A computational method called the local-field response method is proposed, where spins evolve by responding to an effective field consisting of gradually decreasing external fields and spin–spin interactions, similarly to what is carried out in adiabatic quantum computing (AQC). This method is partly quantum–mechanical. That is, spins are treated as classical variables, but the response function of the spins to the effective field is determined a priori by referring to a quantum–mechanical calculation that was carried out for similar problems. This novel response function improves the performance of the ground state being maintained in the time evolution compared with the case without a priori information. The performance is numerically checked in an eight-qubit system by solving random-interaction problems of finding their ground states. The false probability decreases by about half as a result of using a priori information. The operation of this method is classical, but it has a quantum–mechanical advantage through a priori information. This method is practically useful because obtaining a complete quantum system is difficult as it stands.
A computational method called the local-field response method is proposed, where spins evolve by responding to an effective field consisting of gradually decreasing external fields and spin–spin interactions, similarly to what is carried out in adiabatic quantum computing (AQC). This method is partly quantum–mechanical. That is, spins are treated as classical variables, but the response function of the spins to the effective field is determined a priori by referring to a quantum–mechanical calculation that was carried out for similar problems. This novel response function improves the performance of the ground state being maintained in the time evolution compared with the case without a priori information. The performance is numerically checked in an eight-qubit system by solving random-interaction problems of finding their ground states. The false probability decreases by about half as a result of using a priori information. The operation of this method is classical, but it has a quantum–mechanical advantage through a priori information. This method is practically useful because obtaining a complete quantum system is difficult as it stands.
The local-field response (LFR) method is a way of searching for the ground state of an Ising spin Hamiltonian. Spins evolve by responding to an effective field consisting of gradually decreasing external transverse fields and spin–spin interactions. The spins are treated as classical variables, but quantum effects are taken into account in the response of the spins. The conventional LFR method takes a quantum effect into account through an a priori determined response function. This report not only uses the response function but also considers quantum effects more directly. (1) The effect of band anticrossing is considered. (2) The interaction term and local term in the effective field are appropriately balanced by renormalizing the local term. (3) The a priori determined response function is modified on the basis of a self-field correction. In addition to these quantum-mechanical improvements, (4) the transverse field is modulated with a saw shape to increase opportunities of finding the global minimum, and all spins are inverted several times in the middle of the process in order to expand the search space. As a result, the false probability, the inaccuracy of the solution, and the average number of steps before reaching the final solution decrease to about one-tenth of the corresponding values for the case without improvements.
In the standard approach to adiabatic quantum computing (AQC), quantum information stored on qubits are adia- batically evolved to find the lowest energy state of a problem Hamiltonian. Here we investigate a variation of AQC where spin ensembles are used in place of qubits. The use of ensembles duplicates the quantum information, and allows errors to be suppressed during the adiabatic evolution. We show that there are two distinct types of problem Hamiltonians under this mapping, where the first excited state is a single particle perturbation on the ground state (Type I); or a macroscopically distinct fully polarized state (Type II). For Type I problems, we find that the minimum gap for large ensembles is well predicted by mean-field theory and the AQC performance improves with ensemble size, realizing error-suppression. For Type II problems, the performance of the AQC is mixed, and the gap can increase or decrease with ensemble size depending on the problem instance. The incidences of the Type II problems are greatly suppressed with ensemble size, and consequently for randomly chosen problem instances the success probability of the scheme improves with the ensemble size. Our approach shows that it is possible to perform AQC without the necessity of controlling individual qubits, which allows for an alternative route towards implementing AQC.
The local field response (LFR) method is a method of searching for the ground state in a similar manner to quantum annealing. However, it operates on a classical machine, and quantum effects are phenomenologically introduced: (1) an average quantum effect is incorporated in the response function of the spins to the effective field through a priori information; (2) the effect of linear superposition is introduced by taking two successive time points into account in the spin response, reproducing 〈σz〉 ≈ 0 at band anticrossing points; (3) spin correlations, as an effect of entanglement, are introduced by resequencing the calculation order of the spins at their response to the effective field. In this report, we review these phenomenological methods and their effects.
The local field response (LFR) method is a way of searching for the ground state in a similar manner to quantum annealing. However, the LFR method operates on a classical machine, and quantum effects are introduced through a priori information and through phenomenological means reflecting the states during the computations. The LFR method has been treated with a one-body approximation, and therefore, the effect of entanglement has not been sufficiently taken into account. In this report, spin correlations are phenomenologically introduced as one of the effects of entanglement, by which multiple tunneling at anticrossing points is taken into account. As a result, the accuracy of solutions for a 128-bit system increases by 31% compared with that without spin correlations.
Quantum computing promises to efficiently and accurately solve many important problems in quantum chemistry which elude classical solvers, such as the electronic structure problem of highly correlated materials. Two leading methods in solving the ground state problem are the Variational Quantum Eigensolver (VQE) and Adiabatic Quantum Computing (AQC) algorithms. VQE often struggles with convergence due to the energy landscape being highly non-convex and the existence of barren plateaux, and implementing AQC is beyond the capabilities of current quantum devices as it requires deep circuits. Adiabatically-inspired algorithms aim to fill this gap. In this paper, we first present a unifying framework for these algorithms and then benchmark the following methods: the Adiabatically Assisted VQE (AAVQE) (Garcia-Saez and Latorre (2018)), the Variational Adiabatic Quantum Computing (VAQC) (Harwood et al (2022)), and the Adiabatic Quantum Computing with Parametrized Quantum Circuits (AQC-PQC) (Kolotouros et al (2025)) algorithms. Second, we introduce a novel hybrid approach termed G-AQC-PQC, which generalizes the AQC-PQC method, and combines adiabatic-inspired initialization with the low-memory BFGS optimizer, reducing the quantum computational cost of the method. Third, we compare the accuracy of the methods for chemistry applications using the beryllium hydride molecule (BeH2). We compare the approaches across a number of different choices (ansätze types, depth, discretization steps, initial Hamiltonian, adiabatic schedules and method used). Our results show that the G-AQC-PQC outperforms conventional VQE. We further discuss limitations such as the zero-gradient problem and identify regimes where adiabatically-inspired methods offer a tangible advantage for near-term quantum chemistry applications.
In this thesis, I investigate aspects of local Hamiltonians in quantum computing. First, I focus on the Adiabatic Quantum Computing model, based on evolution with a time dependent Hamiltonian. I show that to succeed using AQC, the Hamiltonian involved must have local structure, which leads to a result about eigenvalue gaps from information theory. I also improve results about simulating quantum circuits with AQC. Second, I look at classically simulating time evolution with local Hamiltonians and finding their ground state properties. I give a numerical method for finding the ground state of translationally invariant Hamiltonians on an infinite tree. This method is based on imaginary time evolution within the Matrix Product State ansatz, and uses a new method for bringing the state back to the ansatz after each imaginary time step. I then use it to investigate the phase transition in the transverse field Ising model on the Bethe lattice. Third, I focus on locally constrained quantum problems Local Hamiltonian and Quantum Satisfiability and prove several new results about their complexity. Finally, I define a Hamiltonian Quantum Cellular Automaton, a continuous-time model of computation which doesn't require control during the computation process, only preparation of product initial states. I construct two of these, showing that time evolution with a simple, local, translationally invariant and time-independent Hamiltonian can be used to simulate quantum circuits.