Federated Learning (FL) has been proposed as an appealing approach to handle data privacy issue of mobile devices compared to conventional machine learning at the remote cloud with raw user data uploading. By leveraging edge servers as intermediaries to perform partial model aggregation in proximity and relieve core network transmission overhead, it enables great potentials in low-latency and energy-efficient FL. Hence we introduce a novel Hierarchical Federated Edge Learning (HFEL) framework in which model aggregation is partially migrated to edge servers from the cloud. We further formulate a joint computation and communication resource allocation and edge association problem for device users under HFEL framework to achieve global cost minimization. To solve the problem, we propose an efficient resource scheduling algorithm in the HFEL framework. It can be decomposed into two subproblems: resource allocation given a scheduled set of devices for each edge server and edge association of device users across all the edge servers. With the optimal policy of the convex resource allocation subproblem for a set of devices under a single edge server, an efficient edge association strategy can be achieved through iterative global cost reduction adjustment process, which is shown to converge to a stable system point. Extensive performance evaluations demonstrate that our HFEL framework outperforms the proposed benchmarks in global cost saving and achieves better training performance compared to conventional federated learning.

BackgroundAngiostrongylus cantonensis (A. cantonensis) infection can result in increased risk of eosinophilic meningitis. Accumulation of eosinophils and inflammation can result in the A. cantonensis infection playing an important role in brain tissue injury during this pathological process. However, underlying mechanisms regarding the transcriptomic responses during brain tissue injury caused by A. cantonensis infection are yet to be elucidated. This study is aimed at identifying some genomic and transcriptomic factors influencing the accumulation of eosinophils and inflammation in the mouse brain infected with A. cantonensis.MethodsAn infected mouse model was prepared based on our laboratory experimental process, and then the mouse brain RNA Libraries were constructed for deep Sequencing with Illumina Genome Analyzer. The raw data was processed with a bioinformatics’ pipeline including Refseq genes expression analysis using cufflinks, annotation and classification of RNAs, lncRNA prediction as well as analysis of co-expression network. The analysis of Refseq data provides the measure of the presence and prevalence of transcripts from known and previously unknown genes.ResultsThis study showed that Cys-Cys (CC) type chemokines such as CCL2, CCL8, CCL1, CCL24, CCL11, CCL7, CCL12 and CCL5 were elevated significantly at the late phase of infection. The up-regulation of CCL2 indicated that the worm of A. cantonensis had migrated into the mouse brain at an early infection phase. CCL2 could be induced in the brain injury during migration and CCL2 might play a major role in the neuropathic pain caused by A. cantonensis infection. The up-regulated expression of IL-4, IL-5, IL-10, and IL-13 showed Th2 cell predominance in immunopathological reactions at late infection phase in response to infection by A. cantonensis. These different cytokines can modulate and inhibit each other and function as a network with the specific potential to drive brain eosinophilic inflammation. The increase of ATF-3 expression at 21 dpi suggested the injury of neuronal cells at late phase of infection. 1217 new potential lncRNA were candidates of interest for further research.ConclusionsThese cytokine networks play an important role in the development of central nervous system inflammation caused by A. cantonensis infection.Electronic supplementary materialThe online version of this article (doi:10.1186/s13071-015-0939-6) contains supplementary material, which is available to authorized users.

No abstract

No abstract

No abstract