Commutation-derived Quantum Filters

• The paper demonstrates that maintaining canonical commutation relations is crucial for constructing filters that achieve optimal state estimation and error mitigation in quantum systems.
• It details measurement-based filtering equations derived from commutant algebras and Tomita–Takesaki theory to ensure accurate and representation-independent quantum filtering.
• It shows that applying symplectic constraints and ancilla-assisted circuits allows for efficient error correction and quadratic suppression of infidelity in noisy quantum channels.

Commutation-derived quantum filters are a class of quantum information processing protocols that leverage commutation relations—either at the level of canonical observables, system–probe dynamics, or quantum channel structure—to define, constrain, or implement optimal state estimation, error filtration, and information extraction in open quantum systems. These filters arise in quantum control, measurement theory, and quantum error mitigation, where preservation of operator commutativity is not only a physical constraint but a tool for constructing robust, efficient, and sometimes representation-independent filter architectures.

1. Foundations: Commutation Relations and Filtering Architectures

Central to the construction of quantum filters is the requirement that certain operator commutation relations (CCRs) be preserved throughout the system's evolution. For linear quantum systems, these take the canonical form $[x_j, x_k] = 2 i \Theta_{jk}$, with $\Theta$ a nondegenerate antisymmetric matrix, reflecting the underlying symplectic structure of quantum mechanics. The plant’s evolution is constrained by these CCRs, governing both physical realizability (ensuring dynamics can be derived from a quadratic Hamiltonian and linear coupling) and admissibility of subsequent quantum filtering schemes (Vladimirov et al., 2013).

A commutation-derived quantum filter typically couples a second quantum system (the filter) unidirectionally to the plant’s output. The filter must itself maintain prescribed CCRs, necessitating algebraic constraints on its system matrices. Physical realizability of both plant and filter yields coupled quadratic equations—termed PR constraints—directly associated with the preservation of commutators under quantum stochastic evolution.

2. Measurement and Commutant Algebras: Quantum Filtering Equations

Measurement-based quantum filtering relies on the principle that only system observables commuting with the measurement history (i.e., elements of the commutant of the generated von Neumann algebra) admit conditional expectation and thus statistical estimation (Gough, 6 Oct 2025). This conceptual structure is formalized via the Tomita–Takesaki theory, which ensures a rich and unique commutant structure, especially in the presence of general (e.g., thermal or squeezed) input fields (Gough et al., 18 Jan 2026).

The filtering equations—the quantum analogs of the classical Kushner–Stratonovich and Zakai equations—are derived by projecting the system’s Heisenberg evolution onto the commutant associated with the measurement algebra. For continuous homodyne measurement in a thermal state, the structure of the commutant is determined by the Araki–Woods representation, yielding explicit, representation-independent filters parameterized only by physical noise covariances (Gough, 6 Oct 2025, Gough et al., 18 Jan 2026). For discrete-time or POVM-based measurements, existence and uniqueness of the filter follow from the strong commutativity of the instruments involved, which allows for conditional POVMs and recursive update formulas (Somaraju et al., 2013).

3. Commutation-Derived Quantum Filters in Error Mitigation and Correction

An explicit application of commutation-derived quantum filters is in quantum error mitigation and correction, particularly for quantum channels with structure (e.g., Clifford or Pauli noise). In this context, a quantum filter is realized as a superchannel acting on a noisy channel $\mathcal{E}$, constructed by coherently querying $\mathcal{E}$ in parallel with controlled unitaries that (anti)commute with the target operation (Das et al., 2024).

Typical constructions involve decomposing errors into components commuting or anti-commuting with a chosen involution (e.g., Pauli operators). Ancilla-assisted circuits (using Hadamards, controlled-Paulis, and measurement) perform this filtration deterministically or via post-selection. For $n$-qubit Clifford circuits, full correction is achieved using $2n$ ancilla qubits, mapping any Pauli error to a classically detectable ancilla outcome, after which the appropriate Pauli correction is applied. For non-Clifford gates, only partial filtration is possible, converting generic errors into biased forms compatible with specialized quantum error-correcting codes.

For more resource-constrained scenarios, ancilla-efficient filters employing only two ancillary qubits can eliminate all single-qubit (weight-1) Pauli errors by leveraging their anti-commutation properties with aggregate multi-qubit operators such as $Z^{\otimes n}$ and $X^{\otimes n}$ (Das et al., 2024). This yields quadratic suppression of average infidelity under local depolarizing noise.

4. Mathematical Structure: Symplectic Constraints, Stationarity, and Covariance Control

Commutation-derived quantum filtering in control and estimation contexts is mathematically formulated as a constrained optimization problem. For coherent (measurement-free) filtering of physically realizable linear quantum plants, one seeks to minimize a quadratic mean-squared discrepancy between the plant’s and filter’s output variables. The constraints, arising from CCR preservation, enforce skew-Hamiltonian symplectic structure on the involved state-space matrices.

First-order optimality (stationarity) conditions are obtained by introducing Lagrange multipliers for the PR constraints and differentiating a Lagrangian involving the infinite-horizon cost, PR constraints, and system Gramians. This results in a set of coupled algebraic matrix equations involving the Gramians $P, Q$, the Hankelian $H = QP$, matrices $\Theta$0, and the filter’s internal parameters. The symplectic forms not only ensure viability but permeate the optimality equations, dictating the geometry of the feasible solution set and enabling explicit parametrization or elimination of non