Search Results (125)

Internet of Things(IoT)/edge intrusion-detection systems operate on distributed traffic and system-state data whose distributions vary across gateways, services, and attack conditions. We study a server-assisted federated setting in which a teacher reference is fitted from a permitted server-accessible training pool and explicitly distinguish this simulation assumption from fully decentralized deployment. The proposed framework evaluates progressive local training through boundary stabilization, confidence-weighted decision distillation, representation alignment, and validation-quality-aware aggregation. The evaluation uses a leakage-controlled protocol: server and client validation subsets are held out before federated training, update quality and early stopping use validation data only, and the final-test split is evaluated once. Results on NSL-KDD, CIC-IDS2017, Edge-IIoTset, and the ToN-IoT network dataset show competitive primary performance and stronger robustness in several severe label-skew settings. On the Telemetry of Things(ToN-IoT) with Dirichlet alpha = 0.1, the proposed method achieves 91.46 ± 5.54 F1, compared with 53.73 ± 49.00 for FedAvg and 53.77 ± 48.92 for FedProx. The results do not establish universal superiority or a universally optimal stage order: competing methods remain stronger in selected stable and attack-shift settings. The framework is therefore presented as a bounded, server-assisted robustness-oriented training strategy for heterogeneous IoT/edge intrusion detection. Full article

Efficient task offloading in UAV-assisted heterogeneous mobile edge computing (MEC) networks is increasingly challenged by the co-existence of operationally distinct workload scenarios—including high-demand bursts, resource-constrained periods, and balanced operational states—each demanding fundamentally different assignment strategies. In such networks, mobile executor nodes (e.g., UAVs or vehicle-mounted edge servers) must be dispatched to the vicinity of geographically distributed tasks, making assignment decisions jointly dependent on node mobility, the quality of sensing data, and dynamic resource availability. Conventional approaches based on combinatorial optimization with fixed parameters or greedy heuristics fail to adapt to these varying conditions, leading to resource depletion under sequential workloads or underutilization under high-demand bursts. To address these limitations, this paper proposes SAGE (Scene-Adaptive Graph-Enhanced offloading), a task-offloading framework that combines a heterogeneous graph neural network (HeteroGNN) with a dueling double DQN meta-controller and a mixed-integer linear programming (MILP) solver. At the state-representation level, a heterogeneous bipartite graph is constructed over mobile executor nodes and tasks, with type-specific projection layers encoding the semantic features of each node type and three-dimensional edge features—comprising task success probability, normalized service distance, and link quality—integrated via edge-gated message passing. At the decision level, the meta-controller perceives the current workload scenario through a seven-dimensional situational state vector fused with the graph embedding, selects an appropriate offloading strategy from a learned discrete action space, and drives the MILP solver to perform task-chain assignment under the selected configuration. Experiments on 60 fixed evaluation episodes spanning three representative workload scenarios demonstrate that SAGE achieves an overall reward improvement of 15.9% over the best fixed-strategy baseline, reduces the resource depletion rate to 16.7%, and maintains a high-priority task completion rate of 84.7%. Particularly under resource-constrained conditions, SAGE reduces the reward deficit by 72.3% relative to the best fixed strategy (from $-0.531$ to $-0.147$), demonstrating strong scene-adaptive decision-making capability. Full article

This study presents a multi-tier framework for deploying multi-modal Conversational AI on edge devices, spanning from constrained ultra-low-power systems to high-performance edge workstations. Utilizing an automated model discovery process and a modular benchmarking testbed, the research demonstrates that real-time, fully edge AI execution is feasible through strategic model selection and hardware acceleration. Key outcomes from the performance analysis are as follows. Speech-to-Text: Fun-ASR-Nano achieved the highest transcription accuracy with a Word Error Rate of 0.026, while Moonshine Tiny was the most efficient, recording a Real-Time Factor of 0.036 on the CPU. Scaling up to the high-performance tier, Whisper Large-V3 Turbo demonstrated high speed and robustness on a dedicated GPU, achieving an RTF of 0.093. Language Modeling: The Qwen 2.5 (1.5B Instruct) model, optimized for the Intel edge NPU, delivered robust constrained edge performance with an average generation speed of 20.15 tokens per second and a high semantic accuracy score of 0.86. The non-transformer Liquid LFM-24B model showcased server-level reasoning capabilities on the high-performance edge, reaching an impressive 39.2 tokens per second when fully offloaded to a dedicated GPU, despite its massive VRAM requirements. Text-to-Speech: Piper TTS emerged as the most efficient model for constrained environments (RTF of 0.034). However, Kokoro TTS redefined high-fidelity zero-shot synthesis on the GPU tier, achieving a groundbreaking RTF of 0.024 and far outperforming larger autoregressive audio models like OuteTTS, which remained too slow for real-time use without significant acceleration. Hardware Acceleration and Energy Efficiency: The use of Intel OpenVINO 2026.0 for hardware offloading significantly reduced energy consumption; for example, Whisper Large-V3 Turbo’s energy per audio second dropped from 52.68 Joules on the CPU to just 3.24 Joules on the integrated GPU. Furthermore, dedicated GPU acceleration revealed a critical “race-to-sleep” paradigm, where higher peak wattage is offset by drastically reduced processing times. The study concludes by identifying two optimal cascaded pipelines: a constrained edge tier (Moonshine, Qwen 1.5B, Piper) running on a Khadas NUC (Khadas Technology, Shenzhen, China powered by an Intel processor (Intel Corporation, Santa Clara, CA, USA) maximizing energy efficiency, and a high-performance tier (Whisper V3 Turbo, Liquid LFM-24B, Kokoro) running on an NVIDIA 5060ti, delivering uncompromising accuracy and subsecond latency for privacy-preserving, advanced edge AI. Full article

Telemedicine in emergency and remote care relays medical images from ambulances and rural clinics to a hospital edge-computing server over a congested wireless uplink. Existing work prices such transmissions per bit or per quality-of-experience; neither metric captures the clinical value of a medical transmission. Diagnostic utility vanishes below a modality-specific acceptability floor rather than degrading gracefully, the deadline is determined by triage acuity rather than by the network, and a missed finding is far costlier than a false alarm. A per-bit clearing price therefore disadvantages the node that has expended local compute to produce a compact, diagnostically sufficient stream. We propose SemAuc, a triage-aware truthful mechanism for medical-image admission over a rate-splitting uplink, in which the shared semantic knowledge base rides the common stream, and case-specific residuals ride private streams. SemAuc filters tiers below the diagnostic floor and beyond the clinical deadline, reserves a regulated-price lane for life-threatening cases, and allocates remaining capacity through a single-parameter contestable auction whose bid-independent pre-selection step satisfies the conditions of Myerson’s lemma. The contestable lane is dominant-strategy truthful, individually rational, near-linear in the number of nodes, and achieves a constant-factor density-greedy welfare guarantee; the clinical lanes follow from triage policy without disturbing these properties. Diagnostic value is grounded by an offline kernel fitted on BraTS and CheXpert. On a Rayleigh-faded uplink at two hundred contending nodes, SemAuc preserves the high-acuity diagnostic service-level objective where bit-centric benchmarks fail, and tracks the offline optimum. Full article

Multi-access edge computing (MEC) enables computation-intensive perception and decision-making tasks in vehicular networks to be offloaded to nearby edge servers. Existing approaches usually fix the artificial intelligence (AI) inference model, overlooking how model selection jointly affects latency, energy consumption, and service reliability. We propose a hierarchical model selection and control (HMSC) framework based on deep reinforcement learning (DRL) for MEC-assisted vehicular networks. The framework couples a vehicle-layer MAPPO component that provides a communication interface representation for subchannel assignment and energy accounting with a centralized MEC-layer soft actor-critic (SAC) agent that, under SDN orchestration, adaptively selects lightweight or high-fidelity AI models and allocates computational resources. Accordingly, the core contribution of this paper lies in MEC-side model-aware computation control under an explicitly defined subchannel-contention abstraction, rather than in physical-layer transmit-power optimization. Both layers are guided by a composite objective that integrates normalized end-to-end (E2E) latency, normalized energy consumption, and a deadline-violation penalty. Using a discrete-time simulation framework, HMSC reduces E2E latency compared with static inference and non-hierarchical DRL baselines and sustains a higher deadline satisfaction ratio (DSR) under constrained uplink throughput and varying traffic loads. The learned policy is load-aware, favoring high-fidelity inference under light load and lightweight inference under congestion; a post hoc analysis using YOLOv5-family accuracy reference further quantifies the inference-quality implications of this adaptive selection behavior. These results show that coordinated MEC-side control of AI model selection and computation, under a shared deadline-aware objective, provides a robust latency–energy trade-off for MEC-assisted vehicular networks. Full article

Network Intrusion Detection Systems (NIDS) face a trade-off between detection accuracy and computing efficiency, particularly in the edge environment with strict real-time requirements and limited resources. Current approaches rely on complicated models, which are computationally demanding, or simple ones, which sacrifice detection performance. To deal with this issue, we propose a lightweight cloud-edge cooperative two-stage NIDS architecture, which separates the real-time detection and detailed classification. At the edge, a decision tree based on feature selection is used for rapid binary classification by using only the top 10 most informative features, thus efficiently screening out abnormal traffic with minimum processing cost. Meanwhile, the cloud server identifies attack classification accurately by using a hybrid CNN-BiLSTM-Attention model to capture the spatial structures, temporal relationships, and semantic relevance. This hierarchical design effectively balances detection performance and system efficiency. Experiments conducted on UNSW-NB15, NSL-KDD, and CIC-IDS2017 datasets indicate that our suggested scheme can obtain competitive performance both at the edge and in the cloud. The edge model obtains binary classification accuracy of 86.04%, 95.27%, and 99.01%, respectively, with very low processing cost (less than 100 FLOPs per sample). The cloud model achieves multi-class accuracy of 92.23%, 97.18%, and 98.66%, respectively, with AUC values higher than 0.98. The hierarchical cloud–edge collaborative design provides an efficient and accurate solution for intrusion detection under resource-restricted situations. Full article

The rapid development and expansion of the Internet of Things (IoT) ecosystem require managing increasing computational demands with the aid of advanced hierarchical architectures. The integration of a complete four-tier hierarchy—including Mobile, Edge, Fog, and Cloud layers—and the priority requirements are being overlooked in the literature despite the effort of existing studies. The goal of this study is to fill these gaps by proposing a novel, context-aware task-offloading framework designed for multi-dimensional ecosystems involving multi-server and multi-application environments. A targeted biomimetic approach is utilized at the core of this research. The decentralized foraging behavior of biological swarms is translated into a concrete engineering solution. This solution is designed specifically for computational offloading and resource management. To achieve this, a “Learning-guided Meta-heuristic” hybrid model is developed. Within this framework, bio-inspired Ant Colony Optimization (ACO) is directly integrated with an Upper Confidence Bound (UCB)-inspired exploration mechanism. Natural, pheromone-based imitation is solely relied upon by traditional biomimetic algorithms. In contrast, higher-order cognitive learning is fully incorporated by this hybrid synergy. Consequently, underlying system dynamics are adaptively learned. Local minima traps are also successfully avoided. This avoidance is achieved by dynamically selecting the optimal layer for each individual task. Both energy consumption and latency are optimized simultaneously. Meanwhile, strict operational feasibility is ensured through a dynamic penalty-based mechanism. Battery and deadline constraints are explicitly handled by this mechanism. Extensive simulations demonstrate the superiority of the proposed UCB-ACO model over state-of-the-art meta-heuristics, including Particle Swarm Optimization (PSO), Gray Wolf Optimizer (GWO), ACO, Artificial Bee Colony Optimization (ABO), and Non-dominated Sorting Genetic Algorithm II (NSGA-II). The findings reveal that the proposed framework outperforms the methods compared by achieving 22.5% lower latency and 23% lower energy consumption. This study effectively maps the current literature and then introduces a pioneering solution for next-generation resource management in distributed computing. Full article

We tackle the problem of secure and private feature selection by homomorphically evaluating features’ information gains over the encrypted data of horizontally partitioned private datasets owned by edge IoT (Internet of Things) devices. In the process, we use a powerful cloud server to do the bulk of the costly homomorphic encryption aggregations. We proceeded with secure and private federated learning (training) and Machine Learning (ML) classification over the selected features in the same environmental settings (context). In the process, the participants interact with each other under strict security, privacy, and efficiency requirements. To this end, to each participant’s interaction we provide confidentiality, integrity, and authenticity (CIA) by signing its hashed contents with the corresponding participant’s private key. We assure consistency among interactions by introducing timestamps and linking them with the hashed content(s) of the preceding interaction(s). Those linked blocks of hashed content(s) from each interaction of participants while running the protocols produce the so-called chain-of-blocks (COB) structure, which will be utilized to detect malicious edge IoT dataset owners, unauthorized participants, and network errors. The security of the proposed protocols is proven through rigorous mathematical modeling. Extensive experimental evaluations over benchmark datasets give an advantage to our secure protocols ranging from several times to orders of magnitudes w.r.t to the state of the art in terms of computation and communication costs, as well as security and privacy characteristics. Moreover, since the utilized underlying cryptographic techniques are resilient to quantum computer attacks, the proposed algorithms are applicable to the post-quantum world. Full article

Edge artificial intelligence (Edge AI) requires real-time inference under stringent constraints on computation, memory, energy, and connectivity. Although training can be offloaded to servers, efficient, high-capacity inference and rapid on-device adaptation remain central challenges. Cloud-based inference offers substantial computational power but depends on connectivity, latency, privacy, and reliability conditions that edge deployments cannot always guarantee. Classical model-compression methods—including quantization, pruning, distillation, and neural architecture search—have extended the feasibility of on-device inference, yet they leave largely unchanged the fundamental cost of the linear-algebraic, sampling, and optimization primitives that dominate modern deep learning. Quantum computing has therefore been proposed as a complementary accelerator for selected AI workloads, with theoretical advantages in linear systems, singular value decomposition, sampling, kernel evaluation, and optimization. This review surveys the emerging field of quantum-accelerated AI for edge systems under a hybrid architectural premise: edge devices remain classical, while quantum processors operate as remote, cloud, MEC, or near-edge accelerators. We synthesize advances across quantum learning models, hybrid optimization methods, hardware and deployment architectures, and quantum-inspired approaches suitable for constrained devices. We also assess the practical barriers that currently separate asymptotic quantum advantage from deployable edge intelligence, including data loading, measurement overhead, noise, latency, and benchmarking gaps. Finally, we outline a staged research roadmap from near-term hybrid workflows to fault-tolerant and integrated quantum-edge architectures. Full article

Databases serve as critical repositories of digital evidence in criminal investigations, and the recoverability of deleted data is a key determinant of forensic success. Microsoft SQL Server, one of the most widely deployed relational database management systems, has been the subject of multiple forensic studies examining how deleted records persist in physical database files across different acquisition methods. A previous study established a reference baseline using SQL Server 2008 and 2017, demonstrating that the Database Shrink operation causes version-specific and method-specific behavior: under logical collection with Shrink applied in SQL Server 2017, unallocated deleted data becomes fully initialized, rendering recovery impossible—a pattern not observed in SQL Server 2008 or under physical collection in either version. With the release of SQL Server 2025, the most significant architectural update to the platform in a decade, it remained unknown whether these forensic behaviors persist in the latest version. This study replicates the experimental design of in a controlled SQL Server 2025 environment, applying the same deletion scenario (DELETE command without conditions), the same two acquisition methods (logical and physical collection), and the same Shrink condition. The results demonstrate that SQL Server 2025 does not reproduce the version-specific initialization behavior observed in SQL Server 2017: across all four experimental conditions, deleted data residue in unallocated page space remains recoverable, indicating a fundamental change in the interaction between the Shrink operation and the logical collection mechanism. This recoverability is a double-edged property: while it benefits forensic investigators by preserving deleted evidence, it simultaneously represents a data-sanitization risk from a security and privacy standpoint, as deleted records are not reliably erased. These findings provide updated forensic guidance for digital investigators operating in contemporary SQL Server environments. Specifically, the results inform acquisition-method selection in real-world investigations where a suspect may have deleted records and where only a logical backup (.bak) is available to investigators. Full article

As a distributed approach to Artificial Intelligence (AI) model construction over wireless networks, federated learning (FL) based on multi-device collaborative training can protect data privacy, as well as increase the computing load of local model updates. In contrast, split learning (SL) with proper model splitting can adapt to the computation and transmission capabilities among devices. In this paper, while taking advantage of FL and SL, we concentrate on a semi-decentralized hybrid federated split learning (SD-HFSL) framework, in which we surpass the limitations of a single central server and allow the shared split models to be aggregated among multiple edge servers. To verify the importance of latency optimization for training efficiency, we analyze the convergence performance of SD-HFSL while jointly considering the limited computation and communication resources. Then, aiming at maximizing the long-term training efficiency, we propose an online optimization problem that includes local model splitting and device association. Considering that the training latency is unknown to the system a priori, a context-aware online training algorithm with sublinear regret is proposed based on the framework of contextual multi-armed bandit (CMAB), where the edge servers can observe the context information of device sites for latency estimation, followed by the iterative optimization based on the evaluated information in different contexts. Experiments on several neural network models show that the proposed algorithm reduces training latency and improves test accuracy compared with the selected benchmarks. Full article

Mobile edge computing (MEC) enables computation-intensive and latency-sensitive tasks to be offloaded from mobile devices to nearby edge servers. Most existing MEC task offloading studies formulate offloading as a selection problem over a fixed or fully available set of candidate servers, which is restrictive in heterogeneous MEC scenarios with task-node eligibility constraints. Under such constraints, a task can be processed by an edge server only when task attributes, service requirements, link conditions, and node states jointly satisfy the corresponding eligibility conditions. The feasible action set therefore varies over time, while offloading decisions are further coupled with edge-node-side queue competition and long-term load evolution. To address this problem, this paper proposes Resource-oriented Scheduling Coordination (RoSCo), a load-aware task offloading method with scheduling-level constraint handling for eligibility-constrained MEC systems. In this paper, scheduling coordination refers to the joint use of feasible-action control, priority-aware edge-node service-order modeling, and load-responsive feedback within the task offloading decision process; it does not denote inter-server communication, task aggregation, federated model aggregation, or a distributed coordination protocol. RoSCo constructs a dynamic feasible action set, applies eligibility-aware action masking to exclude infeasible offloading actions, incorporates priority-aware edge-node service-order information to characterize queueing competition among heterogeneous tasks, and designs a load-responsive reward to guide congestion mitigation and load balancing. A dueling double deep Q-network (D3QN) is adopted as the value-learning backbone, while the main methodological contribution lies in embedding task-specific feasible-action control, priority-aware node-side queue information, and load-responsive feedback into the constrained offloading process. Simulation results show that RoSCo reduces the task drop rate and edge-node load imbalance while maintaining competitive task completion delay and energy consumption, especially under high-load and sparse-eligibility conditions. Full article

Precision in estimating the ripeness of fruits is critical in quality control and minimizing losses in supply chains of agricultural produce following harvesting. Conventional ripeness assessment techniques tend to be destructive, time-consuming and unsuited to monitoring in real-time. In order to avoid these drawbacks, this research suggests a cloud-integrated smart electronic nose (E-nose) system to predict fruit ripeness in a non-destructive and real-time manner. The system uses a low-priced, non-selective gas sensor array with an ESP8266-based Internet of Things (IoT) board to record volatile organic compound (VOC) signatures released at various maturation phases of fruits. The obtained sensor data will be sent to a cloud server to be preprocessed centrally and classified using machine learning, thus reducing the computational needs at the edge. There is a collection of 953 samples of the unripe, ripe, and rotten stages of banana under controlled conditions. Several supervised machine learning algorithms are tested, and methods of ensemble boosting proved to be more effective. The Light Gradient Boosting Machine (LightGBM) is the most accurate in terms of classification of 96.50% and weighted F1-score of 96.49%. The confusion matrix analysis shows that the majority of misclassifications are observed among the neighboring stages of ripeness, indicating the gradual biochemical changes. The system is practically applicable as visualization of the predicted ripeness levels occurs in real time via a mobile application. The suggested model provides a scalable, low-cost, and smart solution to precision agriculture, which can allow efficient, automated, and non-destructive measurement of fruit quality. Full article

Federated Learning (FL) is a machine learning technique that preserves data privacy and security by training models directly on decentralized edge network devices. This generates substantial communication overhead due to the repeated exchange of model updates across numerous edge network devices. Quantization has tackled this challenge by reducing communication overhead and computational costs by quantizing model updates. Although selecting the most suitable quantization level to balance communication efficiency and model accuracy is challenging, failing to achieve this balance results in excessive compression, leading to accuracy degradation due to the lossy nature of the quantization technique. This challenge was tackled in this paper via a Quantile-based lossless compression method named Pcodec, which implements lossless compression in the FL context. Pcodec is a Quantile-based lossless compression algorithm designed for numerical data that utilizes mode identification with delta encoding and binning, where binning groups similar values into entropy-coded bins and stores the exact offset within each bin, thus achieving high compression ratios and efficient processing speed. Using MNIST and CIFAR-10 datasets and models such as CNN and ResNet18, we demonstrate that Pcodec achieves up to 58.19% size reduction with no accuracy loss compared to standard quantization methods. The experiments showed that the proposed Quantile-based compression approach in FL reduces up to 2.81× the communication overhead between each server and edge network device while maintaining the accuracy. In comparison to quantization, the Quantile approach reduced the communication overhead by 2.74×, tackling the main challenge of FL context by reducing communication overhead with a remarkably high compression ratio while maintaining the model’s accuracy. Full article

Edge computing on unmanned aerial vehicles (UAVs) enables low-latency wildfire monitoring by performing visual inference onboard; however, practical deployment is constrained by limited labeled data and resource budgets that often preclude reliance on large GPU servers. This work investigates transfer learning (TL) for UAV-based wildfire smoke and flame detection and evaluates its impact on both detection accuracy and edge deployment performance. We introduce the Aerial Fire and Smoke Essential (AFSE) dataset (282 aerial-view images; classes—smoke and fire), compiled from publicly available wildfire footage and FLAME2. Lightweight YOLO models are fine-tuned using heterogeneous (MS COCO) and homogeneous (FASDD) source pretraining and are assessed using mAP@0.5 together with frames per second (FPS), average inference power, energy consumption, and the normalized energy–delay product (EDP) on an edge computing platform. Results show that TL substantially improves detection accuracy on AFSE, achieving up to 79.2% mAP@0.5, while reducing training time, and improving cross-validation stability. On the tested edge platform, TL does not materially change inference speed or energy use, indicating that accuracy gains from TL do not automatically translate to improved efficiency without additional optimization. Among the evaluated lightweight detectors, YOLOv5n achieves the best mAP@0.5 while maintaining the highest edge device throughput, processing images nearly twice as fast as YOLO11n without hardware acceleration. More broadly, the measured throughput and energy differences among lightweight YOLO variants show that edge model selection should be guided by application-specific accuracy, latency, and energy constraints. Full article