Ancilla-driven quantum computation for qudits and continuous variables

Timothy Proctor, Melissa Giulian, Natalia Korolkova, Erika Andersson, Viv KendonPhysical 2017 · DOI

TL;DR

This paper introduces a new framework for quantum computation that uses "ancilla" systems to perform operations on a primary "quantum memory" register. This approach is designed to be universal and work with various quantum systems, including qubits, higher-dimensional qudits, and continuous variables, offering a flexible path towards scalable quantum computing.

What Problem Does This Paper Address?

Building a universal quantum computer is a major challenge, with decoherence and the complexity of controlling quantum systems being key hurdles. While qubits are the standard building blocks, higher-dimensional systems (qudits) and continuous variables (QCVs) offer potential advantages like improved fault-tolerance thresholds and robustness. However, existing quantum computation models often assume qubits or don't fully generalize to these more complex systems. This paper addresses this by proposing a unified method for ancilla-driven quantum computation that works across these different quantum variable types, aiming for a more scalable and flexible quantum computing architecture.

Methods

The core of this work is an "ancilla-driven quantum computation" (ADQC) model. The key idea is to isolate the main quantum memory register and use auxiliary systems (ancillas) to perform the necessary entangling gates. The proposed method relies on:

  • A fixed ancilla-register interaction gate: A single, specific type of interaction between an ancilla and a register quantum variable.
  • Ancilla preparation: Ancillas are prepared in a specific, fixed state (e.g., $|+0\rangle$).
  • Local ancilla measurements: Measurements are performed on the ancillas.

To ensure deterministic computation, the model incorporates classical feed-forward of measurement outcomes, similar to measurement-based quantum computation (MBQC). This allows for the correction of errors introduced by the probabilistic nature of measurements.

The paper also presents a "minimal control" model, which is a variation that replaces the need for complex ancilla measurements with the requirement for ancillas to be prepared in specific states from an orthonormal basis. This is particularly useful in scenarios where high-fidelity ancilla measurements are difficult.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Unified formalism for diverse quantum systems: The paper introduces a "quantum variable" (QV) formalism that elegantly unifies qubits, qudits, and QCVs, allowing for a single computational framework to be applied to all.
  • Minimal ancilla interaction for universality: The ADQC model achieves universal quantum computation with a remarkably simple set of operations: a single fixed ancilla-register interaction, ancilla preparation, and local ancilla measurements. This minimizes the complexity of controlling the primary quantum memory.
  • Hybrid quantum-classical processing: The model leverages classical feed-forward to overcome the probabilistic nature of quantum measurements, mirroring the advantages of MBQC and enabling deterministic computation.
  • Parallelism for Clifford gates: The ADQC model, like MBQC, allows for the implementation of any Clifford circuit in essentially a single layer of quantum computation, offering significant speedups for certain types of quantum algorithms.
  • Alternative for measurement-challenging scenarios: The "minimal control" model provides a viable path to universal computation when high-quality ancilla measurements are not feasible, relying instead on state preparation.

Key Results

The paper demonstrates that the proposed ADQC model, using a fixed ancilla-register interaction (like a generalized controlled-Z gate or a SWAP-based interaction), can implement universal quantum computation for qubits, qudits, and QCVs. Specifically:

  • Entangling gates: The model can implement entangling gates on register QVs by interacting them with ancillas and performing specific measurements.
  • Single-QV gates: Various single-QV gates, including generalized rotation gates, can be implemented by adapting the ancilla measurement or preparation.
  • Universality: The combination of these entangling and single-QV gates, along with classical feed-forward, is sufficient for universal quantum computation.
  • Physical implementations: The paper discusses potential physical realizations in optical systems for QCVs and atomic or superconducting systems for qudits, highlighting the feasibility of the proposed methods.

Limitations and Future Work

  • Physical realization challenges: While the models are theoretically sound, achieving high-fidelity ancilla-register interactions and measurements remains a significant experimental challenge, particularly for QCVs and high-dimensional qudits.
  • Finite squeezing in QCVs: For QCVs, approximating the ideal ancilla state $|+0\rangle$ with squeezed states introduces Gaussian noise, which can accumulate. The paper notes that fault-tolerance thresholds for this noise are known but may be above current experimental capabilities. Extending fault-tolerance techniques to the ADQC model for QCVs is suggested as future work.
  • Hybrid systems: While the paper briefly touches upon hybrid systems (different types of QVs for ancilla and register), achieving universal computation in all such hybrid settings requires further investigation.
  • Parameter optimization: For the "minimal control" model, finding optimal parameters for the interaction gate to ensure universality across all qudit dimensions requires more thorough investigation.

Summary updated

Abstract

Although qubits are the leading candidate for the basic elements in a quantum computer, there are also a range of reasons to consider using higher-dimensional qudits or quantum continuous variables (QCVs). In this paper, we use a general ``quantum variable'' formalism to propose a method of quantum computation in which ancillas are used to mediate gates on a well-isolated ``quantum memory'' register and which may be applied to the setting of qubits, qudits (for $d>2$), or QCVs. More specifically, we present a model in which universal quantum computation may be implemented on a register using only repeated applications of a single fixed two-body ancilla-register interaction gate, ancillas prepared in a single state, and local measurements of these ancillas. In order to maintain determinism in the computation, adaptive measurements via a classical feed forward of measurement outcomes are used, with the method similar to that in measurement-based quantum computation (MBQC). We show that our model has the same hybrid quantum-classical processing advantages as MBQC, including the power to implement any Clifford circuit in essentially one layer of quantum computation. In some physical settings, high-quality measurements of the ancillas may be highly challenging or not possible, and hence we also present a globally unitary model which replaces the need for measurements of the ancillas with the requirement for ancillas to be prepared in states from a fixed orthonormal basis. Finally, we discuss settings in which these models may be of practical interest.
BibTeX
@inproceedings{proctor2017ancilla,
  author = {Timothy Proctor and Melissa Giulian and Natalia Korolkova and Erika Andersson and Viv Kendon},
  title = {{Ancilla-driven quantum computation for qudits and continuous variables}},
  booktitle = {Physical},
  year = {2017},
  doi = {10.1103/physreva.95.052317},
}

Loading related content...

Figure