Abstract
This paper is an introductory review on the selected topics on quantum machine learning and tensor network, based on educational lectures and resources. Quantum computing has shown its ability to surpass classical computers which consists of conventional digital gates on many tasks. However, the major restriction of this technique lies in limited number of physical qubits available in a near-term device and its overwhelming noise when the size and complexity of a quantum circuit increases. Therefore, a hybrid quantum-classical system is proposed to take full advantage of quantum computing while leaving some of the computational burden to classical computers. Additionally, stemmed from the motivation of using tensor network in machine learning, the matrix product state (MPS) may be applied when designing the structure of model circuit part in the hybrid quantum-classical system to further reduce the number of elements required in quantum computing. A classical supervised learning algorithm based on tensor network is implemented, as well as a normal hybrid quantum-classical system with parameterized quantum circuit to examine their effectiveness independently. This work proposes several ideas on designing tensor network based hybrid quantum-classical model which contains strengths from both techniques to achieve quantum resource-efficient algorithm with promising noise resilience.
Add the full text or supplementary notes for the publication here using Markdown formatting.
