Elements of Quantum Neuromorphic Learning and Information Processing on Semiconductor Quantum Dot Arrays
Nikolaos Petropoulos
This dissertation presents a comprehensive study of semiconductor quantum computing, offering novel insights into quantum sensing, information processing and neuromorphic learning within the realm of semiconductor quantum dot arrays. Beginning with an introduction to the foundational mathematical principles of quantum information theory such as qubits, Hilbert spaces and density matrices, it sets the stage for the in-depth analyses that follow. The research progresses by investigating the design and characterization of nanoscale single-electron box (SEB) utilizing a floating lead for enhanced quantum sensing capabilities. Through the adaptation of a multi-orbital Anderson impurity model, the study provides a theoretical framework for understanding the SEB’s behavior, identifying limitations and suggesting future improvements. Further exploration into the dynamical multipartite entanglement formation highlights the critical role of quantum information encoding and the impact of metallic leads. Employing advanced quantum models, the research investigates the interplay between qubit encoding strategies and multipartite entanglement, revealing essential trade-offs and the influence of carrier numbers on quantum information scrambling. The dissertation also investigates the detection of topological order and qubit encoding within Su-Schrieffer-Heeger type quantum dot arrays. By proposing models with odd and even parity and examining topological edge states, the research explores novel approaches to quantum information encoding based on topological invariants. This investigation sheds light on the potential for leveraging quantum correlations and topological properties for quantum computing. Transitioning to quantum neuromorphic learning, the study introduces a novel application of quantum reservoir computing for both quantum and classical optimization tasks. By exploiting the dynamics of higher-dimensional Hilbert spaces, the research demonstrates the potential of quantum reservoir computing to enhance learning performance and address limitations of current quantum computing paradigms. Concluding, the dissertation highlights the significant potential of semiconductor quantum computing for a wide array of applications, from quantum sensing to neuromorphic learning. It shows the promising alternatives provided by emerging quantum computing methodologies, particularly in the Noisy-Intermediate Scale Quantum (NISQ) era, for solving complex quantum and classical problems. Through a series of comprehensive analyses and innovative models, this work contributes not only to the advancement of semiconductor quantum computing technologies, but more broadly in the field of quantum computation and opens new pathways for future research in quantum information processing and quantum-based learning systems.
•Jan 1, 2024•0 citations