亚洲男人的天堂2018av,欧美草比,久久久久久免费视频精选,国色天香在线看免费,久久久久亚洲av成人片仓井空

Training Deep Neural Networks (DNNs) is a widely popular workload in both enterprises and cloud data centers. Existing schedulers for DNN training consider GPU as the dominant resource, and allocate other resources such as CPU and memory proportional to the number of GPUs requested by the job. Unfortunately, these schedulers do not consider the impact of a job's sensitivity to allocation of CPU, memory, and storage resources. In this work, we propose Synergy, a resource-sensitive scheduler for shared GPU clusters. Synergy infers the sensitivity of DNNs to different resources using optimistic profiling; some jobs might benefit from more than the GPU-proportional allocation and some jobs might not be affected by less than GPU-proportional allocation. Synergy performs such multi-resource workload-aware assignments across a set of jobs scheduled on shared multi-tenant clusters using a new near-optimal online algorithm. Our experiments show that workload-aware CPU and memory allocations can improve average JCT up to 3.4x when compared to traditional GPU-proportional scheduling.

We consider a practical cell-free massive multiple-input-multiple-output (MIMO) system with multi-antenna access points (APs) and spatially correlated Rician fading channels. The significant phase-shift of the line-of-sight component induced by the user equipment movement is modeled randomly. Furthermore, we investigate the uplink spectral efficiency (SE) with maximum ratio (MR)/local minimum mean squared error (L-MMSE) combining and optimal large-scale fading decoding based on the phase-aware MMSE, phase-aware element-wise MMSE and linear MMSE (LMMSE) estimators. Then new closed-form SE expressions with MR combining are derived. Numerical results validate our derived expressions and show that the SE benefits from the spatial correlation. It is important to observe that the performance gap between L-MMSE and MR combining increases with the number of antennas per AP and the SE of the LMMSE estimator is lower than that of other estimators due to the lack of phase-shifts knowledge.

In light of the premises of beyond fifth generation (B5G) networks, the need for better exploiting the capabilities of cloud-enabled networks arises, so as to cope with the large-scale interference resulting from the massive increase of data-hungry systems. A compound of several clouds, jointly managing inter-cloud and intra-cloud interference, constitutes a practical solution to account for the requirements of B5G networks. This paper considers a multi-cloud radio access network model (MC-RAN), where each cloud is connected to a distinct set of base stations (BSs) via limited capacity fronthaul links. The BSs are equipped with local cache storage and baseband processing capabilities, as a means to alleviate the fronthaul congestion problem. The paper then investigates the problem of jointly assigning users to clouds and determining their beamforming vectors so as to maximize the network-wide energy efficiency (EE) subject to fronthaul capacity and transmit power constraints. This paper solves such a mixed discrete-continuous, non-convex optimization problem using fractional programming (FP) and successive inner-convex approximation (SICA) techniques to deal with the non-convexity of the continuous part of the problem, and $l_0$-norm approximation to account for the binary association part. A highlight of the proposed algorithm is its capability of being implemented in a distributed fashion across the network's multiple clouds through a reasonable amount of information exchange. The numerical simulations illustrate the pronounced role the proposed algorithm plays in alleviating the interference of large-scale MC-RANs, especially in dense networks.

In this paper, we introduce a Deep Neural Network (DNN) to maximize the Proportional Fairness (PF) of the Spectral Efficiency (SE) of uplinks in Cell-Free (CF) massive Multiple-Input Multiple-Output (MIMO) systems. The problem of maximizing the PF of the SE is a non-convex optimization problem in the design variables. We will develop a DNN which takes pilot sequences and large-scale fading coefficients of the users as inputs and produces the outputs of optimal transmit powers. By consisting of densely residual connections between layers, the proposed DNN can efficiently exploit the hierarchical features of the input and motivates the feed-forward nature of DNN architecture. Experimental results showed that, compared to the conventional iterative optimization algorithm, the proposed DNN has excessively lower computational complexity with the trade-off of approximately only 1% loss in the sum rate and the fairness performance. This demonstrated that our proposed DNN is reasonably suitable for real-time signal processing in CF massive MIMO systems.

In this paper, we investigate a cell-free massive MIMO system with both access points (APs) and user equipments (UEs) equipped with multiple antennas over jointly correlated Rayleigh fading channels. We study four uplink implementations, from fully centralized processing to fully distributed processing, and derive achievable spectral efficiency (SE) expressions with minimum mean-squared error successive interference cancellation (MMSE-SIC) detectors and arbitrary combining schemes. Furthermore, the global and local MMSE combining schemes are derived based on full and local channel state information (CSI) obtained under pilot contamination, which can maximize the achievable SE for the fully centralized and distributed implementation, respectively. We study a two-layer decoding implementation with an arbitrary combining scheme in the first layer and optimal large-scale fading decoding in the second layer. Besides, we compute novel closed-form SE expressions for the two-layer decoding implementation with maximum ratio combining. We compare the SE of different implementation levels and combining schemes and investigate the effect of having additional UE antennas. Note that increasing the number of antennas per UE may degrade the SE performance and the optimal number of UE antennas maximizing the SE is related to the implementation levels, the length of the resource block, and the number of UEs.

This paper studies methods to achieve a high and flexible coverage performance of a large-scale cellular network. The network shall enable unmanned aerial vehicles (UAVs) for non-orthogonal multiple access (NOMA) transmission enhancement so as to simultaneously serve two users. The considered scenario consists of a network with a tier of base stations and UAVs. Each UAV is mounted with an intelligent reflecting surface (IRS) in order to serve as an aerial IRS reflecting signals between a base station and a user in the network. All the UAVs in the network are deployed based on a newly proposed three-dimensional (3D) point process leading to a tractable yet accurate analysis of the association statistics, which is otherwise difficult to analyze due to the mobility of UAVs. In light of this, we also analyze the downlink coverage of UAV-IRS-assisted NOMA transmission for two users and derive the corresponding coverage probabilities. Our coverage analyses shed light on the optimal allocations of transmit power between NOMA users and UAVs to accomplish the goal of ubiquitous and flexible NOMA transmission. We also conduct numerical simulations to validate our coverage analysis results while demonstrating the improved coverage performance achieved by aerial IRSs.

Node clustering is a powerful tool in the analysis of networks. We introduce a graph neural network framework to obtain node embeddings for directed networks in a self-supervised manner, including a novel probabilistic imbalance loss, which can be used for network clustering. Here, we propose directed flow imbalance measures, which are tightly related to directionality, to reveal clusters in the network even when there is no density difference between clusters. In contrast to standard approaches in the literature, in this paper, directionality is not treated as a nuisance, but rather contains the main signal. DIGRAC optimizes directed flow imbalance for clustering without requiring label supervision, unlike existing GNN methods, and can naturally incorporate node features, unlike existing spectral methods. Experimental results on synthetic data, in the form of directed stochastic block models, and real-world data at different scales, demonstrate that our method, based on flow imbalance, attains state-of-the-art results on directed graph clustering, for a wide range of noise and sparsity levels and graph structures and topologies.

In the Minimum Common String Partition Problem (MCSP), we are given two strings on input, and we want to partition both into the same collection of substrings, minimizing the number of the substrings in the partition. This combinatorial optimization problem has applications in computational biology and is NP-hard. Many different heuristic and exact methods exist for this problem, such as a Greedy approach, Ant Colony Optimization, or Integer Linear Programming. In this paper, we formulate the MCSP as a Dynamic Program and develop an exact solution algorithm based on Decision Diagrams for it. We also introduce a restricted Decision Diagram that allows to compute heuristic solutions to the MCSP and compare the quality of solution and runtime on instances from literature with existing approaches. Our approach scales well and is suitable for heuristic solution of large-scale instances.

Sparse code multiple access (SCMA) is a promising multiuser communication technique for the enabling of future massive machine-type networks. Unlike existing codebook design schemes assuming uniform power allocation, we present a novel class of SCMA codebooks which display power imbalance among different users for downlink transmission. Based on the Star-QAM mother constellation structure and with the aid of genetic algorithm, we optimize the minimum Euclidean distance (MED) and the minimum product distance (MPD) of the proposed codebooks. Numerical simulation results show that our proposed codebooks lead to significantly improved error rate performances over Gaussian channels and Rayleigh fading channels.

Adder Neural Networks (ANNs) which only contain additions bring us a new way of developing deep neural networks with low energy consumption. Unfortunately, there is an accuracy drop when replacing all convolution filters by adder filters. The main reason here is the optimization difficulty of ANNs using $\ell_1$-norm, in which the estimation of gradient in back propagation is inaccurate. In this paper, we present a novel method for further improving the performance of ANNs without increasing the trainable parameters via a progressive kernel based knowledge distillation (PKKD) method. A convolutional neural network (CNN) with the same architecture is simultaneously initialized and trained as a teacher network, features and weights of ANN and CNN will be transformed to a new space to eliminate the accuracy drop. The similarity is conducted in a higher-dimensional space to disentangle the difference of their distributions using a kernel based method. Finally, the desired ANN is learned based on the information from both the ground-truth and teacher, progressively. The effectiveness of the proposed method for learning ANN with higher performance is then well-verified on several benchmarks. For instance, the ANN-50 trained using the proposed PKKD method obtains a 76.8\% top-1 accuracy on ImageNet dataset, which is 0.6\% higher than that of the ResNet-50.