A Comparative Study of Classical and Quantum Neural Networks
Year: 2023
Type: article
Abstract: In recent years, classical neural networks have been widely used in various applications and have achieved remarkable success. However, with the advent of quantum computing, there is a growing interest in quantum neural networks (QNNs) as a potential alternative to classical machine learning. In this thesis, we study the architectures of quantum and classical neural networks. We also investigate the performance of QNNs compared to classical neural networks from various aspects, such as vanishing gradient, trainability, expressivity. Our experiments demonstrate that QNNs have the potential to outperform classical neural networks in specific scenarios. While more powerful QNNs exhibit improved performance compared to classical neural networks, our findings also indicate that less powerful QNNs may not always yield significant improvements. This suggests that the effectiveness of QNNs in surpassing classical approaches is contingent on factors such as network architecture, optimization techniques, problem complexity.
Source: Työväentutkimus Vuosikirja
Author: Zhao, Chenhui
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7199577987