Robust Federated Learning (RoFL) extends traditional federated learning, not only by enabling multiple clients to collaboratively train a shared model under the coordination of an edge server, but also by incorporating client-side defense mechanisms (e.g., adversarial training) to defend against adversarial attacks while preserving data privacy. However, recent studies have shown that RoFL also remains vulnerable to the challenges posed by non-independent and identically distributed (non-IID) data distributions across heterogeneous clients, which can degrade overall model generalization and robustness. To mitigate this challenge, in this paper, we propose a novel RoFL framework, called RoFLCCA, to address non-IID challenges while defending against adversarial attacks. In particular, we first introduce a local classifier calibration mechanism that utilizes feature-level augmentation to mitigate the effects of non-IID data. By incorporating global class-wise feature statistics, each client can adjust its classifier using synthetic features derived from these shared representations. Second, we propose a calibrated classifier-guided global adversarial alignment strategy, which enforces consistency between augmented and adversarial predictions to improve robustness. Simulation results demonstrate the effectiveness of the proposed RoFLCCA, which consistently outperforms existing robust federated baselines across different datasets and settings. On average, it achieves a 7.07% improvement in clean accuracy and a 4.71% gain in adversarial robustness, highlighting its ability to enhance both generalization and defense against adversarial threats.

This paper proposes an AI framework that leverages integrated sensing and communication (ISAC), aided by the age of sensing (AoS) to ensure the timely location updates of the users for a holographic MIMO (HMIMO)-assisted base station (BS)-enabled wireless network. The AI-driven framework aims to achieve optimized power allocation for efficient beamforming by activating the minimal number of grids from the HMIMO BS for serving the users. An optimization problem is formulated to maximize the sensing utility function, aiming to maximize the communication signal-to-interference-plus-noise ratio (SINRc) of the received signals and beam-pattern gains to improve the sensing SINR of reflected echo signals, which in turn maximizes the achievable rate of users. A novel AI-driven framework is presented to tackle the formulated NP-hard problem that divides it into two problems: a sensing problem and a power allocation problem. The sensing problem is solved by employing a variational autoencoder (VAE)-based mechanism that obtains the sensing information leveraging AoS, which is used for the location update. Subsequently, a deep deterministic policy gradient-based deep reinforcement learning scheme is devised to allocate the desired power by activating the required grids based on the sensing information achieved with the VAE-based mechanism. Simulation results demonstrate the superior performance of the proposed AI framework compared to advantage actor-critic and deep Q-network-based methods, achieving a cumulative average SINRcimprovement of 8.5 dB and 10.27 dB, and a cumulative average achievable rate improvement of 21.59 bps/Hz and 4.22 bps/Hz, respectively. Therefore, our proposed AI-driven framework guarantees efficient power allocation for holographic beamforming through ISAC schemes leveraging AoS.

Federated learning (FL) enables privacy-preserving model training across decentralized data. However, significant data heterogeneity, common in domains like the Internet of Things (IoT), hinders generalization. Federated Domain Generalization (FDG) extends the FL paradigm by aiming to train models that generalize effectively to unseen domains, without requiring access to data from those domains during training. Current FDG methods primarily use ResNet backbones pre-trained on ImageNet-1K, limiting adaptability due to architectural constraints and limited pre-training diversity. This reliance has created a gap in leveraging advanced architectures and diverse pre-training datasets to address these challenges. To bridge this gap, we present the first comprehensive investigation into the efficacy of advanced pre-trained architectures such as Vision Transformers, ConvNeXt, and Swin Transformers, in enhancing FDG performance. Unlike ResNet, these architectures capture global context and long-range dependencies, making them well-suited for FDG. Beyond architectural evaluation, we systematically assess the impact of diverse pre-training datasets and compare self-supervised and supervised strategies. Our analysis rigorously investigates the influence of architectural depth, parameter efficiency, and the interplay between diverse model families and dataset characteristics on FDG performance. We find that advanced architectures pre-trained on large datasets significantly outperform ResNet models. Specifically, ConvNeXt architectures outperform all other candidates. We find self-supervised methods using masked image patch reconstruction via discrete token prediction outperform their supervised counterparts. We observe that certain advanced model variants with fewer parameters outperform larger ResNet models. This underscores the need for advanced architectures and scalable pretraining to enable efficient and generalizable FDG.

Federated learning (FL) makes it possible to advance towards edge intelligence by enabling collaborative and privacy-preserving model training across distributed edge devices. One of the main challenges in FL is non-IID (not Independent and Identically Distributed) nature of data distribution across edge devices, which results in inconsistent update directions of local and global models, thus hindering model convergence. Moreover, recent studies have shown that FL models can significantly degrade performance under adversarial attacks, which further poses challenges for deployment at edge sides. In this work, we attempt to improve the robustness of FL model under adversarial attacks in non-IID settings by sharing knowledge between a central server and edge devices via knowledge distillation. Specifically, we propose a new knowledge distillation-based federated adversarial training (FAT) framework, termed FedAdv (Federated Adversarial), which involves an edge server collecting global prototypes by aggregating local prototypes obtained from participating devices after adversarial training (AT). These global prototypes are subsequently distributed to the edge devices for regularization. This regularization mechanism aims to encourage each device to align its local representation with the corresponding global prototype. By doing so, it helps prevent significant deviations of local model updates from the global model. Experimental results on MNIST and Fashion-MNIST show that our strategy yields comparable or superior performance gains in both natural and robust accuracy compared to several baselines.

The 6G wireless communication networks need an intelligent networking system to meet the ever-increasing de-mands of various applications and mobile devices to ensure power savings, energy efficiency (EE), high integration of devices, and mass connection. To achieve these aims, an artificial intelligence (AI)-based holographic MIMO (HMIMO)-aided cell-free (CF) network is suggested to allocate desired power for beamforming by activating the required number of grids from the serving HMIMOs for serving the users. An optimization problem is developed to ensure effective power allocation that maximizes the EE of the system. A Transformer-based AI framework is proposed to solve the formulated NP-hard problem that distributes desired power for serving the users by activating the required number of grids from the required number of serving HMIMOs in the CF network. Finally, simulation results represent that the proposed power allocation framework outperforms the gated recurrent unit and long short-term memory-based mechanisms, achieving a combined power savings of 12.5% and 4.06%, and a combined EE improvement of 14.68% and 8.93%, correspondingly. Therefore, our suggested AI-based framework guarantees effective power allocation for beamforming to serve the users.

Federated learning (FL) is a privacy-preserving distributed management framework based on collaborative model training of distributed devices in edge networks. However, recent studies have shown that FL is vulnerable to adversarial examples (AEs), leading to a significant drop in its performance. Meanwhile, the non-independent and identically distributed (non-IID) challenge of data distribution between edge devices can further degrade the performance of models. Consequently, both AEs and non-IID pose challenges to deploying robust learning models at the edge. In this work, we adopt the adversarial training (AT) framework to improve the robustness of FL models against adversarial example (AE) attacks, which can be termed as federated adversarial training (FAT). Moreover, we address the non-IID challenge by implementing a simple yet effective logits calibration strategy under the FAT framework, which can enhance the robustness of models when subjected to adversarial attacks. Specifically, we employ a direct strategy to adjust the logits output by assigning higher weights to classes with small samples during training. This approach effectively tackles the class imbalance in the training data, with the goal of mitigating biases between local and global models. Experimental results on three dataset benchmarks, MNIST, Fashion-MNIST, and CIFAR-10 show that our strategy achieves competitive results in natural and robust accuracy compared to several baselines.

The upcoming 6G wireless communication networks are anticipated to offer extensive mobile connectivity, faster data services with reduced power consumption, and seamless integration among different technologies for providing effective beamforming. To accomplish these aims, a transfer learning empowered AI framework is proposed to allocate the power for serving the users under the coverage areas of the corresponding holographic MIMOs (HMIMOs) by activating the required number of grids from the respective HMIMOs. An optimization problem is formulated with the goal of maximizing the utility function for achievable rate, which in turn maximizes the signal-to-interference-plus-noise ratio (SINR), and achievable rate of the users. The HMIMO that serves the highest number of users is considered as the parent HMIMO and the rest of the HMIMOs are regarded as the child HMIMOs. A Transformer-based AI framework is utilized for allocating the power to the users under the coverage areas of the parent HMIMO and transfers the knowledge of the trained model to the child HMIMOs which requires lower learning cost to allocate power to the corresponding users within the coverage areas of the child HMIMOs. Finally, simulation results show that the proposed AI framework empowered by transfer learning surpasses the baseline methods such as gated recurrent unit and long short-term memory, achieving power savings ranging from 28.14% to 38.92% and achievable rate enhancements from 16.58 bps/Hz to 16.84 bps/Hz.

Federated learning-assisted edge intelligence enables privacy protection in modern intelligent services. However, not independent and identically distributed (non-IID) distribution among edge clients can impair the local model performance. The existing single prototype-based strategy represents a class by using the mean of the feature space. However, feature spaces are usually not clustered, and a single prototype may not represent a class well. Motivated by this, this article proposes a multiprototype federated contrastive learning approach (MP-FedCL) which demonstrates the effectiveness of using a multiprototype strategy over a single-prototype under non-IID settings, including both label and feature skewness. Specifically, a multiprototype computation strategy based on k-means is first proposed to capture different embedding representations for each class space, using multiple prototypes$(k$centroids) to represent a class in the embedding space. In each global round, the computed multiple prototypes and their respective model parameters are sent to the edge server for aggregation into a global prototype pool, which is then sent back to all clients to guide their local training. Finally, local training for each client minimizes their own supervised learning tasks and learns from shared prototypes in the global prototype pool through supervised contrastive learning, which encourages them to learn knowledge related to their own class from others and reduces the absorption of unrelated knowledge in each global iteration. Experimental results on MNIST, Digit-5, Office-10, and DomainNet show that our method outperforms multiple baselines, with an average test accuracy improvement of about 4.6% and 10.4% under feature and label non-IID distributions, respectively.

Sixth-generation wireless networks are required to satisfy the ever-increasing demands of diverse applications to guarantee power savings, energy efficiency (EE), and mass connectivity. To accomplish these goals, in this article, an artificial intelligence (AI)-based holographic MIMO (HMIMO)-empowered cell-free (CF) network is proposed while leveraging integrated sensing and communication (ISAC). The proposed AI-based framework allocates the desired power for beamforming by activating the required number of grids from the serving HMIMO base stations (BSs) in the CF network to serve the users. An optimization problem is formulated that maximizes the sensing utility function, which in turn maximizes the signal-to-interference-plus-noise ratio (SINR) of the received signal, the sensing SINR of the reflected echo signal, and EE, ensuring efficient power allocation. To solve the optimization problem, an AI-based framework is proposed to enable a decomposition of the NP-hard problem into two subproblems: 1) a sensing subproblem and 2) a power allocation subproblem. Initially, a variational autoencoder (VAE)-based scheme is utilized to solve the sensing subproblem that identifies the current location of the users with the sensing information. Then, a transformer-based mechanism is devised to allocate the desired power to users by activating the required grids from the serving HMIMO BSs in the CF network based on the sensing information achieved with the VAE-based scheme. Simulation results demonstrate that the proposed AI-based framework outperforms the long short-term memory and gated recurrent unit-based mechanisms, with cumulative power savings of 8.64% and 16.02%, and cumulative EE of 14.49% and 16.61%, accordingly, considering the ground truth values.

The impending sixth-generation wireless communication networks are anticipated to guarantee mass connectivity, high integration, and lower power consumption for generating the required beamforming. To achieve these goals, an artificial intelligence (AI) framework is proposed by utilizing holographic MIMO-assisted integrated sensing, localization, and communication. The proposed AI framework ensures lower power consumption to activate the minimum number of grids from the holographic grid array for the generation of holographic beamforming. An optimization problem is formulated to maximize the signal-to-interference-plus-noise ratio received by the users, which in turn maximizes the utility function for sensing considering the user distances, beampattern gains, sensing-communication loss, and dense locations controlling parameter. A novel AI-based framework is proposed to solve the formulated NP-hard optimization problem by decomposing it into two subproblems: the sensing problem and the communication resource allocation problem. First, a variational autoencoder (VAE) based mechanism is devised to solve the sensing problem mitigating the disputes to obtain the users’ exact location. Second, a sequential neural network-based scheme is utilized to allocate the communication resources to the heterogeneous users for generating the desired beamforming based on the findings of the VAE-based mechanism. Moreover, an extreme case power allocation strategy is presented once a large number of users enter the system. The extreme case power allocation strategy applies when the total power prediction exceeds the total system power for allocating the communication resources to the users. Finally, simulation results validate that the proposed AI-based framework outperforms the long short-term memory method with a cumulative power savings of 34.02% taking the ground truth power into account. Therefore, the proposed AI framework generates effective beamforming to serve the communication users.

The future sixth-generation (6G) wireless communication networks are expected to provide massive connectivity with lower power requirements for generating the desired beamforming. Therefore, holographic MIMO assisted integrated sensing and communication framework is proposed that ensures lower power requirements to activate the minimum number of grids from the holographic grid array (HGA) for the effective beamforming. An optimization problem is formulated that maximizes the signal to noise-interference ratio (SNIR) of the users which in turn maximizes the utility function for sensing (UFS) considering the beampattern gains, distances, and sensing-communication loss. A novel artificial intelligence (AI) framework is proposed to solve the formulated problem which is a NP-hard problem. First, a variational autoencoder (VAE) based scheme is developed to solve the challenges of determining the exact location of the users and complete data distribution. Then, a sequential neural network-based mechanism is devised to allocate the communication resources to the heterogeneous users for the desired beamforming based on the results obtained from VAE. Finally, simulation results demonstrate that the proposed algorithms confirm 23% power savings compared to long short-term memory (LSTM) method to perform effective beamforming for serving the users.

Federated learning (FL) allows local clients to train a global model by cooperating with a server while ensuring that their raw data is not revealed. However, most existing works usually choose clients randomly, regardless of their capabilities and contributions to training. Additionally, FL client selection mechanisms concentrate on a significant challenge associated with system or statistical heterogeneity. This paper tries to manage both the system and statistical heterogeneity of distributed clients in the networks. First, to manage the system heterogeneity, an optimization objective is first proposed to maximize the number of clients with similar capabilities such as storage, computational, and communication capabilities. Then, a network framework with a logical layer is proposed to logically group similar clients by checking their capabilities. Finally, to manage the statistical heterogeneity among clients, a novel Contribution-based Dynamic Federated training strategy, called CDFed, is designed to dynamically adjust the probability of clients being chosen based on Shapley values in each global round. Experimental results on two baseline datasets: MNIST and FMNIST, demonstrate that our proposal has a faster convergence rate, about 50%, and a higher average test accuracy, at least 1%, than baselines in most cases.

In cognitive radio sensor networks (CRSNs), the sensor devices which are enabled to perform dynamic spectrum access have to frequently sense the licensed channel to find idle channels. The behavior of spectrum sensing will consume a lot of battery power of sensor devices and reduce the network lifetime. In this paper, we aim to answer the question of how many spectrum sensing nodes (SSNs) are required. In order to achieve this, SSN ratio effects on the accuracy of spectrum sensing from the perspective of network energy efficiency are analyzed first. Based on these analyses, the optimal SSN ratio is derived for maximizing the network lifetime by optimizing the cooperative detection probability (CDP). Simulation results show that the optimal SSN ratio can guarantee the spectrum sensing performance in terms of detection and false alarm probabilities and effectively extend the network lifetime.