Federated Learning (FL) has emerged as a powerful paradigm for decentralized model training, yet it remains vulnerable to deep leakage (DL) attacks that reconstruct private client data from shared model updates. While prior DL methods have demonstrated varying levels of success, they often suffer from instability, limited fidelity, or poor robustness under realistic FL settings. We introduce a new DL attack that integrates a generative Flow Matching (FM) prior into the reconstruction process. By guiding optimization toward the distribution of realistic images (represented by a flow matching foundation model), our method enhances reconstruction fidelity without requiring knowledge of the private data. Extensive experiments on multiple datasets and target models demonstrate that our approach consistently outperforms state-of-the-art attacks across pixel-level, perceptual, and feature-based similarity metrics. Crucially, the method remains effective across different training epochs, larger client batch sizes, and under common defenses such as noise injection, clipping, and sparsification. Our findings call for the development of new defense strategies that explicitly account for adversaries equipped with powerful generative priors.

翻译：联邦学习（Federated Learning, FL）已成为一种强大的去中心化模型训练范式，但其仍然容易受到深度泄露（Deep Leakage, DL）攻击，此类攻击能够从共享的模型更新中重建私有客户端数据。尽管先前的深度泄露方法已展现出不同程度的成功，但它们在实际的联邦学习设置下常常存在不稳定性、保真度有限或鲁棒性差的问题。我们提出了一种新的深度泄露攻击方法，该方法将生成流匹配（Flow Matching, FM）先验整合到重建过程中。通过将优化过程引导至真实图像的分布（由一个流匹配基础模型表示），我们的方法在无需知晓私有数据的情况下，提升了重建保真度。在多个数据集和目标模型上进行的大量实验表明，我们的方法在像素级、感知层面以及基于特征的相似性度量上，均持续优于最先进的攻击方法。至关重要的是，该方法在不同的训练轮次、更大的客户端批次大小下，以及在常见的防御措施（如噪声注入、梯度裁剪和稀疏化）下，均保持有效。我们的研究结果表明，需要开发新的防御策略，以明确应对配备了强大生成先验的对手。