Network architectures assisted by Generative Artificial Intelligence (GAI) are envisioned as foundational elements of sixth-generation (6G) communication system. To deliver ubiquitous intelligent services and meet diverse service requirements, 6G network architecture should offer personalized services to various mobile devices. Federated learning (FL) with personalized local training, as a privacy-preserving machine learning (ML) approach, can be applied to address these challenges. In this paper, we propose a meta-learning-based personalized FL (PFL) method that improves both communication and computation efficiency by utilizing over-the-air computations. Its "pretraining-and-fine-tuning" principle makes it particularly suitable for enabling edge nodes to access personalized GAI services while preserving local privacy. Experiment results demonstrate the outperformance and efficacy of the proposed algorithm, and notably indicate enhanced communication efficiency without compromising accuracy.

This paper presents an algorithm named the dependency-aware offloading framework (DeAOff), which is designed to optimize the deployment of Gen-AI decoder models in mobile edge computing (MEC) environments. These models, such as decoders, pose significant challenges due to their inter-layer dependencies and high computational demands, especially under edge resource constraints. To address these challenges, we propose a two-phase optimization algorithm that first handles dependency-aware task allocation and subsequently optimizes energy consumption. By modeling the inference process using directed acyclic graphs (DAGs) and applying constraint relaxation techniques, our approach effectively reduces execution latency and energy usage. Experimental results demonstrate that our method achieves a reduction of up to 20% in task completion time and approximately 30% savings in energy consumption compared to traditional methods. These outcomes underscore our solution's robustness in managing complex sequential dependencies and dynamic MEC conditions, enhancing quality of service. Thus, our work presents a practical and efficient resource optimization strategy for deploying models in resource-constrained MEC scenarios.

The advent of the internet-of-everything era has led to the increased use of mobile edge computing. The rise of artificial intelligence has provided many possibilities for the low-latency task-offloading demands of users, but existing technologies rigidly assume that there is only one task to be offloaded in each time slot at the terminal. In practical scenarios, there are often numerous computing tasks to be executed at the terminal, leading to a cumulative delay for subsequent task offloading. Therefore, the efficient processing of multiple computing tasks on the terminal has become highly challenging. To address the low-latency offloading requirements for multiple computational tasks on terminal devices, we propose a terminal multitask parallel offloading algorithm based on deep reinforcement learning. Specifically, we first establish a mobile edge computing system model consisting of a single edge server and multiple terminal users. We then model the task offloading decision problem as a Markov decision process, and solve this problem using the Dueling Deep-Q Network algorithm to obtain the optimal offloading strategy. Experimental results demonstrate that, under the same constraints, our proposed algorithm reduces the average system latency.

Federated semi-supervised learning (FSSL) faces two major challenges: the scarcity of labeled data across clients and the non-independent and identically distributed (Non-IID) nature of data among clients. To address these issues, we propose diffusion model-based data synthesis aided FSSL (DDSA-FSSL), a novel approach that leverages diffusion model (DM) to generate synthetic data, thereby bridging the gap between heterogeneous local data distributions and the global data distribution. In the proposed DDSA-FSSL, each client addresses the scarcity of labeled data by utilizing a federated learning-trained classifier to perform pseudo labeling for unlabeled data. The DM is then collaboratively trained using both labeled and precision-optimized pseudo-labeled data, enabling clients to generate synthetic samples for classes that are absent in their labeled datasets. As a result, the disparity between local and global distributions is reduced and clients can create enriched synthetic datasets that better align with the global data distribution. Extensive experiments on various datasets and Non-IID scenarios demonstrate the effectiveness of DDSA-FSSL, achieving significant performance improvements, such as increasing accuracy from 38.46% to 52.14% on CIFAR-10 datasets with 10% labeled data.

The intelligent operation management of distribution services is crucial for the stability of power systems. Integrating the large language model (LLM) with 6G edge intelligence provides customized management solutions. However, the adverse effects of false data injection (FDI) attacks on the performance of LLMs cannot be overlooked. Therefore, we propose an FDI attack detection and LLM-assisted resource allocation algorithm for 6G edge intelligence-empowered distribution power grids. First, we formulate a resource allocation optimization problem. The objective is to minimize the weighted sum of the global loss function and total LLM fine-tuning delay under constraints of long-term privacy entropy and energy consumption. Then, we decouple it based on virtual queues. We utilize an LLM-assisted deep Q network (DQN) to learn the resource allocation strategy and design an FDI attack detection mechanism to ensure that fine-tuning remains on the correct path. Simulations demonstrate that the proposed algorithm has excellent performance in convergence, delay, and security.

With the rapid development of network technologies, a large number of deployed edge devices and information systems generate massive amounts of data which provide good support for the advancement of data-driven intelligent models. However, these data often contain sensitive information of users. Federated learning (FL), as a privacy preservation machine learning setting, allows users to obtain a well-trained model without sending the privacy-sensitive local data to the central server. Despite the promising prospect of FL, several significant research challenges need to be addressed before widespread deployment, including network resource allocation, model security, model convergence, etc. In this paper, we first provide a brief survey on some of these works that have been done on FL and discuss the motivations of the Communication Networks (CNs) and FL to mutually enable each other. We analyze the support of network technologies for FL, which requires frequent communication and emphasizes security, as well as the studies on the intelligence of many network scenarios and the improvement of network performance and security by the methods based on FL. At last, some challenges and broader perspectives are explored.

Wireless networks support numerous terminals, manage large data volumes, and provide diverse services, but the vulnerability to environmental changes leads to increased complexity and costs. Situational awareness has been widely applied in network management, but existing methods fail to find optimal solutions due to the high heterogeneity of base stations, numerous metrics, and complex intercell dependencies. To address this gap, this paper proposes a specialized framework for wireless networks, integrating an evaluation model and control approach. The framework expands the indicator set into four key areas, introduces an evaluation method, and proposes the indicator perturbation greedy (IPG) algorithm and the adjustment scheme selection method based on damping coefficient (DCSS) for effective network optimization. A case study in an urban area demonstrates the framework's ability to balance and improve network performance, enhancing situational awareness and operational efficiency under dynamic conditions.

This paper investigates the reconfigurable intelligent surface (RIS)-aided MIMO covert communications in high-speed railway (HSR) scenario. In the scenario, RIS controls the phases of reflection elements dynamically to send the signal in the desired direction, which facilitates the covert communication between base station (BS) and train mobile relay (MR) in the existence of a watchful warden (Willie). To protect the desired transmission, it is assumed that MR sends out jamming signals with a varying power to confuse the Willie. Considering the Doppler spread caused by the time-varying wireless channel, the joint optimization problem of the BS beamforming matrix, MR beamforming matrix, and the RIS phase shifts is established to maximize the covert throughput. An alternating optimization (AO) method for handling non convex problems is proposed based on coupling effects and the constraints of constant modulus, and a semidefinite relaxation method is provided. Finally, we achieve the optimal solutions of the multi-objective optimization problem by interior-point method. The simulation results demonstrate that the proposed algorithm exhibits the superior robustness and covert performances in high-speed railway scenarios.

The beyond fifth-generation Internet of Things requires more capable channel coding schemes to achieve high-reliability, low-complexity and low-latency communications. The theoretical analysis of error-correction performance of channel coding functions as a significant way of optimizing the transmission reliability and efficiency. In this paper, the efficient estimation methods of the block error rate (BLER) performance for rate-compatible polar codes (RCPC) are proposed under several scenarios. Firstly, the BLER performance of RCPC is generally evaluated in the additive white Gaussian noise channels. That is further extended into the Rayleigh fading channel case using an equivalent estimation method. Moreover, with respect to the powerful decoder such as successive cancellation list decoding, the performance estimation is derived analytically based on the polar weight spectrum and BLER upper bounds. Theoretical evaluation and numerical simulation results show that the estimated performance can fit well the practical simulated results of RCPC under the objective conditions, verifying the validity of our proposed performance estimation methods. Furthermore, the application designs of the reliability estimation of RCPC are explored, particularly in the advantages of the signal-to-noise (SNR) estimation and throughput efficiency optimization of polar coded hybrid automatic repeat request.

In this paper, a novel Filter bank multicarrier/quadrature amplitude modulation (FBMC/QAM) scheme which separates the real and imaginary part of the subcarriers in a multipath time-varying fading channel is put forward and analyzed in detail. By applying the methods of mapping the time-domain symbols and reducing the correlation of frequency-domain symbols, the presented scheme can eliminate the intrinsic imaginary interference more thoroughly and greatly mitigate residual interference as well as receive a good peak-to-average power ratio (PAPR) mitigation effect. Theoretical analysis and simulation results indicate that compared with the existing schemes with the methods of pre-coding (FBMC/QAM-CC) and iterative interference cancellation (FBMC/QAM-IIC), our adopted scheme can not only obtain better performances on bit error rate (BER) and out-of-band (OOB) emission with no loss of transmission efficiency, but also achieve a good PAPR mitigation effect with a small increase in complexity.

With the rapid development of wireless techniques, the bandpass filter (BPF) is required to cover microwave and millimeter-wave frequency bands simultaneously with good mid-band suppression. However, it is difficult to implement such BPF due to the large frequency ratio and wideband rejection. This paper presents a superior method to realize a dual-band BPF with a large frequency ratio maintaining compact size and low design complexity. This is contributed by an ultra-wide stopband BPF with inherent discriminating excited degree at spurious frequencies. By properly arranging the feeding position and electrical length ratio of stepped impedance resonator (SIR), the excited degree at specific spurious frequencies can be flexibly adjusted to achieve desired suppression level without affecting characteristics at the fundamental passband. For validation, two BPFs were simulated, fabricated and measured, exhibiting suppression levels of 20.3 dB and 35 dB up to 18$f_0$ and 10.53$f_0$ respectively. Based on this, a dual-band BPF with a large frequency ratio can be easily constructed. For demonstration, a dual-band BPF operating at 3.55 GHz and 43.15 GHz is implemented. A frequency ratio up to 12.15 and mid-band suppression level better than 28 dB had been achieved. Advantages of compactness, simplicity and excellent performance of the proposed work can be observed.

To achieve a low-complexity nonlinearity compensation (NLC) in high-symbol-rate (HSR) systems, we propose a modified weighted digital backpropagation (M-W-DBP) by jointly shifting the calculated position of nonlinear phase noise and considering the correlation of neighboring symbols in the NLC section of DBP. Based on this model, with the aid of neural network optimization, a learned version of M-W-DBP (M-W-LDBP) is also proposed and explored. Furthermore, enough technical details are revealed for the first time, including the principle of our proposed M-W-DBP and M-W-LDBP, the training process, and the complexity analysis of different DBP-class NLC algorithms. Evaluated numerically with QPSK, 16QAM, and PS-64QAM modulation formats, 1-step-per-span (1-StPS) M-W-DBP/LDBP achieves up to 1.29/1.49 dB and 0.63/0.74 dB signal-to-noise ratio improvement compared to chromatic dispersion compensation (CDC) in 90-GBaud and 128-GBaud 1000-km single-channel transmission systems, respectively. Moreover, 1-StPS M-W-DBP/LDBP provides a more powerful NLC ability than 2-StPS LDBP but only needs about 60% of the complexity. The effectiveness of the proposed M-W-DBP and M-W-LDBP in the presence of laser phase noise is also verified and the necessity of using the learned version of M-W-DBP is also discussed. This work is a comprehensive study of M-W-DBP/LDBP and other DBP-class NLC algorithms in HSR systems.

Existing orthogonal space-time block coding (OSTBC) schemes for backscatter communication systems cannot achieve a full transmission code rate when the tag is equipped with more than two antennas. In this paper, we propose a quasi-orthogonal space-time block code (QOSTBC) that can achieve a full transmission code rate for backscatter communication systems with a four-antenna tag and then extend the scheme to support tags with $2^i$ antennas. Specifically, we first present the system model for the backscatter system. Next, we propose the QOSTBC scheme to encode the tag signals. Then, we provide the corresponding maximum likelihood detection algorithms to recover the tag signals. Finally, simulation results are provided to demonstrate that our proposed QOSTBC scheme and the detection algorithm can achieve a better transmission code rate or symbol error rate performance for backscatter communication systems compared with benchmark schemes.

In this paper, we investigate the performance of physical layer security (PLS) over reconfigurable intelligent surfaces (RIS)-aided wireless communication systems, where all fading channels are modeled with Fisher-Snedecor $\mathcal{F}$ distribution. Specifically, we consider a RIS with $N$ reflecting elements between the transmitter and the legitimate receiver to develop a smart environment and also meliorate secure communications. In this regard, we derive the closed-form expressions for the secrecy outage probability (SOP) and average secrecy capacity (ASC). We also analyze the asymptotic behaviour of the SOP and ASC by exploiting the residue approach. Monte-Carlo (MC) simulation results are provided throughout to validate the correctness of the developed analytical results, showing that considering RIS in wireless communication systems has constructive effects on the secrecy performance.

Blockchain-based user-centric access network (UCAN) fails in dynamic access point (AP) management, as it lacks an incentive mechanism to promote virtuous behavior. Furthermore, the low throughput of the blockchain has been a bottleneck to the widespread adoption of UCAN in 6G. In this paper, we propose Overlap Shard, a blockchain framework based on a novel reputation voting (RV) scheme, to dynamically manage the APs in UCAN. AP nodes in UCAN are distributed across multiple shards based on the RV scheme. That is, nodes with good reputation (virtuous behavior) are likely to be selected in the overlap shard. The RV mechanism ensures the security of UCAN because most APs adopt virtuous behaviors. Furthermore, to improve the efficiency of the Overlap Shard, we reduce cross-shard transactions by introducing core nodes. Specifically, a few nodes are overlapped in different shards, which can directly process the transactions in two shards instead of cross-shard transactions. This greatly increases the speed of transactions between shards and thus the throughput of the overlap shard. The experiments show that the throughput of the overlap shard is about 2.5 times that of the non-sharded blockchain.

Software-defined satellite networks (SDSNs) play an essential role in future networks. Due to the diverse service scenarios, SDSN faces the demand of packet processing for heterogeneous protocols. Existing packet switching typically works on one single protocol. For protocol-heterogeneous users, existing packet switch architectures have to construct multiple protocol-specific switching instances, resulting in severe resource waste. In this article, we propose the heterogeneous protocol-independent packet switch architecture (HISA). HISA employs a fast parsing structure to achieve efficient heterogeneous packet parsing and a novel match-action pipeline to achieve shared packet processing among heterogeneous users. HISA can also support the online configuration of switching behaviors. Use cases illustrate the effectiveness of applying HISA in SDSN. Numerical results show that compared to existing packet switching, HISA can significantly improve the resource utilization of SDSN.

To realize dynamic statistical publishing and protection of location-based data privacy, this paper proposes a differential privacy publishing algorithm based on adaptive sampling and grid clustering and adjustment. The PID control strategy is combined with the difference in data variation to realize the dynamic adjustment of the data publishing intervals. The spatial-temporal correlations of the adjacent snapshots are utilized to design the grid clustering and adjustment algorithm, which facilitates saving the execution time of the publishing process. The budget distribution and budget absorption strategies are improved to form the sliding window-based differential privacy statistical publishing algorithm, which realizes continuous statistical publishing and privacy protection and improves the accuracy of published data. Experiments and analysis on large datasets of actual locations show that the privacy protection algorithm proposed in this paper is superior to other existing algorithms in terms of the accuracy of adaptive sampling time, the availability of published data, and the execution efficiency of data publishing methods.

In the field of specific emitter identification (SEI), power amplifiers (PAs) have long been recognized as significant contributors to unintentional modulation characteristics. To enhance signal quality, digital pre-distortion (DPD) techniques are commonly employed in practical applications to mitigate the nonlinear effects of PAs. However, DPD techniques may diminish the distinctive characteristics of individual transmitters, potentially compromising SEI performance. This study investigates the influence of SEI in the presence of DPD applied to PAs. We construct a semi-physical emitter platform using AD9361 and ZYNQ, incorporating memory and non-memory models to emulate an amplification system comprising DPD devices and PAs. Furthermore, we delve into the analysis and evaluation of LMS-based and QRD-RLS-based DPD algorithms to ascertain their efficacy in compensating for amplifier nonlinearity. Finally, we conduct a comprehensive set of experiments to demonstrate the adverse impact of DPD techniques on SEI. Our findings demonstrate a direct correlation between the degree of DPD performance and its impact magnitude on SEI, thereby providing a foundational basis for future studies investigating SEI techniques under DPD.

The COVID-19 pandemic, which was declared by the WHO, had created a global health crisis and disrupted people's daily lives. A large number of people were affected by the COVID-19 pandemic. Therefore, a diagnostic model needs to be generated which can effectively classify the COVID and non-COVID cases. In this work, our aim is to develop a diagnostic model based on deep features using effectiveness of Chest X-ray (CXR) in distinguishing COVID from non-COVID cases. The proposed diagnostic framework utilizes CXR to diagnose COVID-19 and includes Grad-CAM visualizations for a visual interpretation of predicted images. The model's performance was evaluated using various metrics, including accuracy, precision, recall, F1-score, and G-mean. Several machine learning models, such as random forest, dense neural network, SVM, twin SVM, extreme learning machine, random vector functional link, and kernel ridge regression, were selected to diagnose COVID-19 cases. Transfer learning was used to extract deep features. For feature extraction many CNN-based models such as Inception V3, MobileNet, ResNet50, VGG16 and Xception models are used. It was evident from the experiments that ResNet50 architecture outperformed all other CNN architectures based on AUC. The TWSVM classifier achieved the highest AUC score of 0.98 based on the ResNet50 feature vector.

The increase in population and vehicles exacerbates traffic congestion and management difficulties. Therefore, achieving accurate and efficient traffic flow prediction is crucial for urban transportation. For that reason, we propose a graph federated learning-based digital twin traffic flow prediction method (GFLDT) by integrating the benefits of collaborative intelligence and computation of intelligent IoT. Specifically, we construct a digital twin network for predicting traffic flow, which is divided into client twin and global twin. Based on this, we adopt the concept of graph federated learning to learn the temporal dependence of traffic flow using local data from client twins, and the spatial dependence of traffic flow using global information from global twins. In addition, we validate on a real traffic dataset, and the results show that through collaborative training of the client twins and the global twins , GFLDT achieves accurate traffic flow prediction while protecting data security.