Operational Distinguishability and Continuity Bounds

Operational distinguishability is always relative to a set of measurements or inputs. Trace and diamond norms give complete answers for finite systems, but continuum fields require bounded-observable or energy-constrained versions; otherwise small perturbations can be perfectly distinguished by arbitrarily energetic probes.

Required background. Use probability and conditional expectation and Hypothesis Testing and Asymptotic Distinguishability. Helpful background. Fidelity, Chernoff Bounds, and State Overlap supplies complementary bounds.

State bias on a declared algebra

For normal states $\omega,\varphi$ on $\mathfrak M$, the largest bias attainable with bounded effects is controlled by the predual norm:

$$\lVert\omega-\varphi\rVert =\sup_{\lVert A\rVert\leq1} \lvert\omega(A)-\varphi(A)\rvert.$$

With equal priors, the optimal binary success probability is

$$p_{\rm succ}=\frac12+\frac14\lVert\omega-\varphi\rVert.$$

The supremum ranges only over the declared algebra. A detector subalgebra can therefore give a smaller distance than the sharp local algebra. Fidelity bounds the same bias once its square convention is fixed.

The structural diagram emphasizes that operational bias is a resource-selected branch of state comparison.

Bounded-observable bias, hypothesis testing, and channel discrimination answer related but distinct tasks. Channel statements additionally require an input set, ancillary system, and resource constraint. Schematic.

Channels in infinite dimension

For channels $\Phi$ and $\Psi$, the unconstrained diamond norm optimizes over all inputs and ancillas. On a bosonic field it can equal its maximal value even when the channels are close on every experimentally available state. Given an input Hamiltonian $H$ and energy budget $E$, define instead

$$\lVert\Phi-\Psi\rVert_{\diamond E} =\sup_{\rho_{AR}:\,\operatorname{Tr}(H\rho_A)\leq E} \left\lVert[(\Phi-\Psi)\otimes{\rm id}_R](\rho_{AR})\right\rVert_1.$$

The Hamiltonian, energy origin, ancilla class, and input algebra are part of the definition. For an attenuation channel acting on an energy-bounded wavepacket, this norm controls the optimal discrimination bias for ideal versus approximate evolution.

Continuity of entropy also needs restrictions. Finite-dimensional Fannes-type bounds cannot be transplanted to an unbounded field alphabet. Energy-constrained continuity bounds add spectral conditions on $H$ and an explicit tail estimate; Winter 2016, pp. 291–313 gives sharp finite-dimensional bounds that form the regulated starting point.

What a bound does not cover

The figure below makes the resource dependence visible.

An unconstrained diamond norm and an energy-constrained one solve different discrimination problems. Hidden changes to the Hamiltonian, ancilla, cutoff, or postselection rule invalidate the quoted operational bound. Schematic.

Always report the admissible observables, input energy set, ancilla, cutoff error, and output norm. A small constrained distance licenses continuity only for states inside that same set.

References

• Winter, Andreas. “Tight Uniform Continuity Bounds for Quantum Entropies: Conditional Entropy, Relative Entropy Distance and Energy Constraints.” Communications in Mathematical Physics 347 (2016): 291–313. DOI.

Further reading

• Fuchs, Christopher A., and Jeroen van de Graaf. “Cryptographic Distinguishability Measures for Quantum-Mechanical States.” IEEE Transactions on Information Theory 45 (1999): 1216–1227. DOI.
• Shirokov, M. E. “Energy-Constrained Diamond Norms and Their Use in Quantum Information Theory.” Problems of Information Transmission 54 (2018): 20–33. DOI. Open preprint.