Operational Concealment of Measurement Incompatibility by Quantum Channels

Technical Summary: Operational Concealment of Measurement Incompatibility by Quantum Channels

Problem Statement
Measurement incompatibility—the impossibility of jointly measuring certain quantum observables—is a fundamental nonclassical resource. While recent work has generalized incompatibility to channels, this paper addresses the complementary question: how does a quantum channel modify the operational accessibility of measurement incompatibility? The authors identify a phenomenon termed operational concealment, where measurement incompatibility remains intact at the operator level but becomes operationally inaccessible when observations are restricted to the output of a quantum channel. This scenario is motivated by restricted-access settings, such as semi-device-independent certification, where a verifier interacts with a device only through a specific channel output.

Methodology
The paper develops a systematic framework based on the adjoint characterization of quantum channels.

1. Adjoint Kernel Framework: The authors utilize the adjoint channel $E^\dagger$ (Heisenberg picture) rather than the Schrödinger picture, as measurement statistics depend linearly on observables via $E^\dagger$. They define the adjoint kernel, $\ker(E^\dagger)$, as the set of Hermitian operators that vanish under the channel's action.
2. Operational Equivalence: Two observables are defined as operationally equivalent if their difference lies in $\ker(E^\dagger)$. This partitions the space of Hermitian operators into operational equivalence classes.
3. Operational Observable Space: The authors formalize the space of accessible observables as the quotient space $\text{Herm}(\mathcal{H}_{\text{out}}) / \ker(E^\dagger)$. Within this space, measurement incompatibility is determined by whether the equivalence classes of two measurements admit compatible representatives.
4. Quantitative Measures: The framework introduces concealment robustness ($R_c$), defined as the minimal noise required to mix into a measurement pair to make it concealed. This is compared against standard incompatibility robustness ($R_{\text{inc}}$).
5. Approximate Concealment: An approximate framework is developed using $\epsilon$-concealment, quantified by the deviation in statistics, admitting a Semidefinite Programming (SDP) formulation.

Key Contributions and Results

• Structural Characterization: The paper proves that operational concealment is completely characterized by the adjoint kernel. A pair of POVMs is concealed if and only if there exist compatible POVMs within their respective operational equivalence classes (Theorem 1).
• Channel Classification:
  • Kernel Invariance: Channels with identical adjoint kernels are indistinguishable regarding the concealment of POVM pairs (Theorem 3).
  • Kernel Monotonicity: A preorder is established where channels with larger kernels conceal more pairs (Theorem 4).
  • Injectivity: If a channel has an injective adjoint ($\ker(E^\dagger) = \{0\}$), it cannot conceal any incompatible pair; concealment requires non-trivial hidden observable directions (Corollary 1).
• Distinction from Effective Compatibility: The authors distinguish operational concealment (existence of compatible representatives in equivalence classes) from effective compatibility (compatibility of the effective measurements $E^\dagger(M)$ and $E^\dagger(N)$). Concealment implies effective compatibility, but the converse is false. Effective compatibility can occur for injective channels where concealment is impossible.
• Qubit Channel Analysis:
  • Unital Channels: For rank-2 unital qubit channels, a geometric projection criterion is derived. Concealment depends on the projection of Bloch vectors onto the subspace orthogonal to the kernel of $T T_E$ (Theorem 7).
  • Examples: The complete dephasing channel is shown to conceal specific incompatible pairs (e.g., Pauli $X$ and $Z$) because the $X$-measurement becomes operationally equivalent to a trivial measurement. Conversely, the amplitude damping channel (for $\gamma < 1$) has an injective adjoint and conceals nothing.
• Quantitative Hierarchy: The paper establishes that concealment robustness is always less than or equal to standard incompatibility robustness ($R_c \leq R_{\text{inc}}$). For injective channels, they are equal; for non-injective channels, $R_c$ can be strictly smaller, demonstrating that operational restrictions reduce the noise required to hide incompatibility (Proposition 7, Example 2).
• Steering Consequences: The authors show that if a measurement pair is concealed, the resulting assemblage on a remote party admits a Local Hidden State (LHS) model, implying the loss of steerability under restricted access (Corollary 4).

Significance
The paper claims to provide a systematic characterization and quantitative treatment of operationally inaccessible measurement incompatibility. By formalizing the "operational observable space" via the adjoint kernel, the work unifies the adjoint characterization with geometric results and robustness measures. The significance lies in:

1. Clarifying that incompatibility is not an absolute property of a channel but depends on the specific measurement pair and the channel's kernel structure.
2. Providing explicit analytical and geometric criteria for when incompatibility is hidden, particularly for qubit channels.
3. Offering a quantitative tool (concealment robustness) that differs from standard robustness in non-injective scenarios, which has implications for semi-device-independent certification where an adversary or noise might restrict access to the full quantum state space.
4. Linking the phenomenon to restricted-state compatibility, extending previous results by Heinosaari et al. to the specific context of channel outputs.

The authors note that while they provide a complete characterization for unital qubit channels, extending these results to higher dimensions and non-unital channels remains an open direction. They also highlight that the framework assumes tomographically complete input states and single-copy statistics, leaving questions about adaptive or collective strategies open.