[论文解读] Fast-FedUL: A Training-Free Federated Unlearning with Provable Skew Resilience

Thanh Trung Huynh, Trong Bang Nguyen+5 more|arXiv (Cornell University)|May 28, 2024

Brain Tumor Detection and Classification被引用 4

一句话总结

Fast-FedUL 是一种无需训练的联邦遗忘方法，通过解析性地逆转历史梯度，从全局联邦学习（FL）模型中移除目标客户端的影响，实现可证明的偏差鲁棒性。与微调相比，其遗忘时间减少了1000倍，同时将后门攻击成功率降至0.01%，并保持98%的主要任务准确率。

ABSTRACT

Federated learning (FL) has recently emerged as a compelling machine learning paradigm, prioritizing the protection of privacy for training data. The increasing demand to address issues such as ``the right to be forgotten'' and combat data poisoning attacks highlights the importance of techniques, known as extit{unlearning}, which facilitate the removal of specific training data from trained FL models. Despite numerous unlearning methods proposed for centralized learning, they often prove inapplicable to FL due to fundamental differences in the operation of the two learning paradigms. Consequently, unlearning in FL remains in its early stages, presenting several challenges. Many existing unlearning solutions in FL require a costly retraining process, which can be burdensome for clients. Moreover, these methods are primarily validated through experiments, lacking theoretical assurances. In this study, we introduce Fast-FedUL, a tailored unlearning method for FL, which eliminates the need for retraining entirely. Through meticulous analysis of the target client's influence on the global model in each round, we develop an algorithm to systematically remove the impact of the target client from the trained model. In addition to presenting empirical findings, we offer a theoretical analysis delineating the upper bound of our unlearned model and the exact retrained model (the one obtained through retraining using untargeted clients). Experimental results with backdoor attack scenarios indicate that Fast-FedUL effectively removes almost all traces of the target client, while retaining the knowledge of untargeted clients (obtaining a high accuracy of up to 98\% on the main task). Significantly, Fast-FedUL attains the lowest time complexity, providing a speed that is 1000 times faster than retraining. Our source code is publicly available at \url{https://github.com/thanhtrunghuynh93/fastFedUL}.

研究动机与目标

• 解决缺乏高效、无需微调的联邦遗忘方法的问题，实现无需微调即可移除客户端特定数据的影响。
• 克服现有联邦学习遗忘方法的局限性，这些方法依赖于成本高昂的微调，或缺乏理论保证。
• 通过缓解非独立同分布（non-IID）数据设置下梯度反转引入的偏差，确保遗忘后模型的性能保持不变。
• 提供遗忘模型与微调后模型之间差异的理论边界，以确保可靠性。
• 通过仅选择关键历史更新用于遗忘，优化存储与计算开销，降低内存占用。

提出的方法

• 提出一种梯度反转机制，系统性地消除目标客户端历史更新对全局模型的影响，而无需微调。
• 引入一种客户端采样策略，根据更新的重要性优先级选择，以最小化存储与计算成本。
• 构建理论框架，以边界化梯度反转引入的模型偏差，实现准确的性能恢复。
• 采用基于角度的参数偏差分析，量化遗忘模型与微调后模型之间的相似性，确保鲁棒性。
• 采用偏差鲁棒的估计技术，累积多轮的模型偏差，以在遗忘后恢复模型性能。
• 通过感知偏差的梯度校正，隐式应用知识蒸馏原理，避免显式蒸馏步骤。

实验结果

研究问题

• RQ1 是否存在一种联邦遗忘方法，可在无需微调的前提下消除客户端影响，同时保持未污染客户端的模型性能？
• RQ2 在非独立同分布（non-IID）数据设置下，如何理论化边界化并缓解梯度反转引入的模型偏差？
• RQ3 客户端采样策略对遗忘效率与模型准确率有何影响？
• RQ4 与微调相比，所提方法在遗忘效果与后门防御方面表现如何？
• RQ5 无需训练的方法是否能在联邦学习中同时实现高遗忘准确率与低计算成本？

主要发现

• Fast-FedUL 在遗忘模型上的后门攻击成功率仅为0.01%，有效消除了恶意模式。
• 该方法在未污染客户端上保持了高达98%的主要任务准确率，表明知识保留能力出色。
• 与微调相比，Fast-FedUL 将遗忘时间减少了1000倍，显著提升了实际部署效率。
• 理论分析证实，遗忘模型与微调模型之间的差异处于可证明的边界范围内，确保了可靠性。
• 消融实验表明，随机采样与缺乏偏差缓解会使性能下降最高达17%，验证了所提组件的必要性。
• 定性分析表明，Fast-FedUL 的模型参数与微调模型之间的角度偏差在10°以内，且分布集中，表明其保真度高。