Client-wise data heterogeneity is one of the major issues that hinder effective training in federated learning (FL). Since distribution on each client may vary dramatically, selection strategy can significantly influence convergence rate FL process. Active strategies are popularly proposed recent studies. However, they neglect loss correlations between clients and achieve only marginal improvement compared to uniform strategy. In this work, we propose FedCoran FLframework built a...

Mobile edge computing (MEC) has recently emerged as a promising technology to release the tension between computation-intensive applications and resource-limited mobile terminals (MTs). In this paper, we study delay-optimal computation offloading in computation-constrained MEC systems. We consider task queue at server due its constrained capability. case, MT that are strongly coupled cascade manner, which creates complex interdependences brings new technical challenges. model problem an...

The alkylidene transfer reactions of alkenes are particular significance but challenging. Here, we report that enones can serve as diverse sources for catalyst‐controlled selective C–H alkylation and/or alkylidenation various nucleophiles. Treatment a mixture ketone (or lactam), enone and diarylmethanol with catalytic amount Y[N(TMS)2]3, gave the corresponding α‐C–H bond products derived from to ketones/lactams, whereas reaction C‐nucleophiles in presence KOH catalyst resulted...

Cache-enabled network architecture has great potential for enhancing the efficiency of content distribution as well reducing congestion. This, in turn, called joint optimization traffic engineering and caching strategies while considering both congestion demands. In this paper, we present a distributed framework request/data forwarding dynamic cache placement cache-enabled networks. Specifically, to retrieve information about demands over network, establish dual queue system requests data,...

To alleviate the local computation demands from ever-increasing computation-intensive mobile applications, Mobile Edge Computing (MEC) has proved promising. Especially, by opportunistically offloading these tasks to MEC server, delay of computing could be significantly improved through communication. In this paper, we develop an analytical framework for joint communication and resources allocation multi-user systems. Specifically, retrieve combined effect capabilities, establish a dual queue...

The formation of free radicals using visible light as the energy source is great synthetic value. Despite advances in photoredox catalysis process expensive Ir and Ru complexes or organic dyes, discovery employment photocatalysts based on alternative, abundant, inexpensive metal salts remain highly desired. Herein, visible-light-induced decarboxylation α-imino-oxy acids for generation iminyl has been accomplished nontoxic CeCl3·7H2O photocatalyst via LMCT process. iminyl-radical-mediated C–C...

Personalized recommendation systems are widely used in many Internet services. The sparse embedding lookup models dominates the computational cost of inference due to its intensive irregular memory accesses. Applying resistive random access (ReRAM) based process-in-memory (PIM) architecture accelerate processing can avoid data movements caused by off-chip However, naïve adoption ReRAM-based DNN accelerators leads low computation parallelism and severe under-utilization computing resources,...

Offloading the computationally intensive workloads to edge and cloud not only improves quality of computation, but also creates an extra degree diversity by collecting information from devices in service. Nevertheless, significant concerns on privacy are raised as aggregated could be misused without permission third party. Sparse coding, which has been successful computer vision, is finding application this new domain. In paper, we develop a secured face recognition framework orchestrate...

Offloading the computationally intensive workloads to edge and cloud not only improves quality of computation, but also creates an extra degree diversity by collecting information from devices in service, which, turn, has raised significant concerns on privacy as aggregated could be misused without permission third party. Sparse coding, which been successful computer vision, is finding application this new domain. In paper, we develop a secure face recognition framework orchestrate sparse...

Attention-based neural networks have shown superior performance in a wide range of tasks. Non-volatile processing-in-memory (NVPIM) architecture shows its great potential to accelerate the dense attention model. However, unique unstructured and dynamic sparsity pattern sparse model challenges mapping efficiency NVPIM architecture, as conventional uses vector-matrix-multiplication primitives. In this paper, we propose computation attention. We aim improve for both SDDMM SpMM by introducing...

Cooperative multicast has been demonstrated to achieve significant performance gain over the classic source-destination transmission paradigm by exploiting spatial diversity through participation of multiple relay nodes. As a major technical challenge, selection relays for session impact on performance. The challenge is even more pronounced when number channels limited as in this context coupled with channel allocation. goal paper design fair scheme resources. Specifically, we establish an...

Spectrum aggregation (SA) across heterogeneous channels, including both dedicated and shared provides the potential for improving spectrum utilization fulfilling requirement of broadband services. Heterogeneous SA brings new technical challenges on multisystem coexistence channels resource allocation over channels. In this paper, we develop an analytical framework from a queue stability perspective. To make all systems stable, design algorithm multiple systems. Specifically, derive...

Mobile edge computing (MEC) has recently emerged as a promising technology to release the tension between computation-intensive applications and resource-limited mobile terminals (MTs). In this paper, we construct Markov decision process (MDP) framework optimize delay performance in computation-constrained networks. Different most of existing works, consider computation task queue at MEC server due its constrained capability. case, MT that are mutually strongly coupled cascade manner, which...

Edge and cloud computing has recently emerged not only to meet the ever-increasing computation demands, but also provide extra degree of diversity by collecting data from mobile devices in service. However, this, turn, raised new technical challenges on security issue, calls for design frameworks exploit multi-device diversity. In this paper, we take advantage benefit while preserving privacy. Specifically, 1). To address privacy develop a low-complexity encrypting algorithm based random...

Deep neural networks (DNNs) emerge as a key component in various applications. However, the ever-growing DNN size hinders efficient processing on hardware. To tackle this problem, algorithmic side, compressed models are explored, of which block-circulant memory and hardware-friendly; hardware resistive random-access (ReRAM) based accelerators promising for in-situ DNNs. In work, we design an accelerator named ReBoc accelerating DNNs ReRAM to reap benefits light-weight simultaneously. We...

Emerging resistive random-access memory (ReRAM) has recently been intensively investigated to accelerate the processing of deep neural networks (DNNs). Due in-situ computation capability, analog ReRAM crossbars yield significant throughput improvement and energy reduction compared traditional digital methods. However, power hungry analog-to-digital converters (ADCs) prevent practical deployment ReRAM-based DNN accelerators on end devices with limited chip area budget. We observe that due...

Cooperative caching is a promising technology for enhancing user experience and reducing redundant transmissions through the participation of multiple nodes. In this paper, we design clustering algorithm sectionalized caching, in which each divides its space into two parts contents cached these are determined according to individual interest joint all users same cluster respectively. Different most existing works forming clusters based on similarity files, adopt piecewise as criterion...

Cooperative content caching has been demonstrated to achieve significant performance gain over the conventional paradigm by exploiting diversity through participation of multiple cooperative nodes. Although potential increase efficiency, an improper coalition formation may result in severe degradation. Therefore, nodes should be carefully selected according their interests different objects. In this paper, we develop analytical framework for from a coalitional game perspective. The...

Centralized joint processing has been demonstrated to achieve significant performance gain over the conventional per-cell by avoiding inter-cell interference in cloud radio access networks (C-RANs), but full scale cooperation entire network is infeasible due its huge computational complexity. The goal of this paper design remote head (RRH) clustering and associated power allocation algorithm under resource constraint for C-RANs, which challenging combinatorial problem with non-convex...