Quantum computation models and their application
Year: 2019
Type: dissertation
Abstract: Quantum computation is becoming reality.It is likely that in the upcoming years quantum computations of a complexity beyond the capabilities of classical computers will be realized, albeit significant technological and scientific progress will be required.It is therefore paramount to gain a theoretical understanding of a variety of quantum computation models, as a specific model can be more apt to be realized in a given experimental platform.Moreover, different models of quantum computation may enable the discovery of new quantum algorithms or the improvement of other quantum information processing tasks aside from quantum computation.In this thesis, we present a selection of quantum information processing tasks, specifically: characterization of coherent error channels; quantum metrology; quantum mixing of Markov chains (MCs); algorithms for solving the Quantum Linear Systems Problem (QLSP); and applications of QLSP to MC problems.These tasks are solved within three different quantum computation models.Measurement-Based Quantum Computation (MBQC) is employed for solving the first two of the aforementioned problems, which both consist in the estimation of unknown unitary channels acting on a set of physical qubits that comprise a cluster state or, more generically, a graph state.In the first task, the unitaries represent coherent error channels that act in a different way on each qubit.In the second task, the unitaries act in the same way on each qubit and they encode a parameter that has to be estimated as precisely as possible, as is standard in quantum metrology.In both cases, the MBQC framework is beneficial.The standard gate-based model is used to solve the third problem.Specifically, quantum walks are implemented in this framework as a way to quantize classical MCs.Then, new algorithms based on quantum walks are introduced which, in the context of slowly evolving sequences of MCs, improve upon previously known quantum mixing methods.The Randomization Method, a discretized version of Adiabatic Quantum Computation, is employed to solve the QLSP.For this problem, considering Hamiltonian-based methods lead to almost optimal algorithms which are, moreover, considerably simpler than those developed in the gate-based model, readily affording an experimental demonstration in a Nuclear Magnetic Resonance device.Finally, algorithms for the QLSP are applied to solving MC problems.As a consequence of using linear algebraic algorithms, the output quantum states obtained are different from the outputs of methods based on quantum walks.Specifically, the amplitudes of the output quantum states directly encode probabilities, rather than square roots of probabilities.This may provide exponential improvements in statistical mechanics and machine learning settings.Definition 6 (SZK, Statistical Zero-Knowledge).The complexity class SZK can be defined through the following interaction game, involving a computationally unbounded Prover and a Verifier that can only perform polynomial-time probabilistic (classical) computations.The Verifier initially has as input a bit-string x of length n; then, the Prover and the Verifier can exchange messages, for a polynomial number of rounds.At the end of the computation the Verifier:• will output YES with certainty if x ∈ Y and the Prover follows the specified (honest) protocol (completeness);• will output NO with probability at least 2/3 if x ∈ N , for all possible strategies chosen by the Prover (soundness);• if x ∈ Y , then the Verifier does not gain any knowledge about x (statistically) beyond the fact that x ∈ Y (zero-knowledge).
Author: Orsucci, Davide
Institution: Universität Innsbruck
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Keywords: quantum computation models, measurement-based quantum computation, quantum metrology, quantum walks, Markov chain mixing, randomization methods, adiabatic quantum computation, nuclear magnetic resonance, cluster states, graph states, quantum algorithms, statistical zero-knowledge, quantum information processing, machine learning
Open Access status: green
Funder: Austrian Science Fund
OpenAlex ID: W7005401426
