Researchers at the University of Copenhagen and Centrum Wiskunde & Informatica have demonstrated a way to systematically reduce errors during the preparation of complex quantum states. Their work, combining numerical calculations with simulations on a Rydberg atom quantum simulator, reveals that smoother changes to the Hamiltonian, specifically, a time-dependent Hamiltonian with vanishing first to nth order derivatives at the beginning and end of a process, can lower infidelity to a scaling of 1/T^(n+1), where T is the evolution time. The team investigated this effect using a one-dimensional mixed-field Ising model and a chain of Rydberg atoms, finding that careful scheduling of the Hamiltonian suppresses end-to-end transfer error.

Smooth Hamiltonian Boundaries Minimize Quantum State Error

This mathematical relationship demonstrates a systematic way to minimize error by carefully controlling the Hamiltonian’s behavior over time. Numerical calculations and simulations performed on a chain of Rydberg atoms validated this theoretical prediction, establishing a link between smooth Hamiltonian scheduling and suppressed end-to-end transfer error. This level of control is important because the adiabatic theorem, a cornerstone of quantum state preparation, relies on slowly varying Hamiltonians to keep a system in its corresponding instantaneous eigenstate.

However, even small deviations from this ideal can introduce errors, particularly in complex many-body systems. The research extends previous rigorous estimates of error bounds, which previously stood at O(ε^n), to now demonstrate an improvement to O(ε^(n+1)) when the Hamiltonian possesses ‘n’ vanishing time derivatives at both the beginning and end of the evolution.

This means that increasing the smoothness of the Hamiltonian’s initial and final conditions yields a proportionally faster reduction in error as the evolution time increases. A key finding centers on the importance of vanishing derivatives; the more derivatives that equal zero at the beginning and end of the Hamiltonian path, the more effectively errors are suppressed.

“Crucial to this mechanism is the requirement that the adiabatic eigenstates connect smoothly to the initial and final state, i.e., that the time dependent Hamiltonian evolves in a smooth manner throughout the entire process,” the paper states, emphasizing the need for continuity in the system’s evolution. The work reveals that the loss of population from the adiabatic state at large positive times is exponentially small, proportional to exp(-V/α). This exponential suppression, similar to that observed in other analytically solvable two-level problems, is achieved through superadiabatic basis transformations that absorb low-order non-adiabatic corrections.

The researchers proved a rigorous adiabatic theorem showing that ‘n’ vanishing time derivatives of the Hamiltonian at the initial and final times give an error bound of O(ε^(n + 1)). This rigorous approach provides a quantifiable measure of how smoothness directly translates into improved fidelity in adiabatic quantum state preparation, offering a pathway toward more reliable quantum technologies.

Rigorous Adiabatic Theorem Improves Error Scaling to 1/T^(n+1)

The research demonstrates that carefully crafted initial and final conditions for the Hamiltonian can significantly suppress end-to-end transfer error, even in the presence of unavoidable dissipative losses. The team’s approach combined rigorous mathematical proof with simulations performed on a Rydberg atom quantum simulator, allowing for direct comparison between theoretical predictions and experimental observations.

The work builds upon existing estimates of error bounds, previously established at O(ε^n), by demonstrating an improvement to O(ε^(n+1)), where ε represents 1/T. This refinement is not merely theoretical. The simulations show that enforcing vanishing derivatives at the boundaries of the Hamiltonian path reduces infidelity by orders of magnitude in the polynomial regime, the range where the error decays polynomially with time, without negatively impacting performance in the exponential regime, where faster passage is prioritized.

The theoretical foundation rests on a refined adiabatic theorem, which describes how a quantum system evolves when subjected to a slowly changing Hamiltonian. However, the researchers emphasize that the benefit of this approach is not universally guaranteed; the polynomial scaling of the error only dominates for sufficiently small values of ε. For larger ε, the error still decays exponentially, governed by the minimal spectral gap between energy levels and the Hamiltonian’s analytic properties.

“The scaling O(ε^(n + 1)) dominates only for sufficiently small ε,” the paper notes, clarifying the conditions under which the improved error bound is most effective. The team constructed specific schedule functions designed to impose these vanishing boundary derivatives while minimizing disruption to the Hamiltonian’s behavior during the central portion of the evolution. This careful design ensures that the benefits of smoother transitions are realized without compromising the overall process.

The simulations confirm that these boundary conditions significantly reduce final infidelity in a noiseless setting, demonstrating the practical implications of the theoretical findings. The combination of rigorous proof and experimental validation using Rydberg atoms provides a strong foundation for future advancements in adiabatic quantum computation and state preparation techniques.

Mixed-Field Ising Model Demonstrates Boundary Condition Control

The research, focused on a one-dimensional mixed-field Ising model, details how carefully constructed schedule functions can control the number of vanishing derivatives at boundaries while preserving the core dynamics of the system. Investigations into the mixed-field Ising chain, defined by nearest-neighbor couplings and spin-1/2 operators, showed that the peak infidelity occurred earlier in time for certain schedules, specifically those passing through the phase transition more rapidly.

The simulations used matrix product states to model the dynamics, revealing that a linear schedule resulted in an immediate jump in infidelity that never fully recovered, while schedules designed with vanishing boundary derivatives exhibited a smoother increase and decrease in error within the polynomial regime.

This difference is visually apparent in log-log plots of final infidelity versus 1/T for a chain length of 11, where the smoother schedules consistently outperformed the linear approach. The team’s work extends beyond theoretical modeling, incorporating experimental validation on a Rydberg atom chain implemented on the Aquila neutral-atom quantum simulator. These experiments tested the schedule constructions under realistic conditions, including noise and measurement errors, confirming the benefits of smoother transitions in a practical setting.

The simulations and experiments demonstrate that the improvement in fidelity is achievable with minimal modification to existing schedules, offering a readily implementable strategy for enhancing quantum computation. By constructing schedule functions that impose these boundary conditions while minimally altering the Hamiltonian at intermediate times, the researchers were able to reduce final infidelity by orders of magnitude in noiseless simulations.

Rydberg Atom Simulator Validates Schedule Function Performance

Specifically, the work demonstrates that controlling the rate of change of the Hamiltonian, the system’s energy landscape, at the start and end of a process is critical for minimizing the final infidelity, or the deviation from the desired quantum state. Researchers used a Rydberg atom quantum simulator to validate these findings, achieving improved performance through the construction of “smooth” schedule functions.

This means that the energy landscape changes gradually, preventing abrupt transitions that introduce errors. A Hamiltonian with ‘n’ vanishing time derivatives exhibits an infidelity scaling of 1/T^(n+1), where T represents the total evolution time; this precise scaling demonstrates how error systematically decreases with smoother Hamiltonian changes.

Polynomial and Exponential Regimes Govern Adiabatic Error Decay

The precision with which a Hamiltonian changes over time dictates the rate at which errors accumulate during quantum state preparation, according to new work detailing distinct regimes of error decay. The researchers demonstrate that the end-to-end transfer error, for a given process duration and dissipative losses, can be suppressed by adopting smooth initial and final scheduling functions for the Hamiltonian. This relationship between smoothness and error reduction isn’t universally applicable; the scaling O(ε^(n + 1)), where ε equals 1/T, holds only when evolution is sufficiently slow.

For faster evolution, an exponential decay dominates, with the rate influenced by system size and the minimal spectral gap of the Hamiltonian. The number of vanishing derivatives did not significantly affect the exponential decay rate, but demonstrably governed infidelity within the polynomial regime.

The work highlights that the choice of schedule function is critical, and that carefully crafted boundaries can significantly enhance the fidelity of adiabatic state preparation, even without altering the dynamics during the bulk of the process. The study’s results confirm that the polynomial regime is governed by the value of ‘n’, the number of vanishing derivatives, rather than the system size, offering a predictable path towards improved quantum control.

👉 More information
🗞 Adiabatic Preparation of Many-Body Quantum States: Getting the Beginning and the End Right
✍️ Emil T. M. Pedersen, Freek Witteveen, Klaus Mølmer and Matthias Christandl
🧠 DOI: http://link.aps.org/doi/10.1103/hnlr-hm29

More like this