Unitarity Estimation for Quantum Channels

Kean Chen, Qisheng Wang, Peixun Long, Mingsheng Ying

TL;DR

This paper tackles the crucial problem of estimating how "unitary" an unknown quantum channel is, a key task for certifying quantum devices. The authors introduce a unified framework that achieves optimal query complexity for both coherent and incoherent access, significantly improving efficiency and practicality, especially for near-term quantum computers.

What Problem Does This Paper Address?

Quantum computers rely on precise operations, and a fundamental aspect of these operations is their "unitarity." A perfectly unitary quantum channel is one that preserves quantum information and is reversible. However, real-world quantum devices are imperfect and introduce noise, deviating from ideal unitary behavior. Estimating this deviation, or "unitarity," is vital for:

  • Device Certification: Verifying if a quantum device behaves as expected, even without knowing its exact internal workings.
  • Benchmarking: Quantifying the quality of quantum operations and comparing different devices.
  • Understanding Noise: Distinguishing between coherent (e.g., over-rotation) and incoherent (e.g., depolarization) noise, which have different implications for error correction.

The challenge lies in efficiently measuring this unitarity. Traditional methods like quantum process tomography are resource-intensive. Prior work proposed using the "unitarity" as a measure, often relying on the Jamiołkowski state of the channel. However, these methods either required significant ancilla systems or were not practical for near-term devices due to the need for "coherent access" (where the device's output can be used to control subsequent operations). This paper aims to develop more efficient and practical algorithms for unitarity estimation, considering both coherent and incoherent access scenarios.

Methods

The authors introduce a unified framework for unitarity estimation by reformulating the problem. Instead of directly working with the complex Jamiołkowski states, they leverage two key properties of unitary channels:

  1. Purity-Preservation: Unitary channels preserve the purity of pure states.
  2. Orthogonality-Preservation: Unitary channels preserve the orthogonality of orthogonal states.

They define "purity-preservation index" ($p(\mathcal{E})$) and "orthogonality-preservation index" ($o(\mathcal{E})$) to quantify these properties. A crucial insight is that the channel's unitarity ($u(\mathcal{E})$) can be expressed as a linear combination of these two indices: $u(\mathcal{E}) = p(\mathcal{E}) - \frac{1}{d^2} o(\mathcal{E})$.

This reformulation allows for more efficient estimation:

  • Coherent Access: They use a SWAP test on the outputs of the channel applied to carefully chosen states derived from random unitaries. This approach requires only an $O(d)$-dimensional ancilla system, a significant improvement over previous methods needing $O(d^3)$.
  • Incoherent Access: For this more restricted setting (where outputs cannot be directly used for subsequent operations), they adapt a distributed quantum inner product estimation (DQIPE) algorithm. This method is non-adaptive and requires no ancilla system.

The paper also establishes lower bounds for these tasks, demonstrating the optimality of their proposed algorithms. A key part of this involves a novel reduction from a "depolarizing vs. unitary channel" problem to unitarity estimation, leading to improved lower bounds.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Unitarity as Purity/Orthogonality Preservation: Reframing unitarity estimation in terms of purity and orthogonality preservation is a powerful simplification, enabling more efficient algorithms.
  • Ancilla-Efficient Coherent Access: The use of a standard SWAP test on channel outputs, rather than Jamiołkowski states, dramatically reduces the ancilla system requirement for coherent access.
  • Practical Incoherent Access: Developing an ancilla-free, non-adaptive algorithm for incoherent access makes unitarity estimation more feasible for near-term quantum devices.
  • Optimal Complexity: Proving matching lower bounds for both coherent and incoherent access confirms that their algorithms achieve the best possible query complexity.
  • Improved Lower Bound for Distinguishing Channels: The paper provides a new, tighter lower bound for distinguishing between depolarizing and unitary channels, a problem related to unitarity estimation.

Key Results

The paper presents the following optimal query complexities for estimating the unitarity of a $d$-dimensional quantum channel $\mathcal{E}$ to precision $\epsilon$:

  • Coherent Access: $O(\epsilon^{-2})$ calls to $\mathcal{E}$ with an $O(d)$-dimensional ancilla system.
  • Incoherent Access: $O(\sqrt{d} \cdot \epsilon^{-2})$ calls to $\mathcal{E}$ with no ancilla system.

Furthermore, they establish the following lower bounds:

  • Coherent Access: $\Omega(\epsilon^{-2})$ calls to $\mathcal{E}$.
  • Incoherent Access: $\Omega(\sqrt{d} + \epsilon^{-2})$ calls to $\mathcal{E}$.

They also settle the query complexity for distinguishing depolarizing and unitary channels with incoherent access, providing a lower bound of $\Omega(\sqrt{d})$.

Limitations and Future Work

  • Alternative Unitarity Definition: The paper discusses an alternative definition of unitarity ($u'(\mathcal{E})$) related to noise coherence. They show that their developed algorithms and bounds also apply to this alternative definition, but require slightly more sophisticated strategies for the general case.
  • Further Improvements for Incoherent Access: While their incoherent access algorithm achieves $O(\sqrt{d} \cdot \epsilon^{-2})$, they conjecture that it might be possible to improve this further, perhaps to $O(\sqrt{d} \cdot \epsilon^{-\alpha} + \epsilon^{-2})$ for some $\alpha < 2$, with more refined error analysis.
  • Generalization to Other Properties: The techniques used for unitarity estimation might be generalizable to estimating other properties of quantum channels.
  • Practical Implementation: While the algorithms are designed with near-term devices in mind, practical implementation challenges such as state preparation, measurement fidelity, and noise in the estimation process itself would need to be addressed.

Summary updated

Abstract

Estimating the unitarity of an unknown quantum channel $\mathcal {E}$ provides information on how much it is unitary, which is a basic and important problem in quantum device certification and benchmarking. Unitarity estimation can be performed with either coherent or incoherent access, where the former in general leads to better query complexity while the latter allows more practical implementations. In this paper, we provide a unified framework for unitarity estimation, which induces ancilla-efficient algorithms that use $O(\epsilon ^{-2})$ and $O(\sqrt {d}\cdot \epsilon ^{-2})$ calls to $\mathcal {E}$ with coherent and incoherent accesses, respectively, where $d$ is the dimension of the system that $\mathcal {E}$ acts on and $\epsilon $ is the required precision. We further show that both the $d$ -dependence and $\epsilon $ -dependence of our algorithms are optimal. As part of our results, we settle the query complexity of the distinguishing problem for depolarizing and unitary channels with incoherent access by giving a matching lower bound $\Omega (\sqrt {d})$ , improving the prior best lower bound $\Omega (\sqrt [{3}]{d})$ by (Aharonov et al., 2022) and (Chen et al., FOCS 2021).
BibTeX
@article{chen2022unitarity,
  author = {Kean Chen and Qisheng Wang and Peixun Long and Mingsheng Ying},
  title = {{Unitarity Estimation for Quantum Channels}},
  journal = {IEEE Trans. Information Theory},
  year = {2022},
  doi = {10.48550/arxiv.2212.09319},
}
Figure