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 2110.03469cs.LGcs.AIcs.DC
收藏

从小数据集进行联邦学习

Federated Learning from Small Datasets

Institute for AI in medicine (IKIM)(医学人工智能研究所(IKIM));University Hospital Essen(埃森大学医院);Ruhr-University Bochum(波鸿鲁尔大学);Monash University(莫纳什大学);Harvard T.H. Chan School of Public Health(哈佛大学陈曾熙公共卫生学院);CISPA Helmholtz Center for Information Security(CISPA亥姆霍兹信息安全中心)

Michael Kamp, Jonas Fischer, Jilles Vreeken

AI总结:

针对联邦学习中本地数据过小导致训练失败的问题,提出将模型聚合与本地模型排列交织的新方法,通过菊花链式数据暴露提升数据稀疏领域的训练效率,保留隐私与效率优势。

摘要:

联邦学习允许多方在不共享本地数据的情况下协作训练联合模型。这使得机器学习能够应用于本质上分布式、不可公开的数据场景,例如医疗领域。在实践中,联合训练通常通过聚合本地模型来实现,为此,本地训练目标在期望上应与联合(全局)目标相似。然而,本地数据集往往很小,导致本地目标与全局目标差异很大,从而使联邦学习失败。我们提出了一种新颖的方法,将模型聚合与本地模型的排列交织在一起。排列使每个本地模型暴露于本地数据集的菊花链中,从而在数据稀疏领域实现更高效的训练。这使得能够在极小的本地数据集(如跨医院的病人数据)上进行训练,同时保留联邦学习的训练效率和隐私优势。

英文摘要:

Federated learning allows multiple parties to collaboratively train a joint model without sharing local data. This enables applications of machine learning in settings of inherently distributed, undisclosable data such as in the medical domain. In practice, joint training is usually achieved by aggregating local models, for which local training objectives have to be in expectation similar to the joint (global) objective. Often, however, local datasets are so small that local objectives differ greatly from the global objective, resulting in federated learning to fail. We propose a novel approach that intertwines model aggregations with permutations of local models. The permutations expose each local model to a daisy chain of local datasets resulting in more efficient training in data-sparse domains. This enables training on extremely small local datasets, such as patient data across hospitals, while retaining the training efficiency and privacy benefits of federated learning.

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。

↑