Jupiter:a modern federated learning platform for regional medical care

来源 :中国科学:信息科学(英文版) | 被引量 : 0次 | 上传用户:dongshantongak
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
With the emergence of AI technologies,intrinsic value of data is released and takes tremendous effects on numerous industries.In the context of regional medical care,data sharing and cooperating is in high demand,which can bring both financial and societal benefits.At present,however,medical data are locked inside medical facilities owing to legal risks and economic considerations.How to bring AI technologies into full play under this circumstance is a big challenge.In this paper,we propose Jupiter,an easy-to-use,secure,and high-performance platform for federated machine learning.Jupiter constructs a secure and high-performance aggregator cluster with SGX to efficiently aggregate the encrypted model parameters.Jupiter employs a stateful design to cooperate with medical facilities in regional medical systems with a fixed network connection.By providing an innovative programming abstraction,Jupiter makes model development more friendly to developers.The experiments show that with a low memory footprint,the throughput of a single node on an ordinary PC can reach 300 MB/s(with slice size fixed to 64 KB),and the aggregation primitive we built can process 11k aggregations per second.
其他文献
Neural architecture search (NAS) methods have been proposed to relieve human experts from tedious architecture engineering.However,most current methods are constrained in small-scale search owing to the issue of huge computational resource consumption.Mea
Satellite communication is attracting increasing attention owing to its freedom from geographical constraints.However,its spectrum resources are limited,and it is susceptible to interferences.Therefore,cognitive radio technology can be used to detect and
Conventional feature selection methods select the same feature subset for all classes,which means that the selected features might work better for some classes than the others.Towards this end,this paper proposes a new semi-supervised local feature select
Geophysics offers an important means to investigate the physical processes occurring inside the earth.In particular,since the 1960s,electromagnetic(EM)methods have played important roles in mineral exploration and engineering in-vestigation.Such investiga
In cell-free massive multiple-input multiple-output(MIMO)orthogonal frequency division mul-tiplexing(OFDM)systems,user equipments(UEs)are served by many distributed access points(APs),where channels are correlated owing to finite angle-delay spread in rea
The classical proportional-integral-derivative (PID) controller is ubiquitous in engineering sys-tems that are typically nonlinear with various uncertainties,including random noise.However,most of the literature on PID control focused on linear determinis
Many real-world optimization problems involve multiple conflicting objectives.Such problems are called multiobjective optimization problems(MOPs).Typically,MOPs have a set of so-called Pareto optimal solutions rather than one unique optimal solution.To as
In the rational cryptographic protocol,the two rational players often fall into the prisoner\'s dilemma,which is also the case for the rational secret sharing we consider in this paper.First,it is proved that rational secret sharing has a sequential equ
Full-duplex(FD)two-way relaying has the potential to substantially improve the spectral effi-ciency of the system,and is regarded as an enabling technology for the next generation wireless communication systems.Nevertheless,the joint effect of residual in
The high-accuracy drag-free control is one of the key technologies for gravity gradient satellites.Since the range of the gravity gradiometer is limited,the measurement is subject to the saturation constraint.This paper introduces a design method of the a