Fujitsu Research of America has developed a quantum simulation method achieving a cost scaling of O(T^(1+o(1)) (log(1/epsilon))^2 / epsilon) for estimating observables, a quantifiable improvement over previous interpolation methods. The approach avoids requiring additional block-encoding qubits, offering an advantage over Quantum Signal Processing techniques while maintaining scaling with polylog(1/epsilon). This is accomplished through cardinal sine interpolation, which results in a slightly better (log(1/epsilon))^2 scaling compared to methods using Chebyshev interpolation and product formulas. The work provides an alternative to directly running fractional query evolutions, instead extrapolating results from whole-number time steps.

Trotter Method Accuracy Gains via Stable Interpolation

Cardinal sine interpolation offers a refinement in calculating fractional queries for quantum simulations, achieving a (log(1/epsilon))^2 scaling that surpasses the (log(1/epsilon))^3 scaling of Chebyshev interpolation and product formulas. This improvement stems from an optimized and variable order, denoted as p, which adapts to the simulation’s time horizon, T. The technique provides an alternative to Quantum Signal Processing for determining observables or amplitudes following a fractional query evolution, accomplishing this with a cost of O(T log(1/epsilon)).

The work presented builds upon the established Trotter method, a technique that simplifies quantum evolution by breaking it into smaller, manageable steps. While the Trotter method is hardware-friendly, its accuracy is limited by the accumulation of errors with each step, necessitating increasingly complex circuits to maintain precision. This new method circumvents that limitation, achieving higher precision with fewer computational resources, a critical advantage for early error-corrected quantum computers.

This advancement is not merely theoretical. It addresses a practical challenge in quantum simulation, namely the trade-off between accuracy and resource consumption. Researchers have previously explored methods like those detailed by Chuang in 2017 and Gilyén et al. in 2019, but these often require additional qubits or complex circuitry.

The current work, alongside improvements using Chebyshev interpolation as described by Rendon, Watkins and Wiebe in 2024, demonstrates a path toward more efficient and scalable quantum simulations. The simplicity of the approach, combined with its performance gains, positions it as a promising tool for exploring complex quantum systems with limited hardware resources.

Cardinal Sine Interpolation Enables Fractional Query Evolution

The method’s efficiency is further enhanced by its ability to maintain scaling with polylog(1/epsilon) without requiring additional qubits, a critical resource in quantum computing. Rather than directly executing fractional query evolutions on quantum hardware, the technique extrapolates results from whole-number steps, effectively reconstructing the desired fractional operation. A classical computer then combines measurements taken at these whole-number steps using cardinal sine weights, mirroring the process of reconstructing a smooth sound wave from discrete samples in digital audio.

This reconstruction process is stabilized by a Gaussian window, which minimizes the number of samples needed while maintaining accuracy and resilience against measurement noise; the resulting method circumvents the need for direct quantum execution of fractional operations.

“Each vertical line is a circuit that repeats the Trotter step a whole number of times,” explains the research, “and the dots are the measured results.” The team’s approach is particularly valuable for simulating natural phenomena, such as molecular reactions and material conductivity, by allowing researchers to predict system evolution and measure properties of interest without extensive quantum hardware requirements.

Cost Scaling of O(T^(1+o(1)) (log(1/epsilon))^2 / epsilon) for Observable Estimation

This improvement in efficiency stems from the use of cardinal sine interpolation within product formulas, allowing for a more streamlined calculation of fractional queries. This technique distinguishes itself by requiring no additional block-encoding qubits, a benefit compared to Quantum Signal Processing methods. The resulting algorithm estimates quantities after quantum time evolution with a cost that grows nearly linearly with simulation time and polylogarithmically with desired precision, matching the performance of advanced methods.

“These features make the approach well suited to early fault-tolerant quantum computers, where every qubit counts,” the researchers write, highlighting the method’s practicality for near-term quantum hardware. As a byproduct of this work, a simplified method for evaluating “fractional” quantum operations has also been developed, a long-standing component in quantum algorithms. The team’s work, detailed in Quantum, offers a new pathway for observable estimation, relying solely on Trotter circuits and two extra qubits.

Gaussian Windows Optimize Sample Reduction for Reconstruction

By strategically combining measurements from whole-number Trotter steps, the method circumvents the need to directly implement fractional quantum operations, a common bottleneck in quantum simulation. Central to this advancement is the use of a Gaussian window, which shapes the measured results before they are combined on a classical computer. This windowing process, paired with cardinal sine weights, accurately reconstructs the outcome of a fractional operation without requiring the quantum computer to perform it directly.

👉 More information
🗞 All you need is Trotter
✍️ Gumaro Rendón
🧠 DOI: https://quantum-journal.org/papers/q-2026-11-06-2261/

More like this