To improve the quality of Differentially private (DP) synthetic images, most studies have focused on improving the core optimization techniques (e.g., DP-SGD). Recently, we have witnessed a paradigm shift that takes these techniques off the shelf and studies how to use them together to achieve the best results. One notable work is DP-FETA, which proposes using `central images' for `warming up' the DP training and then using traditional DP-SGD. Inspired by DP-FETA, we are curious whether there are other such tools we can use together with DP-SGD. We first observe that using `central images' mainly works for datasets where there are many samples that look similar. To handle scenarios where images could vary significantly, we propose FETA-Pro, which introduces frequency features as `training shortcuts.' The complexity of frequency features lies between that of spatial features (captured by `central images') and full images, allowing for a finer-grained curriculum for DP training. To incorporate these two types of shortcuts together, one challenge is to handle the training discrepancy between spatial and frequency features. To address it, we leverage the pipeline generation property of generative models (instead of having one model trained with multiple features/objectives, we can have multiple models working on different features, then feed the generated results from one model into another) and use a more flexible design. Specifically, FETA-Pro introduces an auxiliary generator to produce images aligned with noisy frequency features. Then, another model is trained with these images, together with spatial features and DP-SGD. Evaluated across five sensitive image datasets, FETA-Pro shows an average of 25.7% higher fidelity and 4.1% greater utility than the best-performing baseline, under a privacy budget $ε= 1$.

翻译：为提升差分隐私（DP）合成图像的质量，多数研究聚焦于改进核心优化技术（如DP-SGD）。近期，我们见证了一种范式转变，即直接采用现有技术并研究如何协同使用它们以获得最佳效果。一项重要工作是DP-FETA，其提出使用“中心图像”对DP训练进行“预热”，随后再使用传统DP-SGD。受DP-FETA启发，我们探索是否存在其他此类工具可与DP-SGD协同使用。我们首先观察到，使用“中心图像”主要适用于包含大量相似样本的数据集。为处理图像差异显著的场景，本文提出FETA-Pro，其引入频率特征作为“训练捷径”。频率特征的复杂度介于空间特征（由“中心图像”捕获）与完整图像之间，从而可为DP训练提供更细粒度的课程学习。为将这两类捷径结合使用，一个挑战在于处理空间特征与频率特征之间的训练差异。为解决此问题，我们利用生成模型的流水线生成特性（即并非让单一模型使用多种特征/目标进行训练，而是让多个模型分别处理不同特征，再将一个模型的生成结果输入另一模型），并采用更灵活的设计。具体而言，FETA-Pro引入一个辅助生成器来产生与带噪频率特征对齐的图像。随后，另一模型使用这些图像，结合空间特征与DP-SGD进行训练。在五个敏感图像数据集上的评估表明，在隐私预算$ε= 1$下，FETA-Pro相比性能最佳的基线方法，平均保真度提升25.7%，效用提高4.1%。