arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.09247cs.LG
收藏

异步私有联邦学习中的动量近似

Momentum Approximation in Asynchronous Private Federated Learning

Amazon(亚马逊);Apple(苹果公司)

Tao Yu, Congzheng Song, Jianyu Wang, Mona Chitnis

AI总结:

针对异步联邦学习中直接套用动量会导致收敛慢、性能降的问题,提出通过历史模型更新最优加权平均最小化隐式偏差的动量近似方法,兼容隐私技术且开销低,可大幅提升收敛速度。

摘要:

异步协议已被证明可提升拥有海量客户端的联邦学习(federated learning, FL)的可扩展性。同时,基于动量的方法能在同步FL中实现最优的模型质量。然而,在异步FL算法中直接套用动量会导致收敛速度变慢、模型性能下降,目前尚不清楚如何有效结合这两种技术以实现双赢。本文发现异步性会给动量更新引入隐式偏差。为解决该问题,我们提出动量近似方法,通过求取所有历史模型更新的最优加权平均来最小化偏差。动量近似与安全聚合、差分隐私均兼容,且仅需少量通信与存储开销即可轻松集成到生产级FL系统中。我们在基准FL数据集上通过实验证明,与直接结合异步FL和动量的方案相比,动量近似可实现1.15–4倍的收敛速度提升。

英文摘要:

Asynchronous protocols have been shown to improve the scalability of federated learning (FL) with a massive number of clients. Meanwhile, momentum-based methods can achieve the best model quality in synchronous FL. However, naively applying momentum in asynchronous FL algorithms leads to slower convergence and degraded model performance. It is still unclear how to effective combinie these two techniques together to achieve a win-win. In this paper, we find that asynchrony introduces implicit bias to momentum updates. In order to address this problem, we propose momentum approximation that minimizes the bias by finding an optimal weighted average of all historical model updates. Momentum approximation is compatible with secure aggregation as well as differential privacy, and can be easily integrated in production FL systems with a minor communication and storage cost. We empirically demonstrate that on benchmark FL datasets, momentum approximation can achieve $1.15 \textrm{--}4\times$ speed up in convergence compared to naively combining asynchronous FL with momentum.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑