On Assessing the Quantum Advantage for MaxCut Provided by Quantum Neural Network Ansätze

Abstract

In this thesis we expand upon the results that led to the paper of Lee et al., arXiv:2105.01114 (2021). In particular, we give more details on the oracular formulation of variational quantum algorithms, and the relationship between properties of Ansätze and the strength of their corresponding oracles. Furthermore, having identified the importance of noncommutativity in parameterized quantum circuits (PQCs) as likely being crucial to achieving a quantum advantage, we compare this notion to similar properties in classical neural networks such as nonlinearity, based on the perspective of the recent moniker for PQCs as quantum neural networks. While this thesis includes much of the figures and content from the aforementioned paper, it should be considered mainly as a self-contained collection of supplementary materials.

arXiv comment61 pages, expansion of Lee et al., arXiv:2105.01114 (2021)

References in corpus (18)

• A variational eigenvalue solver on a quantum processor
• Elementary gates for quantum computation
• Hardware-efficient Variational Quantum Eigensolver for Small Molecules and Quantum Magnets
• Ising formulations of many NP problems
• Barren plateaus in quantum neural network training landscapes
• The theory of variational hybrid quantum-classical algorithms
• A Quantum Approximate Optimization Algorithm
• An adaptive variational algorithm for exact molecular simulations on a quantum computer
• From the Quantum Approximate Optimization Algorithm to a Quantum Alternating Operator Ansatz
• Quantum Approximate Optimization of the Long-Range Ising Model with a Trapped-Ion Quantum Simulator
• Unsupervised Machine Learning on a Hybrid Quantum Computer
• Learning and Generalization in Overparameterized Neural Networks, Going Beyond Two Layers
• Capacity and quantum geometry of parametrized quantum circuits
• Improved success probability with greater circuit depth for the quantum approximate optimization algorithm
• First-order Methods Almost Always Avoid Saddle Points
• The Power of Normalization: Faster Evasion of Saddle Points
• Mutual information-assisted Adaptive Variational Quantum Eigensolver
• How Much Over-parameterization Is Sufficient to Learn Deep ReLU Networks?