Quantum Mpemba Effect's Surprising Noise Resilience

The quantum Mpemba effect (QME) represents a fascinating counterintuitive phenomenon in quantum physics, where systems starting farther from equilibrium can thermalize faster than those closer to it, mirroring the classical Mpemba effect in which hot water freezes quicker than cold water. This quantum analog has garnered significant interest for its potential to revolutionize quantum computing tasks, such as thermodynamic computing, where computation speed is limited by thermalization rates. In the noisy intermediate-scale quantum (NISQ) era, where error-free state preparation is notoriously difficult, the practical utility of QME hinges on its robustness to imperfections. A recent study by Mackinnon and Paternostro investigates how state-preparation errors impact two distinct approaches to achieving QME, revealing stark differences in their susceptibility to noise and offering critical insights for future quantum technologies. By analyzing open and closed quantum systems, the research underscores the importance of error resilience in harnessing quantum effects for real-world applications, potentially accelerating advancements in quantum simulation and computation.

The study employs a rigorous ology to assess the robustness of QME across two primary models: an open quantum system described by the Gorini-Kossakowski-Sudarshan-Lindblad (GKSL) master equation and a closed many-body system using U(1) symmetric random unitary circuits (RUCs). For the open system, the approach involves preparing an initial state orthogonal to the slowest decaying mode of the Liouvillian dynamics, as detailed in prior work, to achieve exponential acceleration in thermalization. This is implemented in specific scenarios like the Dicke model with N=40 spins and a quantum harmonic oscillator coupled to a thermal bath, where unitary rotations are applied to eliminate overlaps with slow modes. In contrast, the closed system model utilizes a chain of spin-1/2 particles under RUCs with U(1) symmetry, where initial states are tilted ferromagnetic states that break symmetry to varying degrees. The researchers introduce state-preparation errors by perturbing these unitary rotations, using parameters like ε to quantify deviations, and employ measures such as the Hilbert-Schmidt distance and entanglement asymmetry to track thermalization dynamics and symmetry restoration, ensuring a comprehensive evaluation of noise impact.

From the open-system model demonstrate a high sensitivity to state-preparation errors, with thermalization acceleration diminishing exponentially as error increases. In the Dicke model, for instance, accurately prepared states show a crossing time tc around 95 units, indicating QME, but even small errors in the rotation parameter s (e.g., ε=0.01) cause a rapid return to slower thermalization rates, as depicted in Figure 2. The thermalization time teq increases exponentially with ε, and the relative speed-up decays sharply, as shown in Figures 3 and 4, with variations attributed to random state coefficients rather than system size. Similarly, in the quantum harmonic oscillator model, diagonalized states achieve exponential speed-ups, but perturbations via QR decomposition lead to a comparable exponential decay in acceleration, confirming that this vulnerability is inherent to the open-system approach and not model-specific. These highlight that minor inaccuracies in state preparation can nullify the benefits of QME in open systems, posing significant s for practical implementation.

In stark contrast, the closed-system model exhibits remarkable robustness to state-preparation errors, with thermalization rates largely unaffected or even enhanced by noise. For tilted ferromagnetic states in U(1) symmetric RUCs, small errors (ε on the order of 10^-2) have minimal impact on entanglement asymmetry and symmetry restoration, as illustrated in Figure 7. Notably, larger errors (ε around 10^0) can increase the rate of thermalization, particularly for states with lower initial symmetry breaking, such as those with tilting angle θ=0.2π. Analysis of overlap probabilities pq with charge sectors reveals that noise drives the system toward higher overlaps with sectors of greater dimension Dq, boosting the weighted average E[D] and accelerating thermalization, as shown in Figure 8. This behavior aligns with the statistical effect of Haar-random unitaries, which skew initial states toward configurations that favor faster coherence spreading and symmetry restoration, underscoring the model's inherent error tolerance and potential for noise-assisted performance improvements.

Of these are profound for quantum computing and simulation, particularly in the NISQ era where noise is ubiquitous. The robustness of the closed-system approach to QME suggests that strategies leveraging symmetry breaking in random circuits could be more feasible for real-world applications, enabling faster thermalization in quantum tasks like thermodynamic computing without stringent error-correction requirements. Conversely, the fragility of the open-system underscores the need for high-precision state preparation, which may be impractical with current technology, urging a shift toward more resilient quantum algorithms. This research not only advances our understanding of non-equilibrium quantum dynamics but also provides a framework for evaluating the practicality of quantum effects, potentially guiding the development of error-robust quantum devices and accelerating progress in fields such as quantum machine learning and material science.

Despite its insights, the study has limitations, including its focus on specific models like the Dicke model and harmonic oscillator, which may not capture all possible quantum systems exhibiting QME. The analysis assumes Markovian dynamics and U(1) symmetry, and generalizations to other symmetries or non-Markovian environments could yield different robustness profiles. Additionally, the use of finite-dimensional approximations for infinite systems, such as the truncated harmonic oscillator, might not fully represent real-world scenarios. Future work should explore a broader range of models and error types, such as coherent errors or environmental noise, to develop a universal characterization of QME robustness and its applications in diverse quantum technologies.