Federated Learning in Unreliable and Resource-Constrained Cellular Wireless Networks

Abstract

Use cases of future wireless networks (e.g. fifth-generation [5G] networks and beyond [B5G]) will have servicequality requirements including higher data rates than today's networks for enhanced mobile broadband (eMBB), minimal latency and high network availability for ultra-reliability low-latency connection (URLLC), and massive access support for machine-type communications (mMTC). Also, 5G and B5G are expected to support communications for highly mobile scenarios with applications in new vertical sectors such as unmanned aerial vehicle (UAV) and autonomous car. Therefore, 5G and B5G cellular systems require a set of new technology enablers and solutions. In this thesis, we address some of the challenges of future wireless networks. In particular, we develop novel analytical models as well as methods, which will enable us to obtain insights into the performance of large-scale cellular networks and optimize network parameters.Non-orthogonal multiple access (NOMA) is a promising multiple access technique that enables massive connectivity and reduces the delay. We develop an analytical framework to derive the distribution of transmission success probabilities, meta distribution, for uplink and downlink NOMA. We also investigate the accuracy of distance-based ranking, instead of instantaneous signal power-based ranking, in the successive interference cancellation (SIC) at the NOMA receiver.Sojourn time, the time duration that a mobile user stays within a cell, is a mobility-aware parameter that can significantly impact the performance of mobile users and it can also be exploited to improve resource allocation and mobility management methods in the network. We derive the distribution and mean of the sojourn time in multi-tier cellular networks.Future wireless networks will exploit data-driven machine learning techniques for improving network management as well as service provisioning. Due to privacy and communication issues, learning at a centralized location (for example, at a base station) by collecting data from the mobile devices may not be always feasible. Federated learning is a machine learning setting where the centralized location trains a learning model using remote devices. Federated learning algorithms cannot be employed in real-world scenarios unless they consider unreliable and resource-constrained nature of the wireless medium. We propose a federated learning algorithm that is suitable for wireless networks.

Federated Learning in Unreliable and Resource-Constrained Cellular Wireless Networks

“…At the beginning of round k, k ∈ {0, 1, 2, ..., K−1}, we set the learning rate as η t = η 0 1+k , where kE ≤ t < (k + 1)E. We measure the performance of our model with regularized cross-entropy loss, where we use 2 -norm regularization with regularization parameter 10 Since we use a smaller learning rate for the synthetic dataset, we consider higher values for E at the time of simulation. Our codes are available at [29].…”

Time-triggered Federated Learning over Wireless Networks

“…where δ and ε are positive constants. (20) can easily be satisfied based on (21) and Cauchy-Schwartz inequality, since the left hand side of the inequality is the inner product of…”

“…Owing to the high communication efficiency gained in synchronized training settings, Sync FL algorithms are widely studied in wireless networks [17][18][19][20][21], and the main challenges of its wireless implementation lie in the resource-constrained and unreliable nature of wireless networks. Due to limited communication resource budgets, the user selection problem was studied to maximize the resource utilization with the focus on energy efficiency [17], training accuracy [19], training time [20], and etc.…”

Wireless for Machine Learning: A Survey

“…Utilizes the resulting upper bound to perform client selection, resource block allocation, and power allocation. [133] Client selection scheme that weighs the ML model update contribution of individual devices based on their probability of successful transmission. [83] Analyzes the convergence of SignSGDbased distributed learning.…”

“…Similar to [35], the authors in [133] consider a transmission success probability, complementary to the probability of error, which impacts the client scheduling policy and convergence analysis. The FL averaging step uses the success probability together with the scheduling policy and sends in the uplink the difference between the local model after E epochs, w t k (E), and the global model of the current communication round, w t .…”