arXiv subjects

Ju-Yeon Gyhm

Publications and source records attributed to Ju-Yeon Gyhm.

8 recordsLinked to original sources

arXiv2025

Diabatic quantum annealing for training energy-based generative models

Energy-based generative models, such as restricted Boltzmann machines (RBMs), require unbiased Boltzmann samples for effective training. Classical Markov chain Monte Carlo methods, however, converge slowly and yield correlated samples, making large-scale training difficult. We address this bottleneck by applying the analytic relation between annealing schedules and effective inverse temperature in diabatic quantum annealing. By implementing this prescription on a quantum annealer, we obtain temperature-controlled Boltzmann samples that enable RBM training with faster convergence and lower validation error than classical sampling. We further identify a systematic temperature misalignment intrinsic to analog quantum computers and propose an analytical rescaling method that mitigates this hardware noise, thereby enhancing the practicality of quantum annealers as Boltzmann samplers. In our method, the model's connectivity is set directly by the qubit connectivity, transforming the computational complexity inherent in classical sampling into a requirement on quantum hardware. This shift allows the approach to extend naturally from RBMs to fully connected Boltzmann machines, opening opportunities inaccessible to classical training methods.

quant-ph↗

arXiv2025

Gaussian Open Quantum Dynamics and Isomorphism to Superconformal Symmetry

Understanding the Lie algebraic structure of a physical problem often makes it easier to find its solution. In this paper, we focus on the Lie algebra of Gaussian-conserving superoperators. We construct a Lie algebra of $n$-mode states, $\mathfrak{go}(n)$, composed of all superoperators conserving Gaussianity, and we find it isomorphic to $\mathbb{R}^{2n^2+3n}\oplus_{\mathrm{S}}\mathfrak{gl}(2n,\mathbb{R})$. This allows us to solve the quadratic-order Redfield equation for any, even non-Gaussian, state. We find that the algebraic structure of Gaussian operations is the same as that of super-Poincar\'e algebra in three-dimensional spacetime, where the CPTP condition corresponds to the combination of causality and directionality of time flow. Additionally, we find that a bosonic density matrix satisfies both the Klein-Gordon and the Dirac equations. Finally, we expand the algebra of Gaussian superoperators even further by relaxing the CPTP condition. We find that it is isomorphic to a superconformal algebra, which represents the maximal symmetry of the field theory. This suggests a deeper connection between two seemingly unrelated fields, with the potential to transform problems from one domain into another where they may be more easily solved.

quant-ph↗

arXiv2025

Fundamental Scaling Limit in Critical Quantum Metrology

Critical quantum metrology aims to harness critical properties near quantum phase transitions to enhance parameter estimation precision. However, critical slowing down inherently limits the achievable precision within a finite evolution time. To address this challenge, we establish a fundamental scaling limit of critical quantum metrology with respect to the total evolution time. We find that the winding number of the system's phase space trajectory determines the scaling bound of quantum Fisher information. Furthermore, we demonstrate that the exponential scaling of the quantum Fisher information can be obtained, and for this, it is necessary to increase the winding number by the total evolution time. We explicitly construct a time-dependent control to achieve optimal scaling from a simple on-off scheme depending on the system's phase and discuss its topological nature. We highlight that such an exponential scaling of quantum Fisher information remains valid even without reaching the critical point and in the presence of thermal dissipation, albeit with a decreased exponent.

quant-ph↗

arXiv2024

Boltzmann Sampling by Diabatic Quantum Annealing

Boltzmann sampling is a central component of many computational frameworks, including numerous algorithms in machine learning. Although quantum annealers have been investigated as potential fast Boltzmann samplers, their dependence on environmental noise makes precise control of the effective temperature difficult, introducing uncertainty into the sampling process. As an alternative, we propose diabatic quantum annealing -- a faster, purely unitary process -- as a controllable Boltzmann sampler in which the effective temperature is determined by the annealing rate. Using the ferromagnetic Ising model and the Sherrington--Kirkpatrick model as test cases, we demonstrate that this method achieves rapid and accurate sampling in the high-temperature regime.

cond-mat.stat-mech↗

arXiv2024

Dynamics of a small quantum system open to a bath with thermostat

We investigate dynamics of a small quantum system open to a bath with thermostat. We introduce another bath, called super bath, weakly coupled with the bath to provide it with thermostat, which has either the Lindblad or Redfield type. We treat the interaction between the system and bath via a rigorous perturbation theory. Due to the thermostat, the bath behaves dissipative and stochastic, for which the usual Born-Markov assumption is not needed. We consider a specific example of a harmonic oscillator system, and a photonic bath in a large container, and a super bath of the Caldeira-Legget oscillators distributed on the inner surface of the container. We use the $P$-representation for the total harmonic system. We derive the reduced time-evolution equation for the system by explicitly finding the correlation between the system and bath beyond the product state, that was not obtainable in the previous theory for the system and bath isolated from environment, and marginalizing bath degrees of freedom. Remarkably, the associated dynamic equation for the system density matrix is of the same form as the Redfield master equation with different coefficients depending on thermostat used. We find steady state does not depend on thermostat, but time-dependent state does, that agrees with common expectation. We expect to apply our theory to general systems. Unlike the usual quantum master equations, our reduced dynamics allows investigation for time-dependent protocols and non-equilibrium quantum stochastic dynamics will be investigated in future.

quant-ph↗

arXiv2024

Minimal time required to charge a quantum system

We introduce a quantum charging distance as the minimal time that it takes to reach one state (charged state) from another state (depleted state) via a unitary evolution, assuming limits on the resources invested into the driving Hamiltonian. For pure states it is equal to the Bures angle, while for mixed states, its computation leads to an optimization problem. Thus, we also derive easily computable bounds on this quantity. The charging distance tightens the known bound on the mean charging power of a quantum battery, it quantifies the quantum charging advantage, and it leads to an always achievable quantum speed limit. In contrast with other similar quantities, the charging distance does not depend on the eigenvalues of the density matrix, it depends only on the corresponding eigenspaces. This research formalizes and interprets quantum charging in a geometric way, and provides a measurable quantity that one can optimize for to maximize the speed of charging of future quantum batteries.

quant-ph↗

arXiv2024

Beneficial and detrimental entanglement for quantum battery charging

We establish a general implementation-independent approach to assess the potential advantage of using highly entangled quantum states between the initial and final states of the charging protocol to enhance the maximum charging power of quantum batteries. It is shown that the impact of entanglement on power can be separated from both the global quantum speed limit associated to an optimal choice of driving Hamiltonian and the energy gap of the batteries. We then demonstrate that the quantum state advantage of battery charging, defined as the power obtainable for given quantum speed limit and battery energy gap, is not an entanglement monotone. A striking example we provide is that, counterintuitively, independent thermalization of the local batteries, completely destroying any entanglement, can lead to larger charging power than that of the initial maximally entangled state. Highly entangled states can thus also be potentially disadvantageous when compared to product states. We also demonstrate that taking the considerable effort of producing highly entangled states, such as W or $k$-locally entangled states, is not sufficient to obtain quantum-enhanced scaling behavior with the number of battery cells. Finally, we perform an explicit computation for a Sachdev-Ye-Kitaev battery charger to demonstrate that the quantum state advantage allows the instantaneous power to exceed its classical bound.

quant-ph↗

arXiv2022

Quantum Charging Advantage Cannot Be Extensive Without Global Operations

Quantum batteries are devices made from quantum states, which store and release energy in a fast and efficient manner, thus offering numerous possibilities in future technological applications. They offer a significant charging speedup when compared to classical batteries, due to the possibility of using entangling charging operations. We show that the maximal speedup that can be achieved is extensive in the number of cells, thus offering at most quadratic scaling in the charging power over the classically achievable linear scaling. To reach such a scaling, a global charging protocol, charging all the cells collectively, needs to be employed. This concludes the quest on the limits of charging power of quantum batteries and adds to other results in which quantum methods are known to provide at most quadratic scaling over their classical counterparts.

quant-ph↗