🤖 AI Summary
To address the prevalent barren plateau (gradient vanishing) and saddle-point stagnation issues in training variational quantum circuits (VQCs) on noisy intermediate-scale quantum (NISQ) devices, this paper proposes a parameter regularization method that jointly leverages data-driven priors and Gaussian noise diffusion. It is the first to synergistically integrate Bayesian prior modeling with controllable noise injection into VQC optimization, regularizing parameter update trajectories to significantly enhance gradient signal strength and improve parameter-space trainability. Experiments across four benchmark quantum machine learning datasets demonstrate that the proposed strategy accelerates convergence by an average factor of 2.1×, boosts final classification accuracy by up to +8.7%, and robustly mitigates barren plateaus. The core contribution is the development of the first data-noise joint regularization framework explicitly designed to enhance the training robustness of VQCs.
📝 Abstract
In the era of noisy intermediate-scale quantum (NISQ), variational quantum circuits (VQCs) have been widely applied in various domains, advancing the superiority of quantum circuits against classic models. Similar to classic models, regular VQCs can be optimized by various gradient-based methods. However, the optimization may be initially trapped in barren plateaus or eventually entangled in saddle points during training. These gradient issues can significantly undermine the trainability of VQC. In this work, we propose a strategy that regularizes model parameters with prior knowledge of the train data and Gaussian noise diffusion. We conduct ablation studies to verify the effectiveness of our strategy across four public datasets and demonstrate that our method can improve the trainability of VQCs against the above-mentioned gradient issues.