Improving Trainability of Variational Quantum Circuits via Regularization Strategies

Regularizing model parameters with prior knowledge of training data and Gaussian noise diffusion can improve the trainability of variational quantum circuits against barren plateaus and saddle points.