Search

Recently, machine learning (ML) has shown its effectiveness in improving communication efficiency by reinstating the semantics of bits. To understand its underlying principles, we propose a novel stochastic model of semantic communication, dubbed semantics-native communication (SNC). Inspired from human communication, we consider a point-to-point SNC scenario where a speaker has an intention of referring to an entity, extracts its semantic concepts, and maps these concepts into symbols that are communicated to a target listener through the traditional Shannon communication channel. Next, we recall rational humans who can communicate more efficiently by reasoning about the others’ contexts before communication, referred to as contextual reasoning. This motivates us to propose a novel SNC framework harnessing agent states as side information in a way that the speaker locally and iteratively communicates with a virtual agent having the listener’s state, and vice versa. Theoretically, we prove the convergence of contextual reasoning, at which it minimizes the bit length while guaranteeing a target reliability. Simulation results corroborate that contextual reasoning based SNC can significantly reduce bit lengths, and be a robust solution to imperfect agent states by quantizing the entity-concept-symbol mapping before contextual reasoning.

Semantic communication (SemCom) aims to convey the meaning behind a transmitted message by transmitting only semantically-relevant information. This semantic-centric design helps to minimize power usage, bandwidth consumption, and transmission delay. SemCom and goal-oriented SemCom (or effectiveness-level SemCom) are therefore promising enablers of 6G and developing rapidly. Despite the surge in their swift development, the design, analysis, optimization, and realization of robust and intelligent SemCom as well as goal-oriented SemCom are fraught with many fundamental challenges. One of the challenges is that the lack of unified/universal metrics of SemCom and goal-oriented SemCom can stifle research progress on their respective algorithmic, theoretical, and implementation frontiers. Consequently, this survey paper documents the existing metrics– scattered in many references– of wireless SemCom, optical SemCom, quantum SemCom, and goal-oriented wireless SemCom. By doing so, this paper aims to inspire the design, analysis, and optimization of a wide variety of SemCom and goal-oriented SemCom systems. This article also stimulates the development of unified/universal performance assessment metrics of SemCom and goal-oriented SemCom, as the existing metrics are purely statistical and hardly applicable to reasoning-type tasks that constitute the heart of 6G and beyond.

Prolonging the lifetime of massive machine-type communication (MTC) networks is key to realizing a sustainable digitized society. Great energy savings can be achieved by accurately predicting MTC traffic followed by properly designed resource allocation mechanisms. However, selecting the proper MTC traffic predictor is not straightforward and depends on accuracy/complexity trade-offs and the specific MTC applications and network characteristics. Remarkably, the related state-of-the-art literature still lacks such debates. Herein, we assess the performance of several machine learning (ML) methods to predict Poisson and quasi-periodic MTC traffic in terms of accuracy and computational cost. Results show that the temporal convolutional network (TCN) outperforms the long-short term memory (LSTM), the gated recurrent units (GRU), and the recurrent neural network (RNN), in that order. For Poisson traffic, the accuracy gap between the predictors is larger than under quasi-periodic traffic. Finally, we show that running a TCN predictor is around three times more costly than other methods, while the training/inference time is the greatest/least.

Beamspace multiple-input-multiple-output (MIMO) as a green technology can efficiently substitute for the conventional massive MIMO, provided that the beamspace channel is acquired precisely. The prior efforts in this area of study, especially the learning-driven ones, however, indicate remarkable performance losses owing to a lack of generalization. In this paper, we propose a modified non-linear auto-regressive exogenous (NARX) model for tracking and predicting the beamspace channel over the sequences of time. Benefiting from bounded generalization error, fast convergence, limited prediction variance, and negligible performance loss, the proposed scheme achieves up to 15% spectral efficiency (SE) gain over its counterparts. We further improve this performance by means of an ensemble learning technique for simultaneously training multiple NARX modules in parallel, thus leading to a 23% SE gain. Relying on the predicted beamspace channel, we propose a beamspace analog beam selection technique through fine-tuning the architecture of a pre-trained off-the-shelf GoogleNet, which brings up to 21% SE gain over similar baselines. With the aid of an ensemble learning technique, it is further indicated numerically that up to 34% SE improvement can be achieved, as compared to the counterparts.

Federated learning (FL) is a key enabler for efficient communication and computing, leveraging devices’ distributed computing capabilities. However, applying FL in practice is challenging due to the local devices’ heterogeneous energy, wireless channel conditions, and non-independently and identically distributed (non-IID) data distributions. To cope with these issues, this paper proposes a novel learning framework by integrating FL and width-adjustable slimmable neural networks (SNN). Integrating FL with SNNs is challenging due to time-varying channel conditions and data distributions. In addition, existing multi-width SNN training algorithms are sensitive to the data distributions across devices, which makes SNN ill-suited for FL. Motivated by this, we propose a communication and energy-efficient SNN-based FL (named SlimFL ) that jointly utilizes superposition coding (SC) for global model aggregation and superposition training (ST) for updating local models. By applying SC, SlimFL exchanges the superposition of multiple-width configurations decoded as many times as possible for a given communication throughput. Leveraging ST, SlimFL aligns the forward propagation of different width configurations while avoiding inter-width interference during backpropagation. We formally prove the convergence of SlimFL. The result reveals that SlimFL is not only communication-efficient but also deals with non-IID data distributions and poor channel conditions, which is also corroborated by data-intensive simulations.

In this paper, we propose a decentralized Newton-type approach to solve the problem of decentralized federated learning (FL). Notably, our proposed algorithm leverages the fast convergence of the second-order methods while avoid sending the hessian matrix at each iteration. Therefore, the proposed approach significantly reduces the communication cost and preserves the privacy. Specifically, we alternate between two problems. The inner problem approximates the inverse Hessian-gradient product which is formulated as a quadratic optimization problem and approximately solved in a decentralized manner using one step of the group alternating direction method of multipliers (GADMM) method. The outer problem learns the model, which is solved by performing one decentralized Newton step at every iteration. Moreover, to reduce the communication-overhead per iteration, a quantized version (leveraging stochastic quantization) is also proposed. Simulation results illustrate that our algorithm outperforms the baselines of GADMM, Q-GADMM, Newton tracking, and Decentralized SGD, and provides energy and communication-efficient solutions for bandwidth-limited systems under different SNR regimes.

In this paper, we propose to solve a regularized distributionally robust learning problem in the decentralized setting, taking into account the data distribution shift. By adding a Kullback-Liebler regularization function to the robust min-max optimization problem, the learning problem can be reduced to a modified robust minimization problem and solved efficiently. Leveraging the newly formulated optimization problem, we propose a robust version of Decentralized Stochastic Gradient Descent (DSGD), coined Distributionally Robust Decentralized Stochastic Gradient Descent (DR-DSGD). Under some mild assumptions and provided that the regularization parameter is larger than one, we theoretically prove that DR-DSGD achieves a convergence rate of O(1/√KT + K/T), where K is the number of devices and T is the number of iterations. Simulation results show that our proposed algorithm can improve the worst distribution test accuracy by up to 10%. Moreover, DR-DSGD is more communication-efficient than DSGD since it requires fewer communication rounds (up to 20 times less) to achieve the same worst distribution test accuracy target. Furthermore, the conducted experiments reveal that DR-DSGD results in a fairer performance across devices in terms of test accuracy.

To meet the growing mobility needs in intra-city transportation, urban air mobility (UAM) has been proposed in which vertical takeoff and landing (VTOL) aircraft are used to provide on-demand service. In UAM, an aircraft can operate in the corridors, i.e., the designated airspace, that link the aerodromes, thus avoiding the use of complex routing strategies such as those of modern-day helicopters. For safety, a UAM aircraft will use air-to-ground communications to report flight plan, off-nominal events, and real-time movements to ground base stations (GBSs). A reliable communication network between GBSs and aircraft enables UAM to adequately utilize the airspace and create a fast, efficient, and safe transportation system. In this paper, to characterize the wireless connectivity performance in UAM, a stochastic geometry-based spatial model is developed. In particular, the distribution of GBSs is modeled as a Poisson point process (PPP), and the aircraft are distributed according to a combination of PPP, Poisson cluster process (PCP), and Poisson line process (PLP). For this setup, assuming that any given aircraft communicates with the closest GBS, the distribution of distance between an arbitrarily selected GBS and its associated aircraft and the Laplace transform of the interference experienced by the GBS are derived. Using these results, the signal-to-interference ra-tio (SIR)-based connectivity probability is determined to capture the connectivity performance of the aircraft-to-ground communication network in UAM. Simulation results validate the theoretical derivations for the UAM wireless connectivity and provide useful UAM design guidelines by showing the connectivity performance under different parameter settings.

In this paper, the content popularity prediction problem in cache-enabled fog radio access networks (F-RANs) is investigated. In order to predict the content popularity with high accuracy and low complexity, we propose a Gaussian process based Poisson regressor to model the content request pattern. Firstly, the relationship between content features and popularity is captured by our developed model. Then, we utilize Bayesian learning to learn the model parameters, which are robust to over-fitting. However, Bayesian methods are usually unable to find a closed-form expression of the posterior distribution. To tackle this issue, we apply a Stochastic Variance Reduced Gradient Hamiltonian Monte Carlo (SVRG-HMC) to approximate the posterior distribution. Two types of predictive content popularity are formulated for the requests of existing contents and newly-added contents. Simulation results show that the performance of our proposed policy outperforms the policy based on other Monte Carlo based method.

Meta-learning is a powerful learning paradigm in which solving a new task can benefit from similar tasks for faster adaption (few shot learning). Stochastic gradient descent (SGD) based meta learning has emerged as an attractive solution in the few-shot learning. However, this approach suffers from significant computational complexity due to the double loop and matrix inversion operations which incurs a significant amount of uncertainty and poor generalization. To achieve lower complexity and better generalization, in this paper, we propose MetaBayes, a novel framework that views the original meta learning problem from a Bayesian perspective where the meta-model is cast as the prior distribution and the task-specific models are viewed as task-specific posterior distributions. The objective amounts to jointly optimizing the prior and the posterior distributions. With this, we obtain a closed-form expression to update the distributions at every iteration, to avoid the high computation cost issue of SGD based meta learning, and produce a more robust and generalized meta-model. Our simulations show that tasks with few training samples achieves higher accuracy when MetaBayes prior distribution is used as an initializer compared to the commonly-used Gaussian prior distribution.

In this paper, we propose an energy-efficient federated meta-learning framework. The objective is to enable learning a meta-model that can be fine-tuned to a new task with a few number of samples in a distributed setting and at low computation and communication energy consumption. We assume that each task is owned by a separate agent, so a limited number of tasks is used to train a meta-model. Assuming each task was trained offline on the agent’s local data, we propose a lightweight algorithm that starts from the local models of all agents, and in a backward manner using projected stochastic gradient ascent (P-SGA) finds a meta-model. The proposed method avoids complex computations such as computing hessian, double looping, and matrix inversion, while achieving high performance at significantly less energy consumption compared to the state-of-the-art methods such as MAML and iMAML on conducted experiments for sinusoid regression and image classification tasks.

Split learning (SL) is a promising distributed learning framework that enables to utilize the huge data and parallel computing resources of mobile devices. SL is built upon a model-split architecture, wherein a server stores an upper model segment that is shared by different mobile clients storing its lower model segments. Without exchanging raw data, SL achieves high accuracy and fast convergence by only uploading smashed data from clients and downloading global gradients from the server. Nonetheless, the original implementation of SL sequentially serves multiple clients, incurring high latency with many clients. A parallel implementation of SL has great potential in reducing latency, yet existing parallel SL algorithms resort to compromising scalability and/or convergence speed. Motivated by this, the goal of this article is to develop a scalable parallel SL algorithm with fast convergence and low latency. As a first step, we identify that the fundamental bottleneck of existing parallel SL comes from the model-split and parallel computing architectures, under which the server-client model updates are often imbalanced, and the client models are prone to detach from the server’s model. To fix this problem, by carefully integrating local parallelism, federated learning, and mixup augmentation techniques, we propose a novel parallel SL framework, coined LocFedMix-SL. Simulation results corroborate that LocFedMix-SL achieves improved scalability, convergence speed, and latency, compared to sequential SL as well as the state-of-the-art parallel SL algorithms such as SplitFed and LocSplitFed.

Integrating low-altitude earth orbit (LEO) satellites (SATs) and unmanned aerial vehicles (UAVs) within a non-terrestrial network (NTN), we investigate the problem of forwarding packets between two faraway ground terminals through SAT and UAV relays using either radio-frequency (RF) or free-space optical (FSO) link. Towards maximizing the communication efficiency, the associations with orbiting SATs and the trajectories of UAVs should be optimized, which is challenging due to the time-varying network topology and a huge number of possible control actions. To overcome the difficulty, we lift this problem to multi-agent deep reinforcement learning with a novel action dimensionality reduction technique. Simulation results corroborate that our proposed SAT-UAV integrated scheme achieves 1.99x higher end-to-end sum throughput compared to a benchmark scheme with fixed ground relays. While improving the throughput, our proposed scheme also aims to reduce the UAV control energy, yielding 2.25x higher energy efficiency than a baseline method only maximizing the throughput. Lastly, thanks to utilizing hybrid FSO/RF links, the proposed scheme achieves up to 62.56x higher peak throughput and 21.09x higher worst-case throughput than the cases utilizing either RF or FSO links, highlighting the importance of co-designing SAT-UAV associations, UAV trajectories, and hybrid FSO/RF links in beyond-5 G NTNs.

In the Internet of Things (IoT) environment, edge computing can be initiated at anytime and anywhere. However, in an IoT environment, edge computing sessions are often ephemeral, i.e., they last for a short period of time and can often be discontinued once the current application usage is completed or the edge devices leave the system due to factors such as mobility. Therefore, in this paper, the problem of ephemeral edge computing in an IoT is studied by considering scenarios in which edge computing operates within a limited time period. To this end, a novel online framework is proposed in which a source edge node offloads its computing tasks from sensors within an area to neighboring edge nodes for distributed task computing, within the limited period of time of an ephemeral edge computing system. The online nature of the framework allows the edge nodes to optimize their task allocation and decide on which neighbors to use for task processing, even when the tasks are revealed to the source edge node in an online manner, and the information on future task arrivals is unknown. The proposed framework essentially maximizes the number of computed tasks by jointly considering the communication and computation latency. To solve the joint optimization, an online greedy algorithm is proposed and solved by using the primal-dual approach. Since the primal problem provides an upper bound of the original dual problem, the competitive ratio of the online approach is analytically derived as a function of the task sizes and the data rates of the edge nodes. Simulation results show that the proposed online algorithm can achieve a near-optimal task allocation with an optimality gap that is no higher than 7.1% compared to the offline, optimal solution with complete knowledge of all tasks.

Recently, it has been shown that analog transmission based federated learning enables more efficient usage of communication resources compared to the conventional digital transmission. In this paper, we propose an effective model compression strategy enabling analog FL under constrained communication bandwidth. To this end, the proposed approach is based on pattern shared sparsification by setting the same sparsification pattern of parameter vectors uploaded by edge devices, as opposed to each edge device independently applying sparsification. In particular, we propose specific schemes for determining the sparsification pattern and characterize the convergence of analog FL leveraging these proposed sparsification strategies, by deriving a closed-form upper boun d of convergence rate and residual error. The closed-form expression allows to capture the effect of communication bandwidth and power budget to the performance of analog FL. In terms of convergence analysis, the model parameter obtained with the proposed schemes is proven to converge to the optimum of model parameter. Numerical results show that leveraging the proposed pattern shared sparsification consistently improves the performance of analog FL in various settings of system parameters. The improvement in performance is more significant under scarce communication bandwidth and limited transmit power budget.

Classical and centralized Artificial Intelligence (AI) methods require moving data from producers (sensors, machines) to energy hungry data centers, raising environmental concerns due to computational and communication resource demands, while violating privacy. Emerging alternatives to mitigate such high energy costs propose to efficiently distribute, or federate, the learning tasks across devices, which are typically low-power. This paper proposes a novel framework for the analysis of energy and carbon footprints in distributed and federated learning (FL). The proposed framework quantifies both the energy footprints and the carbon equivalent emissions for vanilla FL methods and consensus-based fully decentralized approaches. We discuss optimal bounds and operational points that support green FL designs and underpin their sustainability assessment. Two case studies from emerging 5G industry verticals are analyzed: these quantify the environmental footprints of continual and reinforcement learning setups, where the training process is repeated periodically for continuous improvements. For all cases, sustainability of distributed learning relies on the fulfillment of specific requirements on communication efficiency and learner population size. Energy and test accuracy should be also traded off considering the model and the data footprints for the targeted industrial applications.

A mega-constellation of low-altitude earth orbit (LEO) satellites (SATs) are envisaged to provide a global coverage SAT network in beyond fifth-generation (5G) cellular systems. LEO SAT networks exhibit extremely long link distances of many users under time-varying SAT network topology. This makes existing multiple access protocols, such as random access channel (RACH) based cellular protocol designed for fixed terrestrial network topology, ill-suited. To overcome this issue, in this paper, we propose a novel contention-based random access solution for LEO SAT networks, dubbed emergent random access channel protocol (eRACH). In stark contrast to existing model-based and standardized protocols, eRACH is a model-free approach that emerges through interaction with the non-stationary network environment, using multi-agent deep reinforcement learning (MADRL). Furthermore, by exploiting known SAT orbiting patterns, eRACH does not require central coordination or additional communication across users, while training convergence is stabilized through the regular orbiting patterns. Compared to RACH, we show from various simulations that our proposed eRACH yields 54.6% higher average network throughput with around two times lower average access delay while achieving 0.989 Jain’s fairness index.

This paper aims to integrate two synergetic technologies, federated learning (FL) and width-adjustable slimmable neural network (SNN) architectures. FL preserves data privacy by exchanging the locally trained models of mobile devices. By adopting SNNs as local models, FL can flexibly cope with the time-varying energy capacities of mobile devices. Combining FL and SNNs is however non-trivial, particularly under wireless connections with time-varying channel conditions. Furthermore, existing multi-width SNN training algorithms are sensitive to the data distributions across devices, so are ill-suited to FL. Motivated by this, we propose a communication and energy efficient SNN-based FL (named SlimFL) that jointly utilizes superposition coding (SC) for global model aggregation and superposition training (ST) for updating local models. By applying SC, SlimFL exchanges the superposition of multiple width configurations that are decoded as many as possible for a given communication throughput. Leveraging ST, SlimFL aligns the forward propagation of different width configurations, while avoiding the inter-width interference during back propagation. We formally prove the convergence of SlimFL. The result reveals that SlimFL is not only communication-efficient but also can counteract non-IID data distributions and poor channel conditions, which is also corroborated by simulations.

The emergent technology of Reconfigurable Intelligent Surfaces (RISs) has the potential to transform wireless environments into controllable systems, through programmable propagation of information-bearing signals. Techniques stemming from the field of Deep Reinforcement Learning (DRL) have recently gained popularity in maximizing the sum-rate performance in multi-user communication systems empowered by RISs. Such approaches are commonly based on Markov Decision Processes (MDPs). In this paper, we instead investigate the sum-rate design problem under the scope of the Multi-Armed Bandits (MAB) setting, which is a relaxation of the MDP framework. Nevertheless, in many cases, the MAB formulation is more appropriate to the channel and system models under the assumptions typically made in the RIS literature. To this end, we propose a simpler DRL approach for orchestrating multiple metasurfaces in RIS-empowered multi-user Multiple-Input Single-Output (MISO) systems, which we numerically show to perform equally well with a state-of-the-art MDP-based approach, while being less demanding computationally.

In this paper, the cooperative caching problem in fog radio access networks (F-RANs) is investigated to jointly optimize the transmission delay and energy consumption. Exploiting the potential social relationships among fog access points (F-APs), we firstly propose a clustering scheme based on hedonic coalition game (HCG) to improve the potential cooperation gain. Then, considering that the optimization problem is non-deterministic polynomial hard (NP-hard), we further propose an improved firefly algorithm (FA) based cooperative caching scheme, which utilizes a mutation strategy based on local content popularity to avoid pre-mature convergence. Simulation results show that our proposed scheme can effectively reduce the content transmission delay and energy consumption in comparison with the baselines.

Industrial Internet of things (IIoT), one enabler for Industry 4.0 Smart Factories, is a mission-critical and latency-sensitive application of 5G networks. Due to the stringent latency requirements in IIoT, coordinating the simultaneous transmissions of massive entities and knowing the interference they create to each other is not feasible. Additionally, due to the mobility feature of mobile robots and automated guided vehicles, the experienced channel fading may differ from the estimated one. Therefore, some uncertainties exist in IIoT networks while we decide the communication and control mechanisms. Within the context of IIoT, this paper discusses some resource allocation solutions from the perspective of Online Convex Optimization (OCO). OCO is a computationally lightweight and memory-efficient mathematical tool which tackles the optimization problems, given that the network environment is arbitrary and unknown. We first introduce the key performance indicators in IIoT networks and highlight the uncertain factors, which we may encounter while allocating the communication resources in IIoT. Then we provide an overview of main principles of OCO and present the comparison benchmarks and related metrics for performance evaluation. Moreover, we discuss the kind of resource allocation problems in IIoT that can be tackled by OCO. Finally, we summarize the advantages of applying OCO to IIoT networks.

The fog radio access network (F-RAN) is a promising technology in which the user mobile devices (MDs) can offload computation tasks to the nearby fog access points (F-APs). Due to the limited resource of F-APs, it is important to design an efficient task offloading scheme. In this paper, by considering time-varying network environment, a dynamic computation offloading and resource allocation problem in F-RANs is formulated to minimize the task execution delay and energy consumption of MDs. To solve the problem, a federated deep reinforcement learning (DRL) based algorithm is proposed, where the deep deterministic policy gradient (DDPG) algorithm performs computation offloading and resource allocation in each F-AP. Federated learning is exploited to train the DDPG agents in order to decrease the computing complexity of training process and protect the user privacy. Simulation results show that the proposed federated DDPG algorithm can achieve lower task execution delay and energy consumption of MDs more quickly compared with the other existing strategies.

In this paper, the placement strategy design of coded caching in fog-radio access networks (F -RANs) is investigated. By considering time-variant content popularity, federated deep re-inforcement learning is exploited to learn the placement strategy for our coded caching scheme. Initially, the placement problem is modeled as a Markov decision process (MDP) to capture the popularity variations and minimize the long-term content access delay. The reformulated sequential decision problem is solved by dueling double deep Q-learning (dueling DDQL). Then, federated learning is applied to learn the relatively low-dimensional local decision models and aggregate the global decision model, which alleviates over-consumption of bandwidth resources and avoids direct learning of a complex coded caching decision model with high-dimensional state space. Simulation results show that our proposed scheme outperforms the benchmarks in reducing the content access delay, keeping the performance stable, and trading off between the local caching gain and the global multicasting gain.

Semantic communication is an emerging paradigm that focuses on understanding and delivering semantics or meaning of messages. Most existing semantic communication solutions define semantic meaning as the labels of objects recognized from a given form of source signal, while ignoring intrinsic information that cannot be directly observed. Since the models for recognizing labels need to be pre-trained with labelled dataset, the total number of semantic objects are often limited by a fixed set. In this paper, we propose a novel reasoning-based semantic communication architecture in which the semantic meaning is represented by a graph-based knowledge structure in terms of semantic-entity, relationships, and reasoning rules. An embedding-based semantic interpretation framework is proposed to convert the high-dimensional graph-based representation of semantic meaning into a low-dimensional representation, which is efficient for channel transmission. We develop a novel inference function-based approach that can automatically infer hidden information such as missing entities and relations that cannot be directly observed from the message. Finally, we introduce a life-long model updating approach in which the receiver can learn from previously received messages and automatically update the rules for reasoning the hidden information when new unknown semantic entities and relations have been discovered. Extensive experiments are conducted based on a real-world knowledge database and numerical results show that our proposed solution achieves 76% interpretation accuracy of the hidden meaning at the receiver when some entities are missing in the transmitted message.

Multi-access edge computing (MEC) is viewed as an integral part of future wireless networks to support new applications with stringent service reliability and latency requirements. However, guaranteeing ultra-reliable and low-latency MEC (URLL MEC) is very challenging due to uncertainties of wireless links, limited communications and computing resources, as well as dynamic network traffic. Enabling URLL MEC man-dates taking into account the statistics of the end-to-end (E2E) latency and reliability across the wireless and edge computing systems. In this paper, a novel framework is proposed to optimize the reliability of MEC networks by considering the distribution of E2E service delay, encompassing over-the-air transmission and edge computing latency. The proposed framework builds on correlated variational autoencoders (VAEs) to estimate the full distribution of the E2E service delay. Using this result, a new optimization problem based on risk theory is formulated to maximize the network reliability by minimizing the Conditional Value at Risk (CVaR) as a risk measure of the E2E service delay. To solve this problem, a new algorithm is developed to efficiently allocate users’ processing tasks to edge computing servers across the MEC network, while considering the statistics of the E2E service delay learned by VAEs. The simulation results show that the proposed scheme outperforms several baselines that do not account for the risk analyses or statistics of the E2E service delay.

Real-time environment monitoring is a key application in Industrial Internet of Things, where sensors proactively collect and transmit environmental data to the controller. However, due to limited wireless resources, keeping sensors’ sampled data fresh at the controller is critical. This work aims to investigate the trade-off between the sensor’s data-sampling frequency and long-term data transmission energy consumption while maintaining information freshness. Leveraging the entropic risk measure (ERM), we jointly minimize the global transmission energy’s mean and variance subject to probabilistic constraints on information freshness. Furthermore, while jointly saving the model training energy, we adopt the federated learning (FL) paradigm and propose an FL-based two-stage iterative optimization framework to optimize the aforementioned objective. Specifically, we iteratively learn the sampling frequency via Bayesian optimization and minimize the long-term ERM of the global energy consumption via Lyapunov optimization. Numerical results show that the proposed FL-based scheme saves substantial executing energy with less performance loss. Quantitatively, compared with the centralized learning baseline, the proposed FL-based framework saves up to 69% model training energy at the expense of a mere increased objective outcome, i.e., 6.3% in the global data transmission energy consumption ( 9.936×10⁻⁵ in ERM) under 0.4% bias from the global optimal data-sampling frequency.

In this paper, we investigate maximum distance separable (MDS) codes based group coded caching in fog radio access networks (F-RANs). The goal is to minimize the average fronthaul rate under nonuniform file popularity. Firstly, an MDS codes and file grouping based coded placement scheme is proposed to provide coded packets and allocate more cache to the most popular files simultaneously. Next, a fog access point (F-AP) grouping based coded delivery scheme is proposed to meet the requests for files from different groups. Furthermore, a closed-form expression of the average fronthaul rate is derived. Finally, the parameters related to the proposed coded caching scheme are optimized to fully utilize the gains brought by MDS codes and file grouping. Simulation results show that our proposed scheme obtains significant performance improvement over several existing caching schemes in terms of fronthaul rate reduction.

Recent advances in Federated Learning (FL) have paved the way towards the design of novel strategies for solving multiple learning tasks simultaneously, by leveraging cooperation among networked devices. Multi-Task Learning (MTL) exploits relevant commonalities across tasks to improve efficiency compared with traditional transfer learning approaches. By learning multiple tasks jointly, significant reduction in terms of energy footprints can be obtained. This article provides a first look into the energy costs of MTL processes driven by the Model-Agnostic Meta-Learning (MAML) paradigm and implemented in distributed wireless networks. The paper targets a clustered multi-task network setup where autonomous agents learn different but related tasks. The MTL process is carried out in two stages: the optimization of a meta-model that can be quickly adapted to learn new tasks, and a task-specific model adaptation stage where the learned meta-model is transferred to agents and tailored for a specific task. This work analyzes the main factors that influence the MTL energy balance by considering a multi-task Reinforcement Learning (RL) setup in a robotized environment. Results show that the MAML method can reduce the energy bill by at least 2 × compared with traditional approaches without inductive transfer. Moreover, it is shown that the optimal energy balance in wireless networks depends on uplink/downlink and sidelink communication efficiencies.

Next-generation wireless systems are rapidly evolving from communication-only systems to multi-modal systems with integrated sensing and communications. In this paper a novel joint sensing and communication framework is proposed for enabling wireless extended reality (XR) at terahertz (THz) bands. To gather rich sensing information and a higher line-of-sight (LoS) availability, THz-operated reconfigurable intelligent surfaces (RISs) acting as base stations are deployed. The sensing parameters are extracted by leveraging THz’s quasi-opticality and opportunistically utilizing uplink communication waveforms. This enables the use of the same waveform, spectrum, and hardware for both sensing and communication purposes. The environmental sensing parameters are then derived by exploiting the sparsity of THz channels via tensor decomposition. Hence, a high-resolution indoor mapping is derived so as to characterize the spatial availability of communications and the mobility of users. Simulation results show that in the proposed framework, the resolution and data rate of the overall system are positively correlated, thus allowing a joint optimization between these metrics with no tradeoffs. Results also show that the proposed framework improves the system reliability in static and mobile systems. In particular, the highest reliability gains of 10% are achieved in a walking speed mobile environment compared to communication only systems with beam tracking.

The deployment of future intelligent transportation systems is contingent upon seamless and reliable operation of connected and autonomous vehicles (CAVs). One key challenge in developing CAVs is the design of an autonomous controller that can accurately execute near real-time control decisions, such as a quick acceleration when merging to a highway and frequent speed changes in a stop-and-go traffic. However, the use of conventional feedback controllers or traditional learning-based controllers, solely trained by each CAV’s local data, cannot guarantee a robust controller performance over a wide range of road conditions and traffic dynamics. In this paper, a new federated learning (FL) framework enabled by large-scale wireless connectivity is proposed for designing the autonomous controller of CAVs. In this framework, the learning models used by the controllers are collaboratively trained among a group of CAVs. To capture the varying CAV participation in the FL training process and the diverse local data quality among CAVs, a novel dynamic federated proximal (DFP) algorithm is proposed that accounts for the mobility of CAVs, the wireless fading channels, as well as the unbalanced and non-independent and identically distributed data across CAVs. A rigorous convergence analysis is performed for the proposed algorithm to identify how fast the CAVs converge to using the optimal autonomous controller. In particular, the impacts of varying CAV participation in the FL process and diverse CAV data quality on the convergence of the proposed DFP algorithm are explicitly analyzed. Leveraging this analysis, an incentive mechanism based on contract theory is designed to improve the FL convergence speed. Simulation results using real vehicular data traces show that the proposed DFP-based controller can accurately track the target CAV speed over time and under different traffic scenarios. Moreover, the results show that the proposed DFP algorithm has a much faster conver

The newly emerging federated learning (FL) framework offers a new way to train machine learning models in a privacy-preserving manner. However, traditional FL algorithms are based on an event-triggered aggregation, which suffers from stragglers and communication overhead issues. To address these issues, in this paper, we present a time-triggered FL algorithm (TT-Fed) over wireless networks, which is a generalized form of classic synchronous and asynchronous FL. Taking the constrained resource and unreliable nature of wireless communication into account, we jointly study the user selection and bandwidth optimization problem to minimize the FL training loss. To solve this joint optimization problem, we provide a thorough convergence analysis for TT-Fed. Based on the obtained analytical convergence upper bound, the optimization problem is decomposed into tractable sub-problems with respect to each global aggregation round, and finally solved by our proposed online search algorithm. Simulation results show that compared to asynchronous FL (FedAsync) and FL with asynchronous user tiers (FedAT) benchmarks, our proposed TT-Fed algorithm improves the converged test accuracy by up to 12.5% and 5%, respectively, under highly imbalanced and non-IID data, while substantially reducing the communication overhead.

To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents’ local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.

In this paper, the cooperative edge caching problem in fog radio access networks (F-RANs) is investigated. To minimize the content transmission delay, we formulate the cooperative caching optimization problem to find the globally optimal caching strategy. By considering the non-deterministic polynomial hard (NP-hard) property of this problem, a Multi Agent Reinforcement Learning (MARL)-based cooperative caching scheme is proposed. Our proposed scheme applies a double deep Q-network (DDQN) in every fog access point (F-AP), and introduces the communication process in a multi-agent system. Every F-AP records the historical caching strategies of its associated F-APs as the observations of communication procedure. By exchanging the observations, F-APs can leverage the cooperation and make the globally optimal caching strategy. Simulation results show that the proposed MARL-based cooperative caching scheme has remarkable performance compared with the benchmark schemes in minimizing the content transmission delay.

This study demonstrates the feasibility of image inpainting using both visual information and radio frequency (RF) signals. Recent developments in imaging and vision-based technologies using RF signals have revealed the potential of leveraging multimodal information to enhance image inpainting performance. In this context, we propose RF-Inpainter—a novel inpainting method that integrates visual and wireless information by fusing defective RGB images with received signal strength indicator (RSSI) using a deep auto-encoder model. The inpainting performance of RF-Inpainter is evaluated using experimentally obtained images and RSSI datasets in an indoor environment. Image-only inpainting and RSSI-only inpainting models are used as baselines to illustrate the superiority of RF-Inpainter over inpainting methods based on a single modality. The results establish that RF-Inpainter generates satisfactory inpainted images in most experimental scenarios, achieving a maximum improvement of 36.4% and 14.6% in terms of mean peak signal-to-noise ratio (PSNR) and mean structural similarity index (SSIM), respectively.

In this paper, we propose an accurate two-phase millimeter-Wave (mmWave) beamspace channel tracking mechanism. Particularly in the first phase, we train an extreme reservoir machine (ERM) for tracking the historical features of the mmWave beamspace channel and predicting them in upcoming time steps. Towards a more accurate prediction, we further fine-tune the ERM by means of Xavier initializer technique, whereby the input weights in ERM are initially derived from a zero mean and finite variance Gaussian distribution, leading to 49% degradation in prediction variance of the conventional ERM. The proposed method numerically improves the achievable spectral efficiency (SE) of the existing counterparts, by 13%, when signal-to-noise-ratio (SNR) is 15dB. We further investigate an ensemble learning technique in the second phase by sequentially incorporating multiple ERMs to form an ensembled model, namely adaptive boosting (AdaBoost), which further reduces the prediction variance in conventional ERM by 56%, and concludes in 21% enhancement of achievable SE upon the existing schemes at SNR = 15dB.

In this paper, we consider the uplink channel estimation phase and downlink data transmission phase of cell-free millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) systems with low-capacity fronthaul links and low-resolution analog-to-digital converters/digital-to-analog converters (ADC/DACs). In cell-free massive MIMO, a control unit dictates the baseband processing at a geographical scale, while the base stations communicate with the control unit through fronthaul links. Unlike most of previous works in cell-free massive MIMO with finite-capacity fronthaul links, we consider the general case where the fronthaul capacity and ADC/DAC resolution are not necessarily the same. In particular, the fronthaul compression and ADC/DAC quantization occur independently where each one is modeled based on the information theoretic argument and additive quantization noise model (AQNM). Then, we address the codebook design problem that aims to minimize the channel estimation error for the independent and identically distributed (i.i.d.) and colored compression noise cases. Also, we propose an alternating optimization (AO) method to tackle the max-min fairness problem. In essence, the AO method alternates between two subproblems that correspond to the power allocation and codebook design problems. The AO method proposed for the zero-forcing (ZF) precoder is guaranteed to converge, whereas the one for the maximum ratio transmission (MRT) precoder has no such guarantee. Finally, the performance of the proposed schemes is evaluated by the simulation results in terms of both energy and spectral efficiency. The numerical results show that the proposed scheme for the ZF precoder yields spectral and energy efficiency 28% and 15% higher than that of the best baseline.

In this paper, the content popularity prediction problem in fog radio access networks (F-RANs) is investigated. Based on clustered federated learning, we propose a novel mobility-aware popularity prediction policy, which integrates content popularities in terms of local users and mobile users. For local users, the content popularity is predicted by learning the hidden representations of local users and contents. Initial features of local users and contents are generated by incorporating neighbor information with self information. Then, dual-channel neural network (DCNN) model is introduced to learn the hidden representations by producing deep latent features from initial features. For mobile users, the content popularity is predicted via user preference learning. In order to distinguish regional variations of content popularity, clustered federated learning (CFL) is employed, which enables fog access points (F-APs) with similar regional types to benefit from one another and provides a more specialized DCNN model for each F-AP. Simulation results show that our proposed policy achieves significant performance improvement over the traditional policies.

Real-time remote control over wireless is an important yet challenging application in fifth-generation and beyond due to its mission-critical nature under limited communication resources. Current solutions hinge on not only utilizing ultrareliable and low-latency communication (URLLC) links but also predicting future states, which may consume enormous communication resources and struggle with a short prediction time horizon. To fill this void, in this article we propose a novel two-way Koopman autoencoder (AE) approach wherein: 1) a sensing Koopman AE learns to understand the temporal state dynamics and predicts missing packets from a sensor to its remote controller and 2) a controlling Koopman AE learns to understand the temporal action dynamics and predicts missing packets from the controller to an actuator co-located with the sensor. Specifically, each Koopman AE aims to learn the Koopman operator in the hidden layers while the encoder of the AE aims to project the nonlinear dynamics onto a lifted subspace, which is reverted into the original nonlinear dynamics by the decoder of the AE. The Koopman operator describes the linearized temporal dynamics, enabling long-term future prediction and coping with missing packets and closed-form optimal control in the lifted subspace. Simulation results corroborate that the proposed approach achieves a 38X lower mean squared control error at 0-dBm signal-to-noise ratio (SNR) than the nonpredictive baseline.

In this work, hierarchical federated learning (HFL) over wireless multi-cell networks is proposed for large-scale model training while preserving data privacy. However, the imbalanced data distribution has a significant impact on the convergence rate and learning accuracy. In addition, a large learning latency is incurred due to the traffic load imbalance among base stations (BSs) and limited wireless resources. To cope with these challenges, we first provide an analysis of the model error and learning latency in wireless HFL. Then, joint user association and wireless resource allocation algorithms are investigated under independent identically distributed (IID) and non-IID training data, respectively. For the IID case, a learning latency aware strategy is designed to minimize the learning latency by optimizing user association and wireless resource allocation, where a mobile device selects the BS with the maximal uplink channel signal-to-noise ratio (SNR). For the non-IID case, the total data distribution distance and learning latency are jointly minimized to achieve the optimal user association and resource allocation. The results show that both data distribution and uplink channel SNR should be taken into consideration for user association in the non-IID case. Finally, the effectiveness of the proposed algorithms are demonstrated by the simulations.

In this paper, we investigate maximum-distance separable (MDS) codes and weighted graph based coded caching in fog radio access networks (F-RANs). In the placement phase, the redundant MDS based coded placement scheme is used to provide redundant coded packets and homogeneous cached contents. The redundant coded packets can be used to construct multicast opportunities for similar requests. In the delivery phase, the weighted graph based coded delivery scheme is conducted based on homogeneous cached contents, which can induce considerable multicast opportunities. By integrating the above two schemes, a joint MDS codes and weighted graph based coded caching policy is proposed to minimize the fronthaul load. Finally, we theoretically analyze the performance of the proposed policy by deriving the lower and upper bounds of the fronthaul load. Simulation results show that our proposed policy can provide 44% savings in the fronthaul load compared to the MDS-based uncoded delivery policy.

The emerging technology of reconfigurable intelligent surfaces (RISs) is provisioned as an enabler of smart wireless environments, offering a highly scalable, low-cost, hardware-efficient, and almost energy-neutral solution for dynamic control of the propagation of electromagnetic signals over the wireless medium, ultimately providing increased environmental intelligence for diverse operation objectives. One of the major challenges with the envisioned dense deployment of RISs in such reconfigurable radio environments is the efficient configuration of multiple metasurfaces with limited, or even the absence of, computing hardware. In this article, we consider multiuser and multi-RIS-empowered wireless systems and present a thorough survey of the online machine learning approaches for the orchestration of their various tunable components. Focusing on the sum-rate maximization as a representative design objective, we present a comprehensive problem formulation based on deep reinforcement learning (DRL). We detail the correspondences among the parameters of the wireless system and the DRL terminology, and devise generic algorithmic steps for the artificial neural network training and deployment while discussing their implementation details. Further practical considerations for multi-RIS-empowered wireless communications in the sixth-generation (6G) era are presented along with some key open research challenges. Different from the DRL-based status quo, we leverage the independence between the configuration of the system design parameters and the future states of the wireless environment, and present efficient multiarmed bandits approaches, whose resulting sum-rate performances are numerically shown to outperform random configurations, while being sufficiently close to the conventional deep \(Q\) network (DQN) algorithm, but with lower implementation complexity.

Inspired by the multi-head attention (MHA) mechanism in natural language processing, this letter proposes an iterative single-head attention (ISHA) mechanism for multi-UAV path planning. The ISHA mechanism is run by a communication helper collecting the state embeddings of UAVs and distributing an attention score vector to each UAV. The attention scores computed by ISHA identify how many interactions with other UAVs should be considered in each UAV’s control decision-making. Simulation results corroborate that the ISHA-based communication and control framework achieves faster travel with lower inter-UAV collision risks than an MHA-aided baseline, particularly under limited communication resources.

Owing to their fast convergence, second-order Newton-type learning methods have recently received attention in the federated learning (FL) setting. However, current solutions are based on communicating the Hessian matrices from the devices to the parameter server, at every iteration, incurring a large number of communication rounds; calling for novel communication-efficient Newton-type learning methods. In this article, we propose a novel second-order Newton-type method that, similarly to its first-order counterpart, requires every device to share only a model-sized vector at each iteration while hiding the gradient and Hessian information. In doing so, the proposed approach is significantly more communication-efficient and privacy-preserving. Furthermore, by leveraging the over-the-air aggregation principle, our method inherits privacy guarantees and obtains much higher communication efficiency gains. In particular, we formulate the problem of learning the inverse Hessian-gradient product as a quadratic problem that is solved in a distributed way. The framework alternates between updating the inverse Hessian-gradient product using a few alternating direction method of multipliers (ADMM) steps, and updating the global model using Newton’s method. Numerical results show that our proposed approach is more communication-efficient and scalable under noisy channels for different scenarios and across multiple datasets.

Wireless virtual reality (VR), a key 3GPP use case of emerging cellular systems, imposes new visual and haptic requirements directly linked to the Quality of Experience (QoE) of VR users. These QoE requirements can only be met by wireless connectivity that offers high-rate and high-reliability low-latency communications (HR2LLC), unlike the low rates commonly associated with ultrareliable low-latency communication. The high rates for VR over short distances can only be supported by an enormous bandwidth, available in the terahertz (THz)-frequency bands. To explore the potential of THz for meeting HR2LLC requirements, a quantification of the risk for an unreliable VR performance is conducted through a novel and rigorous characterization of the tail of the end-to-end (E2E) delay. Then, a thorough analysis of the Tail-Value-at-Risk (TVaR) is performed to concretely characterize the behavior of extreme wireless events crucial to the real-time VR experience. In particular, the probability distribution function of the THz transmission delay is derived and then used to infer the system reliability scenarios with guaranteed Line of Sight (LoS) as a function of THz network parameters. Numerical results show that abundant bandwidth and low molecular absorption are necessary to improve the reliability. However, their effect remains secondary compared to the availability of LoS, which significantly affects the THz HR2LLC performance. In particular, for scenarios with guaranteed LoS, a reliability of 99.999% (with an E2E delay threshold of 20 ms) for a bandwidth of 15 GHz along with data rates of 18.3 Gbps can be achieved by the THz network, compared to a reliability of 96% for twice the bandwidth, when blockages are considered.

Future 5G/6G wireless networks will be increasingly using millimeter waves (mmWaves), where fast and efficient beamforming is vital for providing continuous service to highly mobile devices in the presence of interference and signal attenuation, manifested by blockage. In this paper, we propose a novel and efficient method for mmWave beamforming in massive multiple-input multiple-output (MIMO) systems to achieve the aforementioned goals with low complexity in such scenarios. In doing so, we utilize deep reinforcement learning (DRL) to maximize the network’s energy efficiency subject to the quality of service (QoS) constraint for each user equipment (UE) and obtain its hybrid beamforming matrices. In doing so, we assume each UE is simultaneously associated with multiple access points (APs), i.e., simultaneous beamforming to/from multiple APs (coordinated multipoints) is needed for each UE. We also propose a low-complexity training algorithm, based on approximate message passing, which is well suited for the network edge. Besides, we develop a distributed scheme to reduce communications overhead via federated DRL. Extensive simulations show significant performance improvement over existing methods.

In this paper, a novel framework is proposed to optimize the downlink multi-user communication of a millimeter wave base station, which is assisted by a reconfigurable intelligent reflector (IR). In particular, a channel estimation approach is developed to measure the channel state information (CSI) in real-time. First, for a perfect CSI scenario, the precoding transmission of the BS and the reflection coefficient of the IR are jointly optimized, via an iterative approach, so as to maximize the sum of downlink rates towards multiple users. Next, in the imperfect CSI scenario, a distributional reinforcement learning (DRL) approach is proposed to learn the optimal IR reflection and maximize the expectation of downlink capacity. In order to model the transmission rate’s probability distribution, a learning algorithm, based on quantile regression (QR), is developed, and the proposed QR-DRL method is proved to converge to a stable distribution of downlink transmission rate. Simulation results show that, in the error-free CSI scenario, the proposed approach yields over 30% and 2-fold increase in the downlink sum-rate, compared with a fixed IR reflection scheme and direct transmission scheme, respectively. Simulation results also show that by deploying more IR elements, the downlink sum-rate can be significantly improved. However, as the number of IR components increases, more time is required for channel estimation, and the slope of increase in the IR-aided transmission rate will become smaller. Furthermore, under limited knowledge of CSI, simulation results show that the proposed QR-DRL method, which learns a full distribution of the downlink rate, yields a better prediction accuracy and improves the downlink rate by 10% for online deployments, compared with a Q-learning baseline.

In this paper, the content popularity prediction problem in fog radio access networks (F-RANs) is investigated. In order to obtain accurate prediction with low complexity, we propose a novel context-aware popularity prediction policy based on federated learning (FL). Firstly, user preference learning is applied by considering that users prefer to request the contents they are interested in. Then, users’ context information is utilized to cluster users efficiently by adaptive context space partitioning. After that, we formulate a popularity prediction optimization problem to learn the local model parameters by using the stochastic variance reduced gradient (SVRG) algorithm. Finally, FL based model integration is proposed to learn the global popularity prediction model based on local models using the distributed approximate Newton (DANE) algorithm with SVRG. Our proposed popularity prediction policy not only can predict content popularity accurately, but also can significantly reduce computational complexity. Moreover, we theoretically analyze the convergence bound of our proposed FL based model integration algorithm. Simulation results show that our proposed policy increases the cache hit rate by up to 21.5 % compared to existing policies.