Design and application of small-scale quantum information processors
Date
2022-07-22
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The field of quantum computing is developing rapidly, with extensive research being undertaken in several topics ranging from designing novel computing architectures to developing algorithms for achieving quantum advantage. I address two research directions from this wide spectrum of topics and define my PhD research goals. My first objective is to design high-fidelity controlled-Z (CZ) gates for neutral atoms, and my second objective is to construct a quantum-assisted machine-learning model for solving non-linear regression problems. The potential of neutral-atom quantum computer stems from its unique ability to coherently control several stable qubits with the possibility of strong, long-range interactions between qubits; however the fidelity of a native two-qubit entangling gate on this platform lags behind competing platforms of superconducting systems and trapped ions. We propose gate procedures that rely on simultaneous driving of a pair of Caesium (Cs) atoms using broadband laser pulses and predict high-fidelity CZ gates. Using smooth and globally-optimized adiabatic pulse shapes, our simulations predict fidelities exceeding 0.997 in the presence of spontaneous emission from excited energy levels of Cs. By transitionless quantum driving of each Cs atom, we yield a CZ gate with fidelity 0.9985 over an operation time of 0.12 μs in the presence of spontaneous emission and major technical imperfections. The support vector regression (SVR) is a widely-used classical machine-learning model for regression tasks, including prediction of weather, stock market and real-estate pricing; yet, a currently-feasible quantum SVR model is missing from literature. We formulate quantum-assisted SVR based on quantum annealing, and compare its empirical performance against classical models for the task of detecting facial landmarks. By training the quantum-assisted model using the state-of-the-art quantum annealer, we demonstrate comparable performance of this model and two classical models for the landmark-detection task.Our results on high-fidelity CZ gates show that our gate procedures carry significant potential for achieving scalable quantum computing using atoms. On the other hand, our quantum-assisted SVR acts as a feasible quantum alternative for non-linear regression tasks.
Description
Keywords
quantum gates, neutral atoms, quantum annealing, machine learning
Citation
Dalal, Ar. (2022). Design and application of small-scale quantum information processors (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.