Performance Analysis of Quantum Federated Learning in Data Classification

Federated learning is a method with the advantage of allowing various institutions to create a global model by sharing model parameters without sharing the data they possess. Also, with the advent of the era of quantum computing, efforts to combine traditional machine learning algorithms and quantum computing are gaining momentum. In this paper, we intend to discuss quantum federated learning (QFL) methods. We examine the fundamentals of quantum computing and the structure of quantum neural networks, and introduce the methods of quantum federated learning. QFL demonstrated a potential for practical application in the real world, achieving up to 92% performance in the MNIST data classification task.