A Family of Hybrid Federated and CentralizedLearning Architectures in Machine Learning

in IEEE Transactions on Cognitive Communications and Networking (2022)

Many of the machine learning tasks focus on cen-tralized learning (CL), which requires the transmission of localdatasets from the clients to a parameter server (PS) entailing hugecommunication overhead ... [more ▼]

Many of the machine learning tasks focus on cen-tralized learning (CL), which requires the transmission of localdatasets from the clients to a parameter server (PS) entailing hugecommunication overhead. To overcome this, federated learning(FL) has been a promising tool, wherein the clients send only themodel updates to the PS instead of the whole dataset. However,FL demands powerful computational resources from the clients.Therefore, not all the clients can participate in training if they donot have enough computational resources. To address this issue,we introduce a more practical approach called hybrid federatedand centralized learning (HFCL), wherein only the clients withsufficient resources employ FL, while the remaining ones sendtheir datasets to the PS, which computes the model on behalf ofthem. Then, the model parameters corresponding to all clientsare aggregated at the PS. To improve the efficiency of datasettransmission, we propose two different techniques: increasedcomputation-per-client and sequential data transmission. TheHFCL frameworks outperform FL with up to20%improvementin the learning accuracy when only half of the clients perform FLwhile having50%less communication overhead than CL since allthe clients collaborate on the learning process with their datasets. [less ▲]

A family of deep learning architectures for channel estimation and hybrid beamforming in multi-carrier mm-wave massive MIMO.

in IEEE Transactions on Cognitive Communications and Networking (2021)

Hybrid analog and digital beamforming transceivers are instrumental in addressing the challenge of expensive hardware and high training overheads in the next generation millimeter-wave (mm-Wave) massive ... [more ▼]

Hybrid analog and digital beamforming transceivers are instrumental in addressing the challenge of expensive hardware and high training overheads in the next generation millimeter-wave (mm-Wave) massive MIMO (multiple-input multiple-output) systems. However, lack of fully digital beamforming in hybrid architectures and short coherence times at mm-Wave impose additional constraints on the channel estimation. Prior works on addressing these challenges have focused largely on narrowband channels wherein optimization-based or greedy algorithms were employed to derive hybrid beamformers. In this paper, we introduce a deep learning (DL) approach for channel estimation and hybrid beamforming for frequency-selective, wideband mm-Wave systems. In particular, we consider a massive MIMO Orthogonal Frequency Division Multiplexing (MIMO-OFDM) system and propose three different DL frameworks comprising convolutional neural networks (CNNs), which accept the raw data of received signal as input and yield channel estimates and the hybrid beamformers at the output. We also introduce both offline and online prediction schemes. Numerical experiments demonstrate that, compared to the current state-of-the-art optimization and DL methods, our approach provides higher spectral efficiency, lesser computational cost and fewer number of pilot signals, and higher tolerance against the deviations in the received pilot data, corrupted channel matrix, and propagation environment. [less ▲]

Centralized Power Control in Cognitive Radio Networks Using Modulation and Coding Classification Feedback

in IEEE Transactions on Cognitive Communications and Networking (2016), 2(3),

In this paper, a centralized Power Control (PC) scheme and an interference channel learning method are jointly tackled to allow a Cognitive Radio Network (CRN) access to the frequency band of a Primary ... [more ▼]

In this paper, a centralized Power Control (PC) scheme and an interference channel learning method are jointly tackled to allow a Cognitive Radio Network (CRN) access to the frequency band of a Primary User (PU) operating based on an Adaptive Coding and Modulation (ACM) protocol. The learning process enabler is a cooperative Modulation and Coding Classification (MCC) technique which estimates the Modulation and Coding scheme (MCS) of the PU. Due to the lack of cooperation between the PU and the CRN, the CRN exploits this multilevel MCC sensing feedback as implicit channel state information (CSI) of the PU link in order to constantly monitor the impact of the aggregated interference it causes. In this paper, an algorithm is developed for maximizing the CRN throughput (the PC optimization objective) and simultaneously learning how to mitigate PU interference (the optimization problem constraint) by using only the MCC information. Ideal approaches for this problem setting with high convergence rate are the cutting plane methods (CPM). Here, we focus on the analytic center cutting plane method (ACCPM) and the center of gravity cutting plane method (CGCPM) whose effectiveness in the proposed simultaneous PC and interference channel learning algorithm is demonstrated through numerical simulations. [less ▲]

Cooperative Estimation of Power and Direction of Transmission for a Directive Source

in IEEE Transactions on Cognitive Communications and Networking (2016)

Reliable spectrum cartography of directive sources depends on an accurate estimation of the direction of transmission (DoT) as well as the transmission power. Joint estimation of power and DoT of a ... [more ▼]

Reliable spectrum cartography of directive sources depends on an accurate estimation of the direction of transmission (DoT) as well as the transmission power. Joint estimation of power and DoT of a directive source using ML estimation techniques is considered in this paper. We further analyze the parametric identifiability conditions of the problem, develop the estimation algorithm, and derive the Cramer-Rao-Bound (CRB) for the two situations: a) where the source signal is known to the sensors, and b) where the sensors are not aware of the source signal but its distribution. Particularly, we devise a specific sensor placement/selection setup for the symmetric antenna patterned sources which leads to identifiability of the problem. Finally, numerical results verifies the efficiency and accuracy of the provided estimation algorithms in this paper. [less ▲]