Matthew Duschenes of Los Alamos National Laboratory, University of Waterloo, and Perimeter Institute for Theoretical Physics, alongside colleagues, has established a new framework to compute moment operators for ensembles of quantum channels for all moment orders. The work gives hierarchies between these ensembles operational meaning with the definition of channel t-designs. Different types of noise impact these operators in opposing ways; depolarizing noise decreases the norm, while amplitude damping noise increases it. This research generalizes noise-induced concentration phenomena to channel-design-induced phenomena and includes the discovery of a block-orthogonal basis for permutations that simplifies moment calculations.

Channel Ensembles and Quantum Dynamics with Randomness

Hierarchies between ensembles of quantum channels are now theoretically defined, allowing for comparison and ranking based on inequalities of their moment operator norms; this gives them operational meaning. Additive error under the -norm provides a clear interpretation of how easily distinguishable states are at the output of channels drawn from different ensembles, with an error of ε under the -norm meaning the output state of any channel from two ensembles, C or C′, is at most ε-away under the 1-norm for algorithms making a single channel query.

This framework extends beyond unitary ensembles, offering a new method for evaluating the performance of quantum data transfer systems. The Depolarize ensemble emerges as a natural reference point for quantum channel ensembles, generalizing the Haar ensemble and possessing unique invariance properties.

Its moment operator is present within the moment operators of all quantum channel ensembles, positioning it at the base of the moment operator norm hierarchy; this suggests a fundamental role for the Depolarize ensemble in characterizing other channel ensembles. This observation highlights the evolving understanding of channel properties and their connection to algorithmic design.

Recent work demonstrates that design-like randomness can arise dynamically in chaotic many-body systems, even without explicit sampling from a Haar ensemble, and this perspective extends to subsystem quantum channels induced by chaotic or noisy dynamics. The current research builds on this by providing a natural channel-level extension, allowing scientists to study these subsystem channels and their properties.

Moment Operators Capture Statistical Properties of Quantum Experiments

The ability to quantify relationships between different quantum channels now benefits from a newly established framework for calculating moment operators for any order, denoted as ‘t’. This hierarchy gives them operational meaning, and defines concepts like channel t-designs, which assess how closely a given ensemble mimics a reference point. This means any quantum channel ensemble will exhibit a moment operator norm equal to or greater than that of the Depolarize ensemble, offering a clear benchmark for comparison.

Researchers used a block-orthogonal basis for permutations to simplify calculations and potentially offer broader utility in moment calculations beyond this specific study. The framework extends the understanding of how randomness impacts quantum systems, moving beyond scenarios where noise is the primary driver of statistical behavior.

The work demonstrates that different types of noise can decrease the norm of the moment operators (e.g., depolarizing noise), as well as increase it (e.g., amplitude damping noise), and generalizes noise-induced concentration phenomena to channel-design-induced phenomena. Inner products between superoperators, the team found, directly correspond to inner products between Choi states, meaning the norm and trace of moment operators reflect the average overlap and entanglement fidelity of these states within the ensemble.

The paper notes, hinting at the potential for further refinement of these concepts. The resulting insights provide a more nuanced understanding of how different types of noise, including depolarizing and amplitude damping, affect the norms of moment operators, and therefore the reliability of quantum data transfer.

Defining Channel t-Designs via Moment Operator Comparison

This framework assigns an operational meaning to these ensembles, moving beyond purely theoretical mathematics to provide a system for evaluating performance characteristics of quantum data transfer. Researchers developed a formalism centered around the moment operator, which encodes statistical properties arising from evolving states within an ensemble and subsequent measurement of the output. This operator is also central to defining t-designs, a concept borrowed from the study of unitary ensembles and now extended to quantum channels.

Theorems accompanying this work have implications for identifying the depolarize ensemble as a natural reference for ensembles of quantum channels, representing a generalization of the Haar ensemble. The authors define an ensemble as an approximate Depolarize t-design with additive error if and only if a specific condition is met, and a relative error if another condition holds, providing operational meaning to these moment operators.

Derivation of the “cHaar” Ensemble for Reference Channels

The cHaar ensemble demonstrates invariance under any ensemble of unitaries, a property important for establishing reference points in quantum channel analysis. This means that transformations applied using these unitaries do not alter the statistical characteristics of the cHaar ensemble itself, providing a stable foundation for comparison. These bounds align with those previously established for unitary ensembles, suggesting a consistent framework for evaluating the accuracy of quantum channel approximations.

Despite expectations, the research indicates the Depolarize ensemble, rather than the cHaar ensemble, emerges as the most suitable reference ensemble for quantum channels, a finding that initially surprised the researchers. The cHaar ensemble possesses a uniform Lebesgue measure over quantum channels when $d_{E}$ = d, but even then, lacks the key characteristics of the Haar or Depolarize ensembles, specifically, invariance under concatenations.

The study’s framework for analyzing moment operators of random quantum channel ensembles led to the proposal of three reference ensembles: Haar, cHaar, and Depolarize. Exact expressions were derived for their spectra, traces, and norms, establishing a clear hierarchy between them. The cHaar ensemble, in particular, interpolates between unitary and depolarizing limits as the environment dimension changes. This outcome underscores the importance of invariance under concatenation as a defining characteristic of effective reference ensembles, even when other properties, such as uniform measure, are present.

Hierarchy of Ensembles: Depolarize Ensemble as a Baseline

The cHaar ensemble’s behavior reveals a surprising connection to both unitary and depolarizing limits, functioning as an interpolating force between them dependent on environment dimension. Different types of noise can decrease the norm of the moment operators, as well as increase it, establishing that the cHaar ensemble converges towards the Depolarize ensemble as system and environment dimensions increase. This convergence isn’t merely a mathematical curiosity; it underpins a strict hierarchy established between ensembles based on their moment operator norms.

This work defines relationships between the Depolarize, cHaar, and Haar ensembles, revealing that the cHaar ensemble acts as an epsilon-approximate Depolarize t-design when the environment dimension, denoted as $d_{E}$ , is considered. Specifically, the approximation holds with an error scaling as O(1/d_E) in the fixed t and the asymptotic limit.

The team’s analysis of eigenvectors further solidifies this hierarchy, showing that the leading eigenvector of the cHaar ensemble converges to the single, fixed point of the Depolarize channel as dimension increases. This convergence is expressed mathematically as a limit, where the eigenvector approaches a state directly related to the depolarizing operation. The authors state, emphasizing its foundational role in the hierarchy.

Impact of Noise on Moment Operator Norms

This framework extends analysis to all moment orders ‘t’, a capability previously unexplored for quantum channels and important for understanding complex quantum dynamics. This behavior is quantified by the scaling of moment operator norms, revealing that unital noise diminishes these norms exponentially with increasing concatenations of unitary layers.

This convergence towards a Depolarize t-design is linked to the support-preserving nature of Haar-distributed unitary ensembles and how unital noise scales locality projectors, reducing their contribution proportionally to the total noise present. Importantly, the derived bounds on moment operator norms also provide limits for the diamond norm and total variational distance, offering a way to assess the fidelity of quantum state transfer.

The analysis shows that, for qudit systems, a depth at least linear in system size is sufficient for Haar ensembles with unital noise to converge to Depolarize t-designs. However, the impact of noise is not always straightforward. Local amplitude damping noise exhibits a surprising effect: moment operator norms initially decrease with small noise strengths, then increase as the noise intensifies. Similarly, local bit-flip errors reduce norms, but in different ways, with the HEA norm converging to the minimum Depolarize ensemble norm while the MAT norm settles at a larger value.

“Certain types of noise can decrease or increase the norm of the moment operator,” the researchers state, highlighting a subtle relationship between circuit design, noise type, and noise strength. These numerical experiments demonstrate that the precise characteristics of a quantum computer and its operating environment are critical considerations for achieving reliable quantum computation.

From Noise-Induced to Channel-Design-Induced Concentration

This defines mirroring the established concept of approximate t-designs for unitary ensembles and providing a benchmark for evaluating the quality and efficiency of quantum data transfer protocols. The analysis reveals that the impact of noise on these moment operator norms is not uniform; certain types of noise can decrease their value, while others increase it. For example, local depolarizing noise channels drive ensemble norms towards the minimum value dictated by the Depolarize ensemble, a result consistent with previous studies of noise-induced phenomena.

Conversely, local amplitude damping noise exhibits the opposite effect, increasing the norms and demonstrating a more complex interplay between noise and channel characteristics. The authors state that phenomena previously attributed solely to noise can be more broadly understood as channel-design-induced phenomena. This generalization suggests that careful channel design can achieve similar effects as mitigating noise, offering an alternative pathway to robust quantum information processing.

Theorem 4, as presented in the work, encompasses prior results on unitary ensembles and extends the analysis to encompass non-unitary concentration phenomena, providing a unified framework for understanding these effects. Numerical studies of parameterized quantum circuits confirm these theoretical findings, demonstrating how both HEA and MAT norms converge with circuit depth towards the Depolarize ensemble under both depolarizing and dephasing noise.

Numerical Studies of Noisy Quantum Circuit Convergence

These findings extend understanding beyond simple noise-induced concentration, suggesting a broader phenomenon of channel-design-induced concentration at play. Experiments focused on parameterized quantum circuits incorporating layers of unitary and fixed noise components, allowing for analysis of statistical convergence as circuit depth increased. Two distinct unitary ansätze, hardware efficient ansätze (HEA) and matchgate ansätze (MAT), were tested under varying noise strengths, providing a comparative assessment of their resilience.

This interplay suggests that optimizing circuits for particular noise profiles could be a key strategy for improving performance and reliability, and that the statistical properties of a circuit are demonstrably affected by its depth and the strength of noise. The work demonstrates that the convergence of a circuit’s statistical properties is a function of circuit depth, and that the type of noise present significantly influences this convergence.

👉 More information
🗞 Moments of Quantum Channel Ensembles
✍️ Matthew Duschenes, Diego García-Martín, Zoë Holmes and M. Cerezo
🧠 DOI: http://link.aps.org/doi/10.1103/grdp-jph9

More like this