Federated learning (FL) is a creative technique that enables multiple edge devices to train a model without revealing raw data. However, several issues hinder the practical implementation of FL, especially in wireless environments. These issues comprise the limited capacity of the upload transmission link between the edge devices and the aggregator, as well as the wireless disturbances. To address these challenges, we develop a zero-order (ZO) communication-efficient framework for FL. While in standard FL, each device must upload a long vector containing the gradient or the model per communication round, our novel ZO method incorporates a two-point gradient estimator, which requires uploading only two scalars. What also sets our approach apart is that it directly incorporates wireless perturbations into the learning, eliminating the need for additional computational resources to remove their impact. In this work, we overcome the technical and analytical challenges associated with FL problems and ZO methods, comprehensively study our algorithm, and prove it converges almost surely under different conditions, convexity and non-convexity, noise-free and noisy environments. We then find theoretical bounds on the convergence rate when the objective is strongly convex, non-convex, and $\kappa$-gradient-dominated that compete with first-order (FO) or centralized methods under the same settings. Finally, we provide experimental results demonstrating the effectiveness of our algorithm, considering relevant examples. We provide an example illustrating the amount of communication saved due to its efficiency compared to its FO counterpart.

Zero-order (ZO) optimization is a powerful tool for dealing with realistic constraints. On the other hand, the gradient-tracking (GT) technique proved to be an efficient method for distributed optimization aiming to achieve consensus. However, it is a first-order (FO) method that requires knowledge of the gradient, which is not always possible in practice. In this work, we introduce a zero-order distributed optimization method based on a one-point estimate of the gradient tracking technique. We prove that this new technique converges with a single noisy function query at a time in the non-convex setting. We then establish a convergence rate of $O(\frac{1}{\sqrt[3]{K}})$ after a number of iterations K, which competes with that of $O(\frac{1}{\sqrt[4]{K}})$ of its centralized counterparts. Finally, a numerical example validates our theoretical results.

The initial access in 5G New Radio (NR) standalone millimeter wave (mmWave) uses the techniques of beam sweeping to Figure out appropriate directions of transmission and reception. Beam sweeping refers to the transmission of cell specific signals by the base station (BS) while switching beam direction in a sequential manner in order to cover the whole cell. The conventional beam sweeping consists of transmitting sequentially and periodically the beams, according to a fixed pattern, which results in an initial access latency. The aim of this paper is to minimize the overall number of timeslots needed to connect all users to the BS by allowing using multiple beams in some slots and then activating the beams in a prioritized and controlled manner rather than sequentially and periodically as in the conventional scheme. The problem is formulated as a combinatorial optimization problem, for which a Semidefinite Relaxation-based (SDR-based) solving approach is proposed. Numerical results demonstrate that our scheme outperforms significantly the conventional beam sweeping.

Reconfigurable intelligent surfaces (RIS) technology, as the name implies, is a grid of many small intelligent surfaces that can be reconfigured. It is a new and promising concept in wireless communications, expected to help realize the requirements of the future cellular generations. Numerous studies have been done to prove its added advantages and to control these surfaces in beneficial ways. However, these control schemes come with difficulties related to efficient and practical implementation. In this paper, we propose to control multiple RISs in a multi-user scheme with an algorithm that leads to a simple implementation. We formulate a stochastic optimization problem and we propose a new method using a distributed stochastic algorithm that allows the optimization to be done locally at each RIS with low computational requirements and without the need for instantaneous channel knowledge at the RIS. Also, the signaling between the base station (BS) and each RIS is limited to the exchange of a scalar only. Simulation results prove the success and efficiency of this algorithm.