最新刊期
- 摘要：In multi-unmanned-aerial-vehicle (UAV)-assisted Internet of Vehicles (IoV) systems, dynamic variations in traffic flows and task loads can cause mismatches between communication and computing resource supply and demand. To support delay-sensitive tasks, a coordinated resource scheduling method based on a flow matching model was proposed. An optimization model considering task delay, system energy consumption, and deadline violations was formulated, and a three-layer trajectory–resource–offloading mechanism was designed. UAV deployment was optimized through convex inner approximation and sequential optimization, resource-priority and transmit-power candidates were generated and screened using the flow matching model, and task partitioning was optimized based on path availability and queueing feedback. In simulations, lower total system delay and energy-delay product, as well as a higher task success rate, were achieved by the proposed method. The proposed method effectively improves service assurance for delay-sensitive tasks.关键词：IoV;UAV-assisted edge computing;flow matching model;coordinated resource scheduling;service assurance2|0|0更新时间：2026-09-20
- 摘要：Multimodal recommender systems are widely used in content delivery and online services. The third-party visual and textual pretrained encoders are often reused to downstream recommenders through feature caches. This decoupled deployment improves efficiency but exposes pretrained components and cached representations to supply-chain backdoors. If encoder outputs or cached features are covertly poisoned, attackers can bias target-user rankings without changing model parameters or training data, while keeping overall performance nearly unchanged. To address this threat, we propose MAGuard, a multi-view anomaly-gated defense framework. MAGuard detects poisoned representations by fusing local consistency, collaborative behavior, ranking sensitivity, and global distributional deviation into an item-level risk score. It then repairs recommendations at inference time by gating untrusted multimodal features and calibrating collaborative filtering scores, without model retraining. Experiments on MovieLens-10K and Amazon-TG show that MAGuard effectively detects embedding-level backdoors and mitigates attacks with minimal performance loss. This work supports pretrained-component auditing and AI supply-chain protection for multimodal recommender systems.关键词：Multimodal recommendation system;pre-trained encoder;supply chain security;backdoor defense3|0|0更新时间：2026-09-20
- 摘要：Existing glaucoma screening methods based on Convolutional Neural Networks (CNNs) suffer from weak global modeling capability and insensitivity to positional information, owing to the limitation of local receptive fields. Meanwhile, Vision Transformers suffer from high computational complexity and are prone to overfitting on small- and medium-scale datasets. To address these issues, this paper proposes a quantum theory-inspired method for glaucoma screening. The core of the proposed method is a quantum-inspired dynamic anchor bidirectional feature interaction mechanism, which dynamically generates complex-valued anchors, performs joint phase-amplitude modulation, and conducts canonical connectivity-driven feature aggregation. This mechanism achieves efficient local-global feature fusion while reducing computational complexity. In addition, a global context enhancement module and a self-interaction module are introduced to effectively improve the localization of key features and the generalization ability for scale variations. Experiments on public datasets such as AIROGS demonstrate that the proposed method achieves relatively good performance in glaucoma screening with fewer parameters and lower computational cost.关键词：Glaucoma screening;Quantum theory;Dynamic anchors;feature interaction;feature fusion3|4|0更新时间：2026-09-18
- 摘要：Temporal knowledge graph completion aims to predict future missing entity links based on historical observed facts. To overcome the issues of remote information truncation and redundant noise, a method named Historical Summary Enhancement (HSE) was proposed. In this method, entity global historical behavior representations were first extracted through a relation graph neural network. Subsequently, compact historical summaries were constructed to condense cross-temporal historical information. Finally, the historical summaries and current query representations were adaptively combined via a gated fusion mechanism for link prediction. Better results were achieved by the proposed method on the ICEWS14 and ICEWS18 datasets, with the MRR and Hits@10 on ICEWS14 reaching 47.41% and 71.00%, respectively. This effectively improves the performance of temporal knowledge graph completion.关键词：Temporal knowledge graph completion;Historical summary enhancement;Long-range temporal dependency;Time encoding;link prediction2|2|0更新时间：2026-09-18
-
Underwater acoustic ISAC beamforming based on deep reinforcement learning with mixture of experts AI导读
摘要：To achieve the trade-off between communication and sensing performance in underwater acoustic integrated sensing and communication systems, an efficient and highly generalizable beamforming scheme was proposed. First, to address the issues of outdated channel state information and cross-time slot echo interference caused by long propagation delays, a deep reinforcement learning framework was introduced for beamforming decision-making. Second, to overcome the limitations of traditional deep reinforcement learning methods in adapting to diverse performance requirements, a neural network based on a multi-task mixture-of-experts model was constructed to adapt to the trade-offs between communication and sensing performance under different objective-preference factors. In addition, to characterize the long-term impact of current beamforming decisions on system performance, a multi-step cross-task experience replay mechanism was designed, which combines multi-task samples and multi-step rewards to improve learning efficiency. Simulation results demonstrated that compared with benchmark schemes, the proposed scheme obtained more Pareto-efficient solutions under different objective-preference factors and exhibited stronger generalization capability.关键词：underwater acoustic networks;integrated sensing and communication;deep reinforcement learning;mixture of experts4|2|0更新时间：2026-09-18 - 摘要：To address the characteristics of complex perception tasks in high-speed railway scenarios, including directed acyclic graph (DAG) dependencies, high-speed mobility, and dynamically varying trackside computing resources, a mobility- and dependency-aware hierarchical proximal policy optimization algorithm (MDH-PPO) was proposed, since conventional offloading methods designed for independent tasks cannot efficiently support online decision-making under subtask dependency constraints. First, complex perception services were modeled as DAG tasks, and a collaborative multi-edge computing model was established to characterize input uploading, inter-server migration, task queueing, and computation. The dynamic offloading problem was then formulated as a Markov decision process. A hierarchical offloading mechanism comprising topology scheduling and server decision-making was developed, together with a joint action mask. An enhanced proximal policy optimization method was further employed to learn an online offloading policy in high-mobility environments. In the simulations, lower average end-to-end latency and higher system computational efficiency were achieved with MDH-PPO than with the benchmark algorithms under varying computing loads, edge computing capacities, and train operating speeds. These results demonstrate the algorithm’s stable training performance and strong adaptability to high-speed mobility.关键词：high-speed railway;edge computing;dependency awareness;task offloading;proximal policy optimization4|2|0更新时间：2026-09-18
- 摘要：The channel estimation problem in high-speed UAV THz UM-MIMO systems under hybrid near-/far-field propagation, wideband beam squint, Doppler spread, and attitude variations was investigated. A wideband time-varying hybrid-field channel model was established, and an FPN-OAMP-DFMN-TWF method was proposed. Adjacent-slot correlation was exploited by TWF, and multi-slot observations were fused using learnable weights. DFMN, the nonlinear estimator in FPN-OAMP, was composed of adaptive multi-scale fusion and cross-block attention modules; multi-scale angular-domain features were extracted and inter-block information interaction was enhanced. Lower NMSE and faster convergence than FPN-OAMP were demonstrated by simulations, and stable performance was maintained under different UAV speeds, path numbers, and undersampling ratios. The proposed method achieves robust channel estimation for high-speed UAV THz UM-MIMO systems.关键词：Terahertz Communications;ultra-massive MIMO;channel estimation;UAV Communications;Hybrid Near-far Field;attention mechanism4|3|0更新时间：2026-09-17
- 摘要：To address the mismatch between TAS-gated transmission and CQF-periodic forwarding caused by dynamic changes in link delay and frequent handovers in low-Earth orbit satellite networks, as well as issues such as sudden increases in delay, out-of-order delivery, and packet loss for time-sensitive services, a handover-aware TAS-CQF cooperative scheduling deterministic guarantee method is proposed. This method jointly adjusts the CQF period offset and TAS transmission phase based on the comprehensive link delay to achieve period alignment under dynamic link conditions. Furthermore, deterministic network calculus is used to obtain the delay boundaries of the old and new paths, quantify the maximum path differential delay, and adaptively configure the MBB dual-path protection interval and reordering buffer accordingly to ensure continuous and orderly transmission of services during handover. Simulation results show that the proposed method can effectively reduce the packet loss rate and maximum end-to-end delay of time-sensitive services, improve effective throughput, maintain good transmission stability under satellite handover and delay disturbance conditions, and reduce redundant protection overhead in handover scenarios.关键词：LEO satellite network;TAS-CQF;Coordinated scheduling;Certainty guarantee;network calculus6|4|0更新时间：2026-09-16
- 摘要：With the rapid growth of AI, data-center energy consumption continues to increase. To balance computing service requirements and renewable-energy fluctuations, a grid carbon-intensity-aware power–computing coordinated scheduling method was proposed. First, a scheduling framework for geographically distributed heterogeneous data centers was developed to characterize how carbon emission attributes on the power supply side influence task allocation on the computing side. Then, considering workflow dependencies, heterogeneous resources, and spatiotemporal variations in grid carbon intensity, a GINEConv-PPO scheduling mechanism was proposed. A temporal planning algorithm extracted critical-chain positions and resource-contention risks to construct temporal-priority features, while GINEConv was integrated with PPO for low-carbon task allocation under QoS constraints. Simulations across workflow types and sizes showed that GINEConv-PPO met job-batch deadlines. Relative to the lowest emissions achieved by the compared algorithms in each scenario, its emission deviation remained near 0%, versus approximately 6%–35% for the baselines.关键词：Power-Computing co-scheduling;task scheduling;carbon emissions;deep reinforcement learning7|3|0更新时间：2026-09-16
- 摘要：To address the key challenges including insufficient terminal samples, heterogeneous edge generation capabilities, and cloud-edge information asymmetry in federated learning for the Industrial Internet of Things, this paper proposes an intent driven edge data augmentation mechanism for federated learning in industrial IoT. The proposed mechanism employs a large language model to interpret user intents and transform unstructured requirements into optimizable structured parameters. Edge-generated data are then allocated according to the distribution deviation of terminal data. In addition, a contract theory-based incentive mechanism is developed for edge augmentation services, which encourages heterogeneous edge nodes to provide effective data augmentation. Simulation results show that the proposed method can improve model training performance in non-independent and identically distributed (Non-IID) scenarios and adaptively adjust the edge augmentation service level under different user intent constraints, making the training outcomes better aligned with users' heterogeneous requirements for accuracy, latency, and cost.关键词：industrial Internet of things;intent-driven networking;federated learning;data augmentation;contract theory;edge computing11|7|0更新时间：2026-09-16
- 摘要：Existing methods for detecting anonymous proxy traffic rely heavily on protocol-specific features and generalize poorly to unseen protocols. To improve cross-protocol generalization, a behavior-invariant state-space model (BI-SSM) is proposed. BI-SSM first uses a context-gated adaptive collaborative embedding to capture dependencies among behavioral attributes in microburst sequences. It then uses the intervals between behavioral units to control the state-evolution step size and dynamically generates state input and output projections to model temporal dependencies on a nonuniform time axis. Finally, statistical moments and quantiles are used to summarize the distribution of hidden states within each session, while classification and supervised contrastive losses jointly optimize the flow-level representation. Leave-one-protocol-out experiments on two datasets covering four proxy protocols show that BI-SSM achieves an average F1 score of 98.87%, outperforming the best proxy-specific baseline by 2.60 percentage points and demonstrating its ability to detect unseen proxy protocols.关键词：anonymous proxy;traffic classification;state-space model;encrypted traffic5|5|0更新时间：2026-09-16
- 摘要：To meet the requirements of fifth-generation (5G) new radio (NR) for high-rate and reliable data transmission, an efficient low-density parity-check (LDPC) decoding system was proposed. The system used odd-even interleaving to decompose each quasi-cyclic matrix into two submatrices, while a differential circular-shift network was employed to further reduce hardware redundancy and alleviate routing congestion. To reduce pipeline stalls caused by inter-layer data dependencies, a three-frame interleaved pipeline and an offline static scheduling strategy were introduced. The strategy replaced complex conflict detection with a pre-generated microcode stream. Field-programmable gate array (FPGA) implementation and fixed-point simulation results show that, for both base graphs and various configurations of code length, code rate, and lifting size defined in the 5G-NR standard, the frame error rate (FER) decreases rapidly with no noticeable error floor. The proposed system achieves a normalized throughput-to-resource ratio comparable to state-of-the-art commercial intellectual property (IP) cores and further integrates a rate-dematching module. These results demonstrate that the proposed system can satisfy the channel decoding requirements of non-terrestrial network (NTN) and related scenarios.关键词：5G new radio;low-density parity-check code;decoder;rate dematching;field-programmable gate array4|3|0更新时间：2026-09-16
- 摘要：Extreme disasters cause severe damage to power facilities and communication disruptions, hindering dispatch centers from obtaining critical operational information and impairing post-disaster decision-making. To address scenarios where terrestrial base stations (BSs) and backhaul links are partially damaged, unmanned aerial vehicles (UAVs) can establish multi-hop integrated access and backhaul (IAB) links to restore power communication connectivity. To further enhance situational awareness and spectrum reuse, this paper proposes a software-defined networking (SDN)-controlled, UAV-assisted integrated sensing and communication (ISAC) method for power grid emergency communication recovery. An air-ground multi-hop network model comprising available terrestrial BSs, control centers, and UAVs is developed, unifying access coverage, sensing accuracy, pilot resources, slice routing, and bandwidth reservation into a single optimization framework. To tackle the high problem complexity and network state uncertainty, a surrogate value function-based multi-timescale rolling optimization (SVF-MTRO) approach is proposed. Specifically, UAV deployment and sensing tasks are determined at a second-level timescale; routing and slice bandwidth reservations are updated based on real-time link states at a hundred-millisecond timescale; and ISAC scheduling alongside traffic forwarding is executed per radio frame based on instantaneous resource block capacities. Simulations on an IEEE 39-bus system demonstrate that the proposed method effectively guarantees the latency performance and service recovery rate of critical power communication slices, while significantly enhancing the data delivery capability of the emergency communication network.关键词：integrated sensing and communication;power grid emergency communication;multi-hop networking;software-defined networking;multi-timescale optimization8|3|0更新时间：2026-09-16
- 摘要：To address the collaboration issue of trajectory planning and communication-sensing tasks in unmanned aerial vehicle integrated sensing and communication (ISAC) system assisted by fluid antenna, a deep reinforcement learning (DRL) based multidimensional resource co-optimization strategy was proposed. Firstly, an unmanned aerial vehicle ISAC model was constructed where fluid antennas were equipped at both transceivers. The unmanned aerial vehicle trajectory, user and target scheduling, and transceivers antenna location selection were jointly optimized. Then the communication rate maximization problem was divided into two sub-problems. A multi-worker advantage actor-critic (MW-A2C) algorithm was introduced, which employed the probabilistic mandatory action to assist the unmanned aerial vehicle in reaching the destination. Finally, the greedy algorithm was adopted to solve the scheduling sub-problem, and an immediately performance feedback mechanism was designed to update the weights of the DRL networks. Simulation results demonstrate that the proposed strategy enables flexible adjustment of the unmanned aerial vehicle trajectory, and can effectively enhance the overall ISAC performance. Moreover, the MW-A2C strategy outperforms the baseline strategies in term of the scalability and robustness.关键词：Fluid antenna;unmanned aerial vehicle networks;ISAC;multidimensional resource;DRL11|4|0更新时间：2026-09-16
- 摘要：In limited-data image classification, freezing a pretrained visual frontend controls the trainable scale, but the low-dimensional backend must balance stable classical representations with nonlinear remapping. Controlled fusion of classical and quantum representations was proposed within a hybrid quantum transfer learning framework, combining a frozen residual network 18 (ResNet18), a six-dimensional bottleneck, a variational quantum circuit, and sample-wise gating. Experiments were conducted on Kaggle chest X-ray binary classification under 100–300 training samples per class and OCTMNIST retinal optical coherence tomography four-class classification with 300 training samples per class, each using six random seeds. The highest mean accuracy at most Kaggle training-set sizes was achieved by the controlled-fusion model based on the strongly entangling template (SET), denoted SET-Mix; more consistent mean accuracy (Acc) and area under the receiver operating characteristic curve (AUC) were also obtained relative to its no-mix direct quantum-replacement counterpart (SET-NM). On OCTMNIST, the highest mean Acc with relatively low variation was achieved by SET-Mix. The structure-matched classical control, sample-wise complementarity, gating, circuit-depth, and noise-perturbation analyses further indicate that controlled dual-branch fusion helps retain a classical anchor while introducing complementary branch information in a stable manner. Overall, the results position the quantum representation as a controlled complement to, rather than a direct replacement for, the classical representation.关键词：hybrid quantum transfer learning;classical–quantum representations;controlled fusion;variational quantum circuit;image classification30|4|0更新时间：2026-09-11
- 摘要：To address the high dimensionality of cascaded channels and the high complexity of polar-domain codebooks in extremely large-scale reconfigurable intelligent surface (XL-RIS)-assisted near-field time-varying channel estimation, a low-complexity Bayesian channel estimation and tracking method was proposed. First, by exploiting the static property of the base station (BS)-RIS link and the double-structured sparsity of the cascaded channel, the BS-side common support was estimated, and the received signals were projected onto the corresponding subspace for dimensionality reduction; consequently, the cascaded channel estimation problem was transformed into an equivalent near-field channel estimation problem at the RIS side. Then, a multiple measurement vector (MMV) model was constructed in the first frame, and multiple measurement vector sparse Bayesian learning (MSBL) was employed to update the adaptive codebook with fixed angular sampling, through which the angular support and Bayesian priors were obtained. Finally, in the subsequent transmission blocks, the adaptive codebook was fixed, and a hierarchical Bayesian Kalman filter (HBKF) was employed to recursively track the time-varying sparse coefficients. The ability of the proposed method to improve near-field time-varying channel estimation accuracy while reducing estimation complexity was demonstrated by the simulation results. These results demonstrate the effectiveness of the proposed Bayesian estimation and tracking framework.关键词：XL-RIS;channel estimation;near-field;time-varying;sparse Bayesian learning10|7|0更新时间：2026-09-11
- 摘要：To address the joint constraints of LED nonlinearity, multicarrier fluctuations, indoor multipath effects, and channel time-variability on reliability and efficiency, an adaptive joint constellation shaping scheme fused with CNN sensing was proposed. Firstly, 3D cubic constellations were constructed via T-F reorganization to expand MED and decision margin; simultaneously, Discrete Fourier tranform spread-based precoding was introduced to smooth the envelope by reshaping temporal correlation. Secondly, nonlinear memory features were implicitly extracted by CNN and combined with a GMI-maximized LUT to implement the closed-loop optimization of the shaping factor under dynamic conditions. Experimental results showed that the CNN-fused 3D-PS reduced clipping distortion by concentrating weights toward the center, which, together with CNN's correction of constellation offsets, effectively eliminated error floors in modulation regions. Significant SNR and GMI gains were achieved compared with uniform mapping while low PAPR was maintained. In summary, this mechanism provides an effective solution for high-efficiency VLC in indoor multipath environments.关键词：visible light communication;3D constellation;probabilistic shaping;channel estimation;neural network11|7|0更新时间：2026-09-11
- 摘要：To achieve the joint optimization of communication, sensing, and computing performance in integrated sensing, communication, and computation-based Internet of Vehicles (ISCC-IoV) systems, this paper investigated a vehicle sensing request-driven end-edge cooperative sensing task in multi-roadside unit (RSU) coverage scenarios, and proposes a joint sensing, communication, and computation resource allocation scheme, where a system utility function was formulated to jointly consider sensing data size, task latency, and energy consumption. The joint optimization of RSU selection and multi-dimensional resource allocation was then modeled as a non-convex and NP-hard mixed-integer nonlinear programming problem. To solve this problem, a Dual-Critic Heterogeneous Multi-Agent Reinforcement Learning (DCHMARL) algorithm was proposed, where vehicles and RSU were modeled as heterogeneous agents. By combining a dual-critic mechanism with a hierarchical Actor structure, DCHMARL improves training stability and learning efficiency. Simulation results demonstrate that the proposed resource allocation scheme achieves superior overall communication, sensing, and computation performance under dynamic system conditions. Furthermore, compared with benchmark RL algorithms, DCHMARL achieves higher system utility, lower task failure rates, and faster, more stable convergence.关键词：integrated sensing;communication and computation;cooperative sensing;resource allocation17|14|0更新时间：2026-09-11
-
A Traceable and Revocable Registered Attribute-Based Encryption Scheme for Medical Data Sharing AI导读
摘要：In medical data sharing, the leakage, transfer, or misuse of internal users’ private keys may lead to illegal access to sensitive data. Therefore, it is necessary to trace the responsibility for leaked private keys and revoke the relevant permissions in a timely manner. Traditional traceable attribute-based encryption usually generates complete user private keys by a central authority and embeds tracing information into them. However, in registered attribute-based encryption, users independently generate the core components of their private keys, and the key manager does not possess complete private keys, making it difficult to directly embed tracing information. To address this problem, this paper proposes a traceable and revocable registered attribute-based encryption scheme, TRRABE. Based on the user-generated private key components, the scheme constructs identity-associated tracing components, enabling the key manager to locate the responsible user through the leaked private key without possessing the complete private key. It combines key update and ciphertext update mechanisms to realize post-tracing revocation and revoke the access permission of the user whose private key has been leaked. It also reduces the identity association risk during the revocation process through user identity hashing. Security analysis demonstrates that TRRABE is secure against chosen-plaintext attacks and provides private-key leakage traceability and revocation security. Performance evaluation shows that the proposed scheme incurs acceptable computational and storage overheads.关键词：attribute-based encryption;registered attribute-based encryption;private key tracing;user revocation;identity protection10|9|0更新时间：2026-09-11 - 摘要：To address UAV-enabled edge computing service interruptions, tail-task backlogs, and delay accumulation in constrained airspace due to flight-corridor constraints, link intermittency, and energy scarcity, a Lyapunov optimization-based service recovery algorithm is proposed. A collaborative model of service UAVs, charging UAVs, and outpost recharging nodes characterizes the coupling among task execution, energy consumption, predictive recharging, and task migration. Energy-deficit, outpost recharging-load, and migration-task backlog virtual queues transform the long-term service recovery problem into per-slot drift-plus-penalty minimization. LKH-based path construction, queue-aware recharging matching, and a global migration pool enable interrupted tasks to resume across UAVs. Simulations show that, for 100–500 task points, the full strategy reduces the average resumption and service completion delays of interrupted tasks by approximately 65.7% and 30.8%, respectively, relative to its multi-SUAV variant without the global migration pool, thereby improving service continuity and recovery efficiency.关键词：UAV-enabled edge computing;service recovery;Lyapunov optimization;task resumption5|9|0更新时间：2026-09-11
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
- Total Visits：634731 Visits Today：579 Online Users：14
0