauli operator provide a simple noisy measurement. A sharpness parameter between zero and one controls the contrast between the two eigenstates. At zero, both effects equal half the identity and the result is a fair bit independent of the state. At one, the effects become the sharp projectors. Intermediate values reduce information about the corresponding spin component.

The same outcome probabilities can arise from different mechanisms, such as a sharp measurement followed by classical record errors, or a direct system–auxiliary coupling with a different disturbance. Their effects agree, but their output states may differ. Describing this family as weak in its statistics does not determine a unique minimally disturbing implementation. The measurement instrument, introduced next, carries the additional conditional-state information.

E_\pm=\frac12(I\pm\eta Z),\qquad 0\le\eta\le1,\qquad \rho=\frac12(I+\mathbf r\cdot\boldsymbol\sigma)\ \Longrightarrow\ p(\pm)=\frac{1\pm\eta r_z}{2}

Z is the dimensionless Pauli matrix diag(1,−1), η is dimensionless sharpness, and rz is the state's Bloch z coordinate. These effects describe an unsharp Z probability rule; η is not automatically a coupling duration or a unique disturbance strength.↔ Scroll wide equations horizontally.

General effects fit into ordinary quantum dynamics

A POVM can be realized by adjoining an auxiliary system, applying an appropriate interaction, and measuring a sharp observable on a larger space. The enlarged-space construction explains how nonorthogonal effects on the original system can coexist with ordinary projective readout elsewhere. It is an existence statement with a specified dilation, not a claim that every laboratory detector uses the same auxiliary dimension or interaction.

Some effect sets span enough independent operator directions to reconstruct a density matrix from many identically prepared copies. Merely having many outcomes does not guarantee such informational completeness: effects confined to one Bloch plane miss the perpendicular component. A single outcome also does not identify an unknown state. Tomography uses frequencies from repeated preparations together with a trusted measurement model and a statistical uncertainty analysis.

Work it through

A noisy Z readout

Take η=3/5 and $|\psi\rangle=(\sqrt3|0\rangle+|1\rangle)/2$ . Write the two effects explicitly, verify their validity, and compute both outcome probabilities.

The effects are $E_+=\operatorname{diag}(4/5,1/5)$ and $E_-=\operatorname{diag}(1/5,4/5)$ . Their eigenvalues are positive and their sum is identity, so they form a valid POVM even though neither is a projector.
The state populations are 3/4 and 1/4, giving $r_z=3/4-1/4=1/2$ . The relative phase would not affect these diagonal effects, although it could affect a measurement with off-diagonal sensitivity.
Directly, $p(+)= (4/5)(3/4)+(1/5)(1/4)=13/20=0.65$ . Completeness gives $p(-)=7/20=0.35$ , agreeing with $(1\pm\eta r_z)/2$ .

The result

The two valid nonprojective effects predict probabilities 0.65 and 0.35. These numbers describe the recorded outcomes but leave the conditional output states unspecified.

The living notebook

Explore the model

Interactive lab

Interactive lab · quantum

Read a state coordinate before computing a POVM probability

Reset and set both controls to θ=60° and φ=0°; predict displayed Bloch z=0.5 and therefore p(+)=0.65 for η=0.6. For a fresh comparison, Reset and set both controls to θ=120° and φ=0°; predict z=−0.5 and derived p(+)=0.35. The instrument displays its ordinary basis probabilities and Bloch coordinates; the POVM probabilities are calculated separately.

Bloch vector / P(0)P(1)Phase latitude

Probability |0⟩: 71.1%
Probability |1⟩: 28.9%
Bloch (x, y, z): 0.78, 0.45, 0.42

Mixing angle θ65°Relative phase φ30°

t = 0 s

A normalized two-level pure state, not an interacting field simulation. Playback adds one radian per second to the relative phase; basis probabilities stay fixed. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forAn effect converts state information into a probability through the generalized Born rule. This pure two-level lab has no sharpness control, POVM apparatus, or measurement sampling and conditioning.

Pause & reason

Check your understanding

Interactive checks need JavaScript. You can still read the explanation.

Read the checkpoint explanation

Positivity ensures nonnegative Born probabilities, and completeness ensures they sum to one. Orthogonal projectors describe a special case. Individual traces need not equal one, and the effect set is defined independently of the particular input state.

Your turn

Try it yourself

Let three equatorial qubit states have Bloch directions at phases 0°, 120°, and 240°, and define effects $E_j=(2/3)|n_j\rangle\langle n_j|$ . Verify completeness and find outcome probabilities for input |+⟩ along the first direction.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

Write each effect as $(I+\mathbf n_j\cdot\boldsymbol\sigma)/3$ . The three unit direction vectors sum to zero.

Compare with the solution

Each effect is positive with eigenvalues 2/3 and zero. Their sum is I because the three vector contributions cancel and the three identity contributions add. For the input Bloch vector n0, $p(j)=(1+\mathbf n_0\cdot\mathbf n_j)/3$ . The dot products are 1, −1/2, and −1/2, yielding probabilities 2/3, 1/6, and 1/6. Three outcomes on a qubit are valid because these effects are not orthogonal projectors. This equatorial set alone cannot determine an arbitrary state's z coordinate.

Keep the idea

Take it with you

POVM effects are positive and complete, with no requirement of mutual orthogonality.
The generalized Born rule determines outcome probabilities for a specified state.
A POVM alone leaves measurement disturbance and conditional output states undetermined.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.