Federated graph learning (FGL) empowers distributed training of subgraphs across multiple institutions, overcoming the challenges of data silos and inter-institutional data sharing. Existing federated subgraph methods achieve collaborative training by training local models on clients and uploading them to a server for parameter aggregation. However, most existing methods do not take into account for the heterogeneity of client subgraph data, leading to suboptimal results. To deal with this challenging problem, we propose a personalized FGL method named FedPGT. FedPGT employs a single-layer transformer to capture long-term dependencies between client nodes, and pools client subgraphs into an averaged node representation for subgraph similarity calculation. We validate the effectiveness of FedPGT through extensive experiments on two subgraph scenarios and five datasets. The results demonstrate that FedPGT significantly outperforms the baseline methods and mitigates the adverse impact of heterogeneity.

With the advances of deep learning, efficient model inference is crucial. Deep learning compilers optimize inference by decomposing models into subgraphs and searching schedules for them, whose evaluation relies on accurate cost models. Existing methods suffer from high transformation overheads or limited prediction accuracy caused by insufficient structural representation of subgraphs and schedules. To address these limitations, we propose CAST, a framework that predicts schedule performance based on Abstract Syntax Trees (ASTs). CAST proposes AST classification based on structural similarity and class-specific cost models. Experiments show CAST achieves significantly reduced prediction errors and up to$13 \times$higher efficiency than prior methods.

With the widespread application of deep neural networks, improving inference efficiency has become increasingly critical. To speed up the inference, deep learning compilers search for high-performance schedules for the DNN tensor programs. During the process, cost models have been extensively studied to evaluate the performance of the schedules, such that high-performance ones can be efficiently obtained. However, existing methods suffer from either high overhead or low accuracy of performance evaluation, both of which limit the efficiency of the final schedule. To address these issues, we propose FASP, a fast and accurate framework for schedule performance evaluation, based on Abstract Syntax Trees (ASTs). First, we propose a redundancy-aware ASTs reduction method to generate compact ASTs for more accurate feature extraction. Second, we propose a feature extraction method based on compact ASTs, which extracts features by accounting for computation nodes, loop nodes, and their structural relationships. Third, we propose a composition-similarity-driven ASTs classification method and a class-specific cost model architecture for more accurate performance evaluation. FASP overcomes the limitations of prior methods by significantly reducing evaluation errors. Experiments show its excellent performance in both single-model and cross-model evaluation, with errors ranging from 6 % to$\mathbf{1 3 \%}$. Moreover, FASP can obtain high-performance schedules with$13 \times$lower latency.

The integration of sensing, communication, and computation (ISCC) is a critical technology that will support various emerging wireless services in future 6G networks. The unmanned aerial vehicles (UAVs) equipped with edge servers can be used as an aerial service platform in intelligent transportation systems (ITSs) to offer ISCC services to vehicles. This article studies an aerial UAV network comprising a central UAV and secondary UAVs to realize sensing of the global ITS environment and data fusion computation through collaborative UAVs. To enhance the service performance of ISCC, we maximize the success rate of ISCC services and the energy efficiency of UAVs by jointly optimizing bandwidth allocation, power allocation, and computing capacity control while ensuring the sensing and data processing latency requirements. Leveraging the network architecture and collaboration requirements of UAVs, we propose the multi-UAV collaborative Air-ISCC (MCAI) algorithm based on the asynchronous advantage actor-critic algorithm, which obtains the optimal ISCC service policy by co-training a deep reinforcement learning model with multiple UAVs. Sufficient experimental results show that MCAI enhances energy efficiency by 10.51% to 80.12% compared with the baselines. Moreover, MCAI exhibits good scalability, strengthening its feasibility in real scenarios.

We investigate fast data aggregation via over-the-air computation (AirComp) over wireless networks. In this scenario, an access point (AP) with multiple antennas aims to recover the arithmetic mean of sensory data from multiple wireless devices. To minimize estimation distortion, we formulate a mean-squared-error (MSE) minimization problem that considers joint optimization of transmit scalars at wireless devices, denoising factor, and receive beamforming vector at the AP. We derive closed-form expressions for the transmit scalars and denoising factor, resulting in a non-convex quadratic constrained quadratic programming (QCQP) problem concerning the receive beamforming vector. To tackle the computational complexity of the beamforming design, particularly relevant in massive multiple-input multiple-output (MIMO) AirComp systems, we explore the optimal structure of receive beamforming using successive convex approximation (SCA) and Lagrange duality. By leveraging the proposed optimal beamforming structure, we develop two efficient algorithms based on SCA and semi-definite relaxation (SDR). These algorithms enable fast wireless aggregation with low computational complexity and yield almost identical mean square error (MSE) performance compared to baseline algorithms. Simulation results validate the effectiveness of our proposed methods.

Hierarchical federated learning (HFL) has emerged as the architecture of choice for multi-level communication networks, mainly because of its data privacy protection and low communication cost. However, existing studies on the convergence analysis for HFL are limited to the assumptions of full worker participation and/or i.i.d. datasets across workers, both of which rarely hold in practice. Motivated by this, we in this work propose a unified convergence analysis framework for HFL covering both full and partial worker participation with non-i.i.d. data, non-convex objective function and stochastic gradient. We correspondingly develop a three-sided learning rates algorithm to mitigate data divergences issue, thereby realizing better convergence performance. Our theoretical results provide key insights of why partial participation of HFL is beneficial in significantly reducing the data divergences compared to standard FL. Besides, the convergence analysis allows certain individualization for each cluster in HFL indicating that adjusting the worker sampling ratio and round period can improve the convergence behavior.

Unmanned aerial vehicle (UAV)-based edge computing has been widely applied in intelligent transportation systems (ITSs) owing to its ease of deployment and high mobility. In this article, we study intelligent over-the-air computation (AirComp) in UAV swarm-assisted ITS. To develop a holistic service framework for UAV swarm, we consider the heterogeneity of Internet of Things Devices (IoTDs) and UAVs. We model the 3-D deployment of UAVs, service configuration, bandwidth allocation, the control of computing capacity, and transmission power as a joint optimization problem. To tackle this complex problem, we first propose a dual time-scale architecture based on deep reinforcement learning (DRL). This architecture enables UAVs to achieve seamless coverage of IoTDs on larger time scales, while collaborative UAVs dynamically provide services on smaller time scales. Next, we propose an intelligent AirComp algorithm D2IAC based on distributed DRL to obtain the optimal UAV deployment and dynamic service policies on different time scales. The D2IAC algorithm consists of three subalgorithms, i.e., TD3-based UAV deployment (TBUD), UAV services configuration (USC), and REINFORCE-based dynamic service (RBDS). Sufficient experimental results show that the proposed algorithm can achieve 3-D deployment of UAVs with coverage improvement from 9% to 36% compared to clustering, center layout, and random algorithms. Regarding dynamic services, compared with the deep deterministic policy gradient algorithm, greedy, fixed, and random strategies, the service durations of UAV swarm are improved by 32.95%–93.72% and the resource utilization is improved by 36.19%–49.61%.

Hybrid beamforming has been widely considered in millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) system, which can greatly reduce power consumption and hardware cost of data paths. Compared to the fully-connected hybrid beamforming architecture, the sub-connected architecture is more practical for its reduced complexity. However, optimal precoding design for the sub-connected architecture is not straightforward due to the specific block-diagonal structure of analog phase shifter network. Algorithms on fully-digital or fully-connected hybrid beamforming architecture cannot be directly applied to sub-connected architecture. Meanwhile, most existing precoding algorithms in such case can only solve the approximate problem, which results in significant performance loss. In this paper, we study the sum rate maximization problem in the sub-connected architecture. We first relax the objective function and derive a relaxed upper bound of the original problem. Then we propose an algorithm to solve the original problem with a local-optimal solution. Simulation results show that the proposed local-optimal algorithm outperforms the baseline algorithms with better sum rate and energy efficiency performance. Besides, the proposed algorithm also converges quickly and is robust.

Accurately credit default prediction faces challenges due to imbalanced data and low correlation between features and labels. Existing default prediction studies on the basis of gradient boosting decision trees (GBDT), deep learning techniques, and feature selection strategies can have varying degrees of success depending on the specific task. Motivated by this, we propose Tab-Attention, a novel self-attention-based stacked generalization method for credit default prediction. This approach ensembles the potential proprietary knowledge contributions from multi-view feature spaces, to cope with low feature correlation and imbalance. We organize multi-view feature spaces according to the latent linear or nonlinear strengths between features and labels. Meanwhile, the f1 score assists the model in imbalance training to find the optimal state for identifying minority default samples. Our Tab-Attention achieves superior Recall1 and f11 of default intention recognition than existing GBDT-based models and advanced deep learning by about 32.92% and 16.05% on average, respectively, while maintaining outstanding overall performance and prediction performance for non-default samples. The proposed method could ensemble essential knowledge through the self-attention mechanism, which is of great significance for a more robust future prediction system.

Federated learning (FL) leverages the private data and computing power of multiple clients to collaboratively train a global model. Many existing FL algorithms over wireless networks adopting synchronous model aggregation suffer from the straggler issue, due to the heterogeneity of local computing power and channel conditions. To address this issue, we in this paper advocate an asynchronous FL framework with adaptive client selection for training latency minimization, taking into account the client availability and long-term fairness. We consider a practical scenario, where the channel conditions and the locally available computing power are not known in prior. This makes the client selection problem challenging, as the training latency consists of the uplink/downlink transmission time and the local training time. To this end, we tackle the asynchronous client selection problem in an online manner by converting the latency minimization problem into a multi-armed bandit problem, and leverage the upper confidence bound policy and virtual queue technique in Lyapunov optimization to solve the problem. We theoretically show that the proposed algorithm achieves sub-linear regret performance, ensures long-term fairness, and guarantees training convergence. Results show that the proposed algorithm can reduce the training time by up to 50% when compared to the baseline algorithms.

This work proposes a diagnosis method based on mutual dimensionless indices (MDIs) and receiver operating characteristic (ROC) analysis for the detection of rolling bearing faults, which is of great importance to maintain the functionality of rotating machines. The proposed method consists of five steps. Firstly, the mutual dimensionless technique is used to extract five MDIs from the raw vibration signal. Secondly, the principal components analysis (PCA) is employed to reduce the five MDIs to a one-dimensional feature. Thirdly, we obtain the areas under the ROC curve (AUC) and associated variances using two sliding windows along the one-dimensional feature sequence. Fourthly, the potential fault occurring time is estimated via comparing the AUC and the associated variances with the corresponding detection thresholds. Finally, a parameter K is introduced to delete the false alarms, and then the predicting fault occurring time is chosen from the local extrema of the potential fault occurring times. Experimental results demonstrate that our proposed approach is capable to detect fault occurring time with high accuracy and a low false-positive rate.

Over-the-air computation (AirComp) and millimeter wave (mmWave) communications have the feasibility to perform fast wireless data aggregation (WDA) by allowing simultaneous transmissions and providing abundant spectral resources, respectively. However, AirComp is limited by the link with the worst channel condition, while mmWave communications are vulnerable to the blockages. To address these issues, this paper proposes to leverage reconfigurable intelligent surface (RIS) aided AirComp for WDA in mmWave communication networks. To enhance the system performance, we formulate an optimization problem to minimize the mean-squared error (MSE) of WDA by jointly optimizing the receive beamforming vector of the access point, the transmit scalars of devices, and the phase-shift matrix of the RIS. To this end, we derive the closed-form expression of transmit scalars and then propose a Riemannian conjugate gradient algorithm, which can efficiently tackle the unit-modulus constraints with a low computational complexity. Compared to the baseline algorithms, simulation results reveal that the proposed algorithm achieves a faster convergence rate and a smaller MSE.

The Unmanned aerial vehicle (UAV) assisted Internet of Things (IoT) has attracted substantial attention as it is capable of collecting scattered data to meet the stringent demands of emerging IoT applications. Dispatching UAV to collect data from IoT devices (IoTDs) can significantly improve data freshness, which can be measured by Age of Information (AoI). On the other hand, the quantity of IoTDs increases and existing UAV navigation algorithms for dozens of IoTDs can not be applied to massive IoTDs scenarios directly. In this paper, we investigate the AoI minimization problem in massive IoTDs scenarios. Considering unknown traffic patterns of IoTDs, we reformulate the AoI minimization problem as a Markov decision process (MDP). Then we propose a twin delayed deep deterministic policy gradient (TD3) based UAV navigation algorithm to minimize the average AoI of data collected from IoTDs. Simulation results demonstrate that the proposed algorithm can significantly reduce the average AoI in massive IoTDs scenarios when compared with baseline algorithms.

With the expected explosive use of the Internet of Everything in sixth generation (6G), the cybertwin network is able to convert user information to digital assets and provide extensive services. However, protecting and enhancing privacy of the processed and transmitted data in cybertwin-driven 6G is still in its infancy. Federated learning (FL) is a nascent distributed machine learning paradigm that is able to facilitate privacy protection in cybertwin networks. In a cybertwin network, imbalanced data distribution of the clients can increase the bias of the global model and sacrifice the performance of the FL model. Prior research work dealing with imbalanced data requires extra data information exchanged between clients and the server, which increases the risk of privacy leakage. To avoid privacy leakage, we design an estimation algorithm to determine the distribution of local data collected at the clients without the awareness of specific raw data. We consider two scenarios in FL: 1) the server could receive the individual trained model for each selected device and 2) the server could receive the aggregated model from the selected clients. We formulate two device selection problems to improve the training performance of the aforementioned scenarios. We develop two online learning algorithms to tackle the selection problems for both individual model uploading and aggregated model uploading. The proposed algorithms are conducted on the server, thereby avoiding privacy leakage and extra computation at the clients. We validate the effectiveness of the proposed client selection algorithms with sufficient experiments in cybertwin-driven 6G networks.

Federated learning (FL) is emerging as a promising paradigm for achieving distributed machine learning while protecting users' privacy. The accuracy and convergence speed of the global model benefit from involving as many clients as possible during the model training. On the other hand, the scarcity of wireless spectrum restricts the number of clients involved at each round. In this paper, we aim to maximize the number of participating clients in each round with fixed wireless bandwidth. Instead of assuming that the prior information about wireless channel state is available, we consider a more practical scenario under the absence of prior information. We first reformulate the client selection problem with limited bandwidth as a combinatorial multi-armed bandit (CMAB) problem and then propose an online learning algorithm with elegant bandwidth allocation based on the framework of combinatorial upper confidence bound. The proposed algorithm can make full use of the scarce bandwidth to increase the number of involved clients in each round and minimize the training latency for a given training accuracy. Numerical results validate the efficiency of the proposed algorithm.

Comparing to cloud computing, fog computing performs computation and services at the edge of networks, thus relieving the computation burden of the data center and reducing the task latency of end devices. Computation latency is a crucial performance metric in fog computing, especially for real-time applications. In this article, we study a peer computation offloading problem for a fog network with unknown dynamics. In this scenario, each fog node (FN) can offload its computation tasks to neighboring FNs in a time slot manner. The offloading latency, however, could not be fed back to the task dispatcher instantaneously due to the uncertainty of the processing time in peer FNs. Besides, peer competition occurs when different FNs offload tasks to one FN at the same time. To tackle the above difficulties, we model the computation offloading problem as a sequential FN selection problem with delayed information feedback. Using the adversarial multiarm bandit framework, we construct an online learning policy to deal with delayed information feedback. Different contention resolution approaches are considered to resolve peer competition. Performance analysis shows that the regret of the proposed algorithm, or the performance loss with suboptimal FN selections, achieves a sublinear order, suggesting an optimal FN selection policy. Besides, we prove that the proposed strategy can result in a Nash equilibrium (NE) with all FNs playing the same policy. Simulation results validate the effectiveness of the proposed policy.

Fog computing provides computation and services to the edge of networks to support real-time applications. The latency performance is a crucial metric in fog computing. In this article, we consider a computation offloading problem in a fog network with unknown dynamics. In this network, mobile users can offload their computational tasks to neighborhood fog nodes (FNs) in each time slot. The queue of arrival tasks at each FN follows a Markov model with unknown statistics. In order to provide a satisfactory quality of experience, the network latency needs to be minimized. In this article, we construct an offloading policy with interleaved exploration and exploitation epochs to solve the sequential FN selection problem. An upper bound of regret is derived to show the effectiveness of the proposed method. The proposed policy is optimal in the sense that it achieves a regret with sublinear order. In addition, the proposed policy can be applied to both single-user setting and multiuser setting. Simulation results show that when compared with the existing offloading algorithms, the proposed algorithm can reduce the average latency by 7%–47% in the single-user setting, and 91% in the multiuser setting.

Fog computing brings computation and services to the edge of networks enabling real time applications. In order to provide satisfactory quality of experience, the latency of fog networks needs to be minimized. In this paper, we consider a peer computation offloading problem for a fog network with unknown dynamics. Peer competition occurs when different fog nodes offload tasks to the same peer FN. In this paper, the computation offloading problem is modeled as a sequential FN selection problem with delayed feedback. We construct an online learning policy based on the adversary multi-arm bandit framework to deal with peer competition and delayed feedback. Simulation results validate the effectiveness of the proposed policy.

Fog computing is emerging as a promising paradigm to perform distributed, low-latency computation. Efficient computation peer offloading is critical to fully utilize the computational resources in fog networks. In this paper, we consider computation peer offloading problem in a fog network with time-varying stochastic time of arrival tasks and channel conditions. Such time-varying conditions are not available to all fog nodes. In order to minimize the latency of accomplishing arrival tasks, we propose an online algorithm based on combinatorial upper confidence bounds algorithm with two uncertain variables under the non-stationary bandit model. The proposed computation offloading policy is optimized based on historical feedback. The performance of the proposed scheme is validated through numerical simulations.

In wireless sensor networks, sensor nodes have limited energy budget, in general. Energy efficiency is a critical issue which is directly related to the network lifetime. On the other hand, stringent latency requirements are enforced in some applications. To save energy, ordering transmissions is an effective approach in which sensor nodes transmit more informative data to the fusion center earlier. Ordering, however, does not perform well in latency-sensitive scenarios. In this paper, we propose a distributed method based on the framework of ordered transmissions. The proposed algorithm is illustrated in the discretized estimation problem with latency constraint. In our proposed method, each sensor node has specific time slots to transmit data, and can determine its transmission order independently. The proposed algorithm can greatly reduce latency without loss of estimation accuracy, while the increased number of transmissions is negligible. Simulation results validate its effectiveness.

Fog computing is a promising architectural to alleviate increasingly intensive transmission over the network. In addition to the data transmission capability, a fog node (FN) also has spare resources of data storage and computing. In this paper, we study the computation offloading scenario that takes advantage of fog architecture and utilizes the FN resources. A social welfare maximization problem is formulated to distribute the data among FNs based on the trade-off between the considered computational cost and communication cost. A distributed adaptation algorithm is developed based on a Jacobi-Proximal alternating direction method of multipliers (ADMM) algorithm. The computational burden for solving the optimization problem is fully distributed to FNs, software defined network (SDN) controller, where local variables of FNs are updated in parallel. Performance of the proposed algorithm is validated with simulation results.

In wireless sensor networks (WSN s), energy is always precious for sensor nodes. To save energy, censoring is introduced to cut the total number of transmission by only transmitting informative data. This algorithm, however, ignores the energy consumption during the delivery of parameters, which can be significant comparing to the saved power. In this paper, we consider the adaptive censoring from the energy perspective. A distributed censoring algorithm with energy constraint is developed that allows sensor nodes to make autonomous decisions on whether to transmit the incoming data. We show that with the proposed algorithm, the overall energy consumption of the WSN s is reduced, while the performance loss in terms of the estimation error is negligible. Simulation results validate its effectiveness.

In the era of big data, adaptive censoring (AC) provides us a natural option of trimming data by only keeping the statistical informative data. However, the data chosen by AC may arrive in clusters, which do not relieve the computational resource requirement as expected. In this letter, we exploit queuing theory to model a single sink node with abundant sensor nodes. By adding a buffer to censored distributed wireless sensor networks (WSNs), the uncensored data can be modeled as a queue. With the buffer, the new algorithm entails simple, closed-form updates, and has no loss in terms of estimation accuracy comparing to the original AC method. The proposed model can further reduce the communication cost of distributed WSNs. The proposed model is illustrated in a linear regression setting. Numerical results validate the effectiveness of the proposed model in dealing with data congestion problem.

As more than 2.5 quintillion bytes of data are generated every day, the era of big data is undoubtedly upon us. Running analysis on extensive datasets is a challenge. Fortunately, a significant percentage of the data accrued can be omitted while maintaining a certain quality of statistical inference in many cases. Censoring provides us a natural option for data reduction. However, the data chosen by censoring occur non-uniformly, which may not relieve the computational resource requirement. In this paper, we propose a dynamic, queuing method to smooth out the data processing without sacrificing the convergence performance of censoring. The proposed method entails simple, closed-form updates, and has no loss in terms of accuracy comparing to the original adaptive censoring method. Simulation results validate its effectiveness.

In this paper, we propose a coupled level set (LS) framework for segmentation of bladder wall using T(1)-weighted magnetic resonance (MR) images with clinical applications to virtual cystoscopy (i.e., MR cystography). The framework uses two collaborative LS functions and a regional adaptive clustering algorithm to delineate the bladder wall for the wall thickness measurement on a voxel-by-voxel basis. It is significantly different from most of the pre-existing bladder segmentation work in four aspects. First of all, while most previous work only segments the inner border of the wall or at most manually segments the outer border, our framework extracts both the inner and outer borders automatically except that the initial seed point is given by manual selection. Secondly, it is adaptive to T(1)-weighted images with decreased intensities in urine, as opposed to enhanced intensities in T(2)-weighted scenario and computed tomography. Thirdly, by considering the image global intensity distribution and local intensity contrast, the defined image energy function in the framework is more immune to inhomogeneity effect, motion artifacts and image noise. Finally, the bladder wall thickness is measured by the length of integral path between the two borders which mimic the electric field line between two iso-potential surfaces. The framework was tested on six datasets with comparison to the well-known Chan-Vese (C-V) LS model. Five experts blindly scored the segmented inner and outer borders of the presented framework and the C-V model. The scores demonstrated statistically the improvement in detecting the inner and outer borders.

In our surrounding environment, we may often see many various miscible liquid-liquid mixture phenomena, like pouring honey or ink into water, Coca Cola into strong wine etc., while few papers have devoted to the simulation of the phenomena. In this paper, we use a two-fluid lattice Boltzmann method (TFLBM) to simulate the underlying dynamics of miscible mixtures. By the method, a subgrid model is applied to improve its numerical stability so that the free surface of the mixture, accompanying with higher Reynolds number, can be simulated. We also apply control forces to the mixture with interesting animation created. By optimizing the memory structure and taking the advantage of dual-core or multi-core systems, we achieve real time computation for a domain in 643 cells full of fluid mixtures.

Through interaction with surroundings, the fluids may change their properties such as shapes, temperature vastly, and the same would happen to the surroundings simultaneously. On the other hand, different surroundings characterize different interactions, and may change the shapes and motions of the fluids in different ways. Therefore, it is of importance in physically-based simulation of fluids to build physically correct models to represent the varying interactions between fluids and the environments. In this paper, we make a simple summation on the interactions, and in particular focus on those most interesting to us, and model them with various physical solutions. In some of the methods, advantage is taken with the graphics processing unit (GPU) to achieve real-time computation for medial-scale simulation

Abstract Miscible fluid mixtures, like pouring honey into water, Coca Cola into strong wine, are common phenomena in our daily life. While two miscible fluids are mixed together, their appearances in terms of colors and shapes will change due to their mixing interaction. The interaction between the mixture components could be regarded as a combination of the diffusing process and demixing process. If the former dominates the interaction, it is miscible; otherwise, it is immiscible. The complex microscopic interplay between the mixture components makes the simulation highly challenging. So far, there have been some dedicated research in computer graphics dealing with immiscible mixtures, but few works have been done focusing on miscible mixtures. In this paper, for the first time, we introduce a two‐fluid lattice Boltzmann method (LBM), called TFLBM, applied to miscible binary mixtures. Different from other similar methods, the viscous and diffusing properties of the fluid in our work are considered separately, so that the physical insight is exposed more clearly and rationally. In addition, the operation of LBM is mostly a linear local computation, and graphics processing unit (GPU) has been utilized to achieve real‐time simulation. Copyright © 2006 John Wiley & Sons, Ltd.