Quasiprobability Decomposition (QPD)

• QPD is a framework in quantum theory that extends classical probabilities by using negative and complex weights to capture nonclassical phenomena.
• Its formulation, exemplified by the Kirkwood–Dirac distribution, enables operational applications such as state tomography, error mitigation, and dynamic process analysis.
• QPD negativity acts as a resource witness, linking measurement uncertainty with quantum contextuality and optimizing sampling strategies for efficient simulations.

A quasiprobability decomposition (QPD) is a formalism in quantum theory that generalizes the notion of state or process representation via classical probability distributions to accommodate negative or non-real weights, enabling a complete and operational characterization of nonclassical phenomena that cannot be captured probabilistically. This framework underlies a wide array of concepts, protocols, and resource theories in quantum information, including state and channel tomography, resource identification, circuit error mitigation, and the measurement of quantum thermodynamic quantities. QPDs manifest in diverse technical guises such as the Kirkwood–Dirac (KD) distribution, extended quasi-stochastic representations, and algorithmically-optimized variational decompositions, and are foundational for both theoretical and experimental investigations of nonclassicality, contextuality, and quantum computational advantage.

1. Mathematical Formulation of QPD

Let $\mathcal{H}$ be a $d$-dimensional Hilbert space and $\rho$ a density operator. The Kirkwood–Dirac (KD) QPD is defined with respect to two orthonormal bases $\{\lvert a_i\rangle\}$ and $\{\lvert b_j\rangle\}$ as

$$Q_{ij} = \langle b_j \mid a_i \rangle \langle a_i \mid \rho \mid b_j\rangle$$

for all $i, j$. This construction satisfies $\sum_{i,j} Q_{ij} = 1$ but admits negative or complex entries; QPDs in general encode operator decompositions beyond positive-semidefinite mixtures.

A central result is the reconstruction formula, which—if $\langle b_j \mid a_i\rangle \neq 0 \ \forall i,j$—allows inverting the QPD: $$\rho = \sum_{i,j} Q_{ij} \frac{|b_j\rangle\langle a_i|}{\langle b_j \mid a_i\rangle}$$ This formal decomposition generalizes to arbitrary sets of operators (including for quantum channels) via QPDs over minimal informationally complete POVMs and their duals, establishing categorical functorial embeddings of quantum theory into quasi-stochastic processes (Wetering, 2017).

2. QPD as a Universal Resource Identifier

In the general theory of quantum resources, one distinguishes between a convex closed set $d$0 of “free” (classical) states and resourceful (nonclassical) states $d$1. It is proven that for every $d$2, there always exist incompatible bases such that the corresponding QPD exhibits negativity: at least one $d$3 (Tan et al., 2024). Conversely, for every $d$4, all $d$5 in those bases.

The total negativity $d$6 can be made equal to the minimal Frobenius-norm distance from $d$7 to $d$8, providing a geometric resource measure. Thus, QPD negativity is a necessary and sufficient witness of resourcefulness for arbitrary convex resource theories.

3. Operational Implementations and Experimental Reconstruction

QPDs play a critical role in quantum information processing and experimental physics. Many protocols—such as the interferometric construction of the KD QPD of quantum work—systematically reconstruct QPDs by combining controlled-unitary dynamics with ancillary probes (Hernández-Gómez et al., 2024). The expectation values accessed via such protocols—typically through Ramsey interference fringes or sequential/weak measurements—directly yield the real and imaginary components of QPDs. The outcomes serve both as quantitative witnesses of nonclassical correlations and as full, informationally complete state (or process) tomography frameworks.

In error mitigation, QPDs enable virtual simulation of ideal circuits by sampling from implementable noisy primitives. For a given circuit element or channel $d$9, one writes

$\rho$0

where each $\rho$1 is experimentally accessible, and the overhead $\rho$2 quantifies the cost of negative-weight sampling (Piveteau et al., 2021). Key improvements, such as robust semidefinite-programming approaches and noise-aware iterative basis construction, have substantially reduced practical overheads for quantum error mitigation.

4. QPDs in Quantum Channels, Circuits, and Sampling Complexity

QPDs are not limited to state decompositions but also provide complete representations of quantum operations and processes. Any completely positive trace-preserving (CPTP) map can be embedded as a quasi-stochastic matrix acting on a vector of generalized “probabilities” (via IC-POVMs), with composition realized by standard matrix multiplication (Wetering, 2017). This embedding clarifies the connection to classical stochastic processes and shows that quantum theory forms a strict subcategory of quasi-stochastic process theory.

In large-scale quantum circuit simulation (e.g., circuit knitting or distributed computation), QPD-based wire cutting decomposes mid-circuit identity channels into weighted sums of measure-and-prepare channels, incurring exponential overhead in the number of cuts. Advances such as recursive tomography-based decomposition have achieved exponential-to-polynomial reductions in sampling cost on tree-structured circuits, supplanting the global QPD overhead with locally controlled bias at each interface (Harada et al., 22 Dec 2025). For simulating quantum state transfer over noisy interconnects, QPDs combine depolarizing twirls of the available channel with minimal measure-and-prepare sets, parametrized by an experimentally calibrated fidelity (Bechtold et al., 2 Jul 2025).

5. QPDs, Negativity, Contextuality, and Weak Values

QPD negativity is intimately connected with quantum contextuality. In measurement scenarios, the KD QPD of a state with respect to a POVM $\rho$3 and a PVM $\rho$4 is $\rho$5. Both negativity (i.e., $\rho$6 for negative $\rho$7) and nonreality (nonzero imaginary part) provide necessary and sufficient criteria for quantum contextuality certified via weak measurement with postselection; the real part of the weak value for particular projectors coincides with negative KD QPD outcomes (Budiyono, 2024, Tan et al., 2024). The quantum uncertainty so quantified is directly linked to the disturbance exerted by nonselective measurements.

This framework extends naturally to statistical fluctuation quantities, e.g., the QPD of work, where noncommuting projectors in the initial and final bases lead to negative and complex weightings, with experimental observables precisely tracking QPD moments. In Keldysh-type QPDs, the negativity is a signature of interference (off-diagonal trajectory contributions) and underlies nonclassicality in dynamical processes (Hofer, 2017).

6. Structured Sampling and Monte Carlo Implementation

Sampling inefficiency arising from QPD negativity is a fundamental bottleneck in quantum algorithms that exploit QPD-based error mitigation or simulation. Advanced variance reduction strategies, such as stratified sampling based on permutation-invariant configuration strata (e.g., counts vectors), provably never increase estimator variance relative to naïve Monte Carlo and typically yield large constant-factor reductions (up to 60–80% in idealized models) (Dai et al., 11 Feb 2026). These approaches are compatible with product-form QPDs as found in many practical protocols (e.g., probabilistic error cancellation, probabilistic angle interpolation), and their computational cost remains tractable for moderate system sizes.

7. Generalizations and Conceptual Unification

QPD theory encompasses and unifies a range of quasiprobability representations: phase-space distributions (Wigner, Husimi Q), resource-theoretic decompositions, and multipartite and operator-valued extensions (Sperling et al., 2018, Sperling et al., 2018). For dynamic systems, KD-type and Keldysh QPDs serve as universal tools for joint fluctuation analysis of non-commuting observables, enabling calculation of moments, cumulants, and direct experimental tests of dynamical nonclassicality.

QPDs thus serve as a universal operational structure for recognizing, quantifying, and harnessing nonclassical resources in quantum information science. They enable both qualitative distinctions—classicality vs. quantum—based on negativity, and quantitative metrics for resource strength, cost, and experimental accessibility. Their foundational role extends across quantum tomography, resource theory, contextuality testing, error mitigation, distributed computation, and quantum thermodynamics.

---

References:

• "Kirkwood-Dirac Type Quasiprobabilities as Universal Identifiers of Nonclassical Quantum Resources" (Tan et al., 2024)
• "Quasiprobability decompositions with reduced sampling overhead" (Piveteau et al., 2021)
• "Stratified Sampling for Quasi-Probability Decompositions" (Dai et al., 11 Feb 2026)
• "Separation of measurement uncertainty into quantum and classical parts based on Kirkwood-Dirac quasiprobability and generalized entropy" (Budiyono, 2024)
• "Quantum Theory is a Quasi-stochastic Process Theory" (Wetering, 2017)
• "Exponential-to-polynomial scaling of measurement overhead in circuit knitting via quantum tomography" (Harada et al., 22 Dec 2025)
• "Simulating Quantum State Transfer between Distributed Devices using Noisy Interconnects" (Bechtold et al., 2 Jul 2025)
• "Interferometry of quantum correlation functions to access quasiprobability distribution of work" (Hernández-Gómez et al., 2024)
• "Quasiprobability representation of quantum coherence" (Sperling et al., 2018)
• "Quasi-probability distributions for observables in dynamic systems" (Hofer, 2017)
• "Quasistates and quasiprobabilities" (Sperling et al., 2018)
• "The quasiprobability behind the out-of-time-ordered correlator" (Halpern et al., 2017)
• "Overhead for simulating a non-local channel with local channels by quasiprobability sampling" (Mitarai et al., 2020)