Quantum Simulations Gain Accuracy with Newly Found ‘Trotter Scar’ States

Bozhen Zhou and colleagues at the Institute of Theoretical Physics show that specific initial states, termed ‘Trotter scars’, sharply suppress error growth during quantum simulations and exhibit predictable revivals. Their analysis, utilising interaction-picture perturbation theory, reveals a link between a state’s spectral properties and its resilience to ‘Trotter errors’, which commonly plague digital quantum simulation. Furthermore, they present a variational framework to identify these error-minimising states, offering a practical method to improve the accuracy and reliability of quantum computations. Trotter scar identification enables durable quantum simulations with suppressed error growth Error rates dropped to over three orders of magnitude beyond average-case expectations in quantum simulations utilising newly identified ‘Trotter scars’. These states, existing within a vanishingly small fraction of the Hilbert space, the complete set of all possible quantum states for a system, demonstrate a previously unattainable level of durability to ‘Trotter errors’. These errors accumulate when approximating the time evolution of a quantum system using the Trotter decomposition, a standard technique in digital quantum simulation. The Trotter decomposition breaks down a complex quantum process into a sequence of simpler, more manageable steps, introducing inaccuracies with each step. This discovery clarifies how specific initial states can dramatically reduce these errors, a long-standing challenge in digital quantum simulation. They also provide a variational framework, a method of optimisation, to actively seek out these error-minimising states. The identified ‘Trotter scars’ not only suppressed error growth but also induced persistent Loschmidt revivals, the reappearance of the initial quantum state after time evolution, indicating durability against decay and a stable, predictable response over time. Interaction-picture perturbation theory derived an analytical expression linking Trotter error directly to the Hamiltonian’s eigenbasis, revealing that states residing on ‘spectrally commensurate energy ladders’, levels with harmonically related energies, exhibit this enhanced stability. Alastair Peoples (University of Strathclyde) and colleagues applied a newly developed variational framework to spin models like the Heisenberg chain and PXP model, discovering optimised states aligning with these predictions and demonstrating a clear connection between spectral support and reduced error. The significance of this lies in the potential to perform longer and more accurate quantum simulations, crucial for modelling complex physical and chemical systems. Leading-order Trotter error analysis via interaction-picture perturbation theory This work proved central to dissecting how errors accumulate during quantum simulations using interaction-picture perturbation theory. The technique effectively isolates the impact of the Trotter decomposition, which approximates complex quantum movements by breaking them into smaller steps, much like animating a walk with a sequence of still poses. The accuracy of the simulation is directly related to the size of these steps; smaller steps yield greater accuracy but require more computational resources. Analysing the leading-order Trotter error within the eigenbasis of the Hamiltonian, a mathematical description of a quantum system’s total energy, revealed a direct link between a state’s spectral properties and its susceptibility to error. The Hamiltonian governs the time evolution of the quantum system, and its eigenbasis provides a fundamental framework for understanding the system’s behaviour. This approach dissects errors by isolating the effects of approximating quantum movements with discrete steps, differing from alternative methods like linear combinations of unitaries due to its simplicity and low hardware overhead. Linear combinations of unitaries, while potentially more accurate, require significantly more complex quantum circuits and are therefore more demanding on current quantum hardware. Several spin models, including the Heisenberg chain and PXP model, were examined, and the analysis centres on the leading-order Trotter error within the Hamiltonian’s eigenbasis, revealing a connection between a state’s spectral properties and its vulnerability to error. The leading-order error term provides a crucial starting point for understanding and mitigating the overall error accumulation in the simulation. Spectral characteristics define error resilience in quantum systems Despite advances in identifying initial states that minimise errors during quantum simulation, a fundamental challenge remains. Thus far, ‘Trotter scars’, states with specific spectral properties, have been found using variational methods, essentially a trial-and-error approach guided by theory. While effective, this relies on optimising states for a given Hamiltonian and doesn’t explain whether such states are naturally abundant or if their discovery is contingent on a fortunate search. The variational approach involves defining a cost function that quantifies the Trotter error and then iteratively adjusting the parameters of the initial state to minimise this cost function. This process can be computationally expensive, particularly for large quantum systems. Acknowledging that identifying these ‘Trotter scars’ currently relies on somewhat arbitrary initial state selection does not diminish the value of this work. A clear link between a state’s spectral properties and its durability to errors in quantum simulations has been demonstrated for the first time, representing a key step forward. Understanding why certain states perform better allows scientists to move beyond simply finding them, towards designing more robust quantum systems. These states, residing on ‘spectrally commensurate energy ladders’, a series of evenly spaced energy levels, exhibit remarkable durability against errors during complex calculations, demonstrating sharply suppressed error growth and repeatable behaviour known as Loschmidt revival, indicating a stable, predictable response over time. The work reveals that a state’s spectral properties, or its energy level structure, directly influences its ability to withstand inaccuracies introduced by the Trotter decomposition, a common technique used to break down quantum processes into manageable steps. Specifically, states with energy levels that are harmonically related, meaning their energy differences form a simple ratio, appear to be particularly resilient to Trotter errors. This suggests that the structure of the Hamiltonian itself plays a crucial role in determining the existence and stability of these error-resilient states. Further research is needed to determine the prevalence of these ‘Trotter scars’ in more complex systems and to develop more efficient methods for identifying and exploiting them, potentially paving the way for fault-tolerant quantum computation. The research demonstrated that initial states supported on spectrally commensurate energy ladders exhibit suppressed error growth and persistent Loschmidt revivals during quantum simulation. This finding reveals a spectral origin for resilience to Trotter errors, which are inaccuracies arising from a common technique used to simplify quantum calculations. Researchers developed a variational framework to identify error-minimizing initial states for a given Hamiltonian, successfully finding optimized states that align with theoretical predictions. This work provides a strategy for discovering states that are more durable to errors, improving the reliability of digital quantum simulation. 👉 More information 🗞 Trotter Scars: Trotter Error Suppression in Quantum Simulation 🧠 ArXiv: https://arxiv.org/abs/2603.29857 Tags: