Surface code sampling may need half the quantum circuit depth

Ivan H.C. Shum and Angela Capel of the University of Cambridge have developed hybrid computation algorithms that significantly reduce the quantum circuit depth needed for Gibbs sampling of stabilizer codes. The work demonstrates circuits achieving approximately L/2 quantum circuit depth to prepare the Gibbs state of the rotated surface code Hamiltonian, where L is the side of the square lattice.

Researchers also report a circuit depth of L for the toric code Hamiltonian, improving upon existing methods. These algorithms use local quantum algorithms and, with non-local gates, can prepare the Gibbs state of the periodic 1D Ising model in logarithmic depth using linearly many simultaneous measurements.

Rotated Surface and Toric Codes Map to Decoupled Ising Models

This approach uses the equivalence between these codes and classical statistical models, enabling the use of well-established sampling methods to generate the desired quantum states. The algorithms build upon prior work establishing connections between stabilizer codes and classical statistical mechanics, specifically focusing on preparing the Gibbs distribution for the rotated surface code and the toric code. This decoupling simplifies the sampling process, reducing it to a classical sampling problem with subsequent quantum gate operations. Specifically, the process involves obtaining the syndrome, followed by classical XOR manipulations, a layer of Hadamard gates, and then O(L) layers of controlled-NOT (CX) gates. The researchers note that the equivalence of the rotated surface code to a uniform magnetic field, achievable through SWAP gates, simplifies ground state preparation compared to direct logical encoding of information in qubits.

This optimization stems from the ability to use the inherent quantum properties of the algorithms, which act on superpositions of computational basis states and create entanglement, while still relying on a relatively shallow circuit. “There is an O(L) depth local mapping from both the 2D toric code and the 2D rotated surface code, to two decoupled systems that can be identified as 1D Ising models,” the paper states, highlighting the core of the efficiency gain. The team’s work demonstrates a method for efficiently simulating quantum systems, potentially accelerating progress in quantum error correction and fault-tolerant quantum computation.

O(L) Depth Circuits Encode Excited Stabilizer Code States

Algorithms for preparing Gibbs states of stabilizer codes now require potentially half the quantum circuit depth previously needed, according to new work detailing a hybrid computation approach. The research demonstrates circuits achieving O(L) depth for encoding excited states, given a sample of the syndrome, a reduction stemming from deferring state preparation to classical pre-processing. This decoupling allows for efficient classical simulation before quantum entanglement, streamlining the process and reducing the computational load. For the rotated surface code, the team’s circuits achieve a depth of O(L), improving upon the previously known O(L^2) benchmark, while the 2D toric code benefits from an O(L) depth circuit to encode excited states, a significant advancement over the prior O(L^3) result. This improvement is rooted in identifying surface code Hamiltonians as unitarily equivalent to decoupled classical Ising models, one each for the X and Z subsystems, simplifying the sampling process.

Extending beyond these codes, the algorithms also yield a logarithmic depth circuit for Gibbs sampling the one-dimensional Ising chain, but with a trade-off. This requires linearly many simultaneous measurements of all relevant qubits. This strategy allows for operations to remain within the computational basis, facilitating efficient classical simulation. The team acknowledges that enforcing locality in the algorithms, as seen in prior work on toric code encoding, simplifies realization, even though it may not always be optimal. The reduction in circuit depth, the researchers claim, is due to deferring most of the state preparation before entanglement to classical pre-processing, staying in the computational basis and hence being perfectly possible to simulate efficiently classically.

Classical Pre-processing Minimizes Quantum Circuit Depth

Deferring most quantum state preparation to classical pre-processing can reduce circuit depth for stabilizer code Hamiltonians, according to a new analysis of surface code sampling algorithms. The work demonstrates that, for the rotated surface code, a quantum local circuit with approximately half the depth of current lowest-depth circuits is sufficient for classical sampling results. This reduction stems from performing initial state preparation classically while remaining within the computational basis, a strategy enabling efficient simulation. This decoupling simplifies the quantum portion of the algorithm, focusing it on entanglement rather than complex state preparation. This is followed by a post-processing circuit of O(L) depth, streamlining the overall process.

Logarithmic Depth Gibbs Sampling of the 1D Ising Model

Allowing for non-local gates enables the preparation of the Gibbs state of a periodic 1D Ising model with logarithmic quantum circuit depth, alongside a linearly scaling number of simultaneous measurements. The research details a method to achieve this efficiency by reversing the typical process of Gibbs sampling, beginning with the system’s Hamiltonian and decoupling it into subsystems acting on distinct qubits. This decoupling focuses on creating subsystems composed exclusively of X and Z operators, simplifying the quantum portion of the algorithm and shifting the computational burden towards classical processing.

Local Measurements Generate Bernoulli Variables for Quantum Sampling

Researchers report achieving a circuit depth of roughly L/2, where L represents the side length of the square lattice, for this preparation, a potentially significant reduction in computational cost for error correction schemes. This efficiency stems from a hybrid computation approach that combines quantum processing with classical pre-processing steps, streamlining the sampling process. A key element of the approach is the generation of statistics for multiple independent Bernoulli variables using a non-Clifford circuit of depth one, coupled with simultaneous measurements across all qubits. The researchers demonstrated a local algorithm capable of sampling the Gibbs state of the toric code with an O(L) depth pre-processing circuit, followed by 2L2 + O(1) measurements, of which a substantial portion, 2L2–2, can be performed simultaneously. The choice to enforce locality in these algorithms isn’t merely a practical consideration for realizing logical state encoding; it also simplifies the process, as is demonstrated in previous work.

Stabilizer Hamiltonian Decoupling Simplifies Gibbs State Preparation

This reduction in computational complexity stems from decoupling the stabilizer Hamiltonian into subsystems acting on disjoint qubits, simplifying the sampling process. This decoupling relies on transforming the Hamiltonian into a form where the computational basis aligns with its eigenbasis, shifting a substantial portion of the computation to classical processing. The team’s method uses controlled-NOT (CX) gates to separate the X and Z operator subsystems, followed by Hadamard gates applied to qubits hosting X operators, resulting in a Hamiltonian readily sampled using classical techniques.

Stay current

See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.

Avatar of The Quant

The Quant

The Quant possesses over two decades of experience in start-up ventures and financial arenas, brings a unique and insightful perspective to the quantum computing sector. This extensive background combines the agility and innovation typical of start-up environments with the rigor and analytical depth required in finance. Such a blend of skills is particularly valuable in understanding and navigating the complex, rapidly evolving landscape of quantum computing and quantum technology marketplaces. The quantum technology marketplace is burgeoning, with immense growth potential. This expansion is not just limited to the technology itself but extends to a wide array of applications in different industries, including finance, healthcare, logistics, and more.

Latest Posts by The Quant: