No abstract

Federated learning is a recent approach for distributed model training without sharing the raw data of clients. It allows model training using the large amount of user data collected by edge and mobile devices, while preserving data privacy. A challenge in federated learning is that the devices usually have much lower computational power and communication bandwidth than machines in data centers. Training large-sized deep neural networks in such a federated setting can consume a large amount of time and resources. To overcome this challenge, we propose a method that integrates model pruning with federated learning in this paper, which includes initial model pruning at the server, further model pruning as part of the federated learning process, followed by the regular federated learning procedure. Our proposed approach can save the computation, communication, and storage costs compared to standard federated learning approaches. Extensive experiments on real edge devices validate the benefit of our proposed method.

PurposeThe Shanghai Preconception Cohort (SPCC) was initially established to investigate the associations of parental periconceptional nutritional factors with congenital heart disease (CHD) but has further analysed child growth and development and paediatric diseases.ParticipantsPreparing-for-pregnancy couples who presented at Shanghai preconception examination clinics and early-pregnancy women before 14 gestational weeks were enrolled to comprise the periconceptional baseline study population. General characteristics, routine clinical data and consumption of diet supplements, such as folic acid and multivitamins, were collected. Blood samples were obtained at preconception and early, middle and late gestations using standard procedures. Multiple nutritional factors, including folate, homocysteine, vitamin A, vitamin D, vitamin E and metals, in the blood samples of participants selected using a case–control design were examined. Genomic DNA was extracted.Findings to dateThe baseline population included 8045 preconception couples, 3054 single women and 15 615 early-pregnancy women. Data from 12 402 births were collected, and follow-up of the cohort for other outcomes is ongoing. Currently, 151 cases of CHD were identified after birth. The pilot analysis in a small subgroup showed that approximately 20.0% of preconception women and 44.9% of early-pregnancy women had red blood cell (RBC) folate levels that met the international recommendation for preventing neural tube defects.Future plansOnce a sufficient number of CHD cases are achieved, we will investigate the quantitative association of preconception RBC folate levels with CHD using a nested case–control design. The SPCC will be followed up for 18 years to investigate extensive outcomes of growth, development, obesity, and common and rare diseases during childhood and adolescence according to our plan. Blood nutritional factors will be examined in participants selected for specific aims. The SPCC will also allow for prospective cohort studies on extensive research questions.Trial registration numberNCT 02737644

In this paper, we introduce the Theory of Bottleneck Ordering, a mathematical framework that reveals the bottleneck structure of data networks. This theoretical framework provides insights into the inherent topological properties of a network in at least three areas: (1) It identifies the regions of influence of each bottleneck; (2) it reveals the order in which bottlenecks (and flows traversing them) converge to their steady state transmission rates in distributed congestion control algorithms; and (3) it provides key insights into the design of optimized traffic engineering policies. We demonstrate the efficacy of the proposed theory in TCP congestion-controlled networks for two broad classes of algorithms: Congestion-based algorithms (TCP BBR) and loss-based additive-increase/multiplicative-decrease algorithms (TCP Cubic and Reno). Among other results, our network experiments show that: (1) Qualitatively, both classes of congestion control algorithms behave as predicted by the bottleneck structure of the network; (2) flows compete for bandwidth only with other flows operating at the same bottleneck level; (3) BBR flows achieve higher performance and fairness than Cubic and Reno flows due to their ability to operate at the right bottleneck level; (4) the bottleneck structure of a network is continuously changing and its levels can be folded due to variations in the flows' round trip times; and (5) against conventional wisdom, low-hitter flows can have a large impact to the overall performance of a network.

No abstract