Search Results (55)

With the continuous expansion of cross-border agricultural supply chains, reliable food-safety risk screening under decentralized data ownership, heterogeneous sensing equipment, and jurisdiction-dependent, non-stationary regulatory limits pose a fundamental challenge to intelligent inspection systems. To address these limitations, this paper proposes an LLM-assisted regulation-aware heterogeneous multimodal federated learning framework that establishes a three-tier decoupling architecture. First, an available-modality and quality-adaptive fusion mechanism dynamically weights RGB images, near-infrared (NIR) spectra, and environmental time-series while enforcing strict zero weighting on unavailable modalities and continuous suppression of degraded observations. Second, a regulation-independent state-prototype collaboration and personalized federated learning strategy is developed on a unified continuous concentration scale, enabling cross-institutional transfer of physical contamination representations while avoiding the use of jurisdiction-dependent compliance labels to define shared prototype groups and preserving client-specific adapters against non-IID drift. Third, offline large language model (LLM) parsing transforms unstructured multi-jurisdictional trade regulations into machine-executable rule bases, which are coupled with an uncertainty-aware risk calibration module to dynamically determine exceedance probabilities and reinspection recommendations without retraining front-end sensing encoders when regulatory standards shift. Experimental evaluations on a multisource maize and aflatoxin B1 screening dataset in a workstation-based federated simulation demonstrate that the proposed framework achieves a mean absolute error (MAE) of $0.054\pm 0.002$ and a root mean square error (RMSE) of $0.078\pm 0.003$, while elevating the area under the curve (AUC) and F1-score to $0.862\pm 0.008$ and $0.833\pm 0.009$, respectively, reducing the false negative rate (FNR) to $0.121\pm 0.007$ ($p<0.001$). Under severe client heterogeneity and missing-modality scenarios, the proposed approach maintains an AUC of $0.823\pm 0.010$ and an F1-score of $0.796\pm 0.011$. Ablation and ownership-level analyses assess robustness, calibration (expected calibration error of 0.032), and regulatory adaptation with fixed sensing encoders, providing an end-to-end, traceable decision-support paradigm for international agricultural trade screening. Full article

Distributed sensing systems increasingly rely on geographically distributed devices that collect data under different operating conditions. Federated continual learning (FCL) enables a global model to be trained across distributed sensing clients whose local data evolve over time, while preserving previously acquired knowledge. Existing FCL studies have mainly addressed continual task evolution or class-based client heterogeneity, while federated domain-incremental learning has investigated adaptation to changing domains under a shared label space. Less attention has been paid to scenarios where continual task evolution and domain-shifted client participation occur simultaneously, leading to coupled temporal and spatial interference. In this paper, we study a dynamic-domain federated continual learning setting, where new clients share the same label space with source clients but introduce shifted feature distributions during training. This setting reveals a challenging form of spatio-temporal catastrophic forgetting: the model must learn evolving class-incremental tasks while preventing new-domain knowledge injection from degrading source-domain performance. To address this problem, we propose STP-FCL, a coordinated client–server framework. On the client side, STP-FCL preserves historical task semantics and source-domain discriminative structures through source-aware prototype-space optimization, reducing representation drift under evolving tasks and domains. On the server side, STP-FCL designs domain-projected aggregation to remove conflicting components from new-domain updates and adaptively control their fusion weight. Experiments under dynamic-domain federated continual learning scenarios show that STP-FCL achieves stronger source-domain knowledge retention and higher mixed-domain accuracy than representative baselines, with a trade-off in target-domain accuracy. Full article

Heterogeneous client distributions and differential-privacy noise can slow learning. This study proposes time-varying privacy-scheduled differentially private federated learning (TVP-DPFL), combining a public Gaussian multiplier schedule, bounded reliability-weighted signals, and control-variate drift correction. Inverse-square normalization gives the time-varying and fixed schedules identical final Gaussian Rényi differential-privacy costs. On the Open University Learning Analytics Dataset (OULAD) and the Predict Students’ Dropout and Academic Success dataset from the University of California, Irvine repository (UCI-Dropout), the three-seed endpoint results are close to a differentially private variant of Stochastic Controlled Averaging for Federated Learning (DP-SCAFFOLD) under the common 30-round aggregate-update budget of ε = 8. The ten-seed scheduling comparison shows earlier attainment of the 95% validation target: 11.1 versus 12.4 rounds on OULAD and 16.4 versus 18.7 rounds on UCI-Dropout (two-sided Wilcoxon p = 0.002). The normalized area under the validation learning curve over 30 rounds (AULC30) also favors the time-varying schedule. The privacy guarantee applies to aggregate updates conditional on fixed bounded weights and control states, with auxiliary metadata and control-state releases requiring separate protection and accounting. Communication efficiency is measured by coordinated rounds and model/control-vector payloads on retrospective public-data silos. Full article

Quantum federated learning (QFL) enables collaborative training of variational quantum circuits across decentralized clients, but client drift under non-IID data distributions degrades standard Federated Averaging (FedAvg). Existing aggregation methods either introduce tunable hyperparameters that depend on unknown data heterogeneity or incur structural costs such as order-dependent error propagation. We propose Kernel-Weighted Aggregation (KWA), a hyperparameter-free aggregation strategy that constructs a non-parametric kernel density estimate over the received client parameter vectors at each communication round. Clients in high-density regions receive greater voting power, and isolated clients are automatically attenuated. The kernel bandwidth is the median of all pairwise squared distances. We evaluate KWA against six state-of-the-art QFL methods and five classical robust baselines on five binary MNIST digit-pair tasks under IID, Dir($0.5$), and Dir($0.1$) distributions with a unified four-qubit benchmark. Under IID data, KWA recovers the performance of FedAvg. Under Dir($0.5$), KWA attains the highest average accuracy ($0.7120$), followed by FedAvg ($0.7093$). The margin at $N=4$ clients is not statistically significant. It grows to $+0.023$ at $N=16$ in the client scaling experiment. Under this distribution, the five classical robust baselines also rank below FedAvg. Under Dir($0.1$), KWA exceeds FedAvg by $0.028$ on average, with a pooled paired t-test $p=0.068$. Larger-circuit and three-class experiments show that the advantage does not yet transfer consistently beyond the four-qubit binary setting. KWA thus provides a hyperparameter-free aggregation strategy that recovers FedAvg under IID conditions, adapts to client drift, and requires no manually tuned hyperparameters. Full article

Background: Skin cancer encompasses a diverse range of malignancies and remains a significant global health challenge. Accurate machine-learning-assisted diagnosis can substantially improve patient outcomes through early detection and timely clinical intervention. Federated Learning (FL) enables privacy-preserving collaborative model training across multiple healthcare institutions while ensuring that sensitive patient data remain decentralized. However, deploying advanced architectures such as Vision Transformers (ViTs) in clinical environments is challenging due to the high computational demands of self-attention mechanisms. Methods: This work proposes FedSwin-LHTP, an efficient federated learning framework for skin lesion classification that integrates a Swin Transformer backbone with a Lightweight Hessian-Inspired Token Pruning (LHTP) mechanism. LHTP estimates token importance using a second-order Taylor approximation around converged local model parameters to identify less informative patch tokens, enabling the early pruning of redundant representations without explicitly computing the Hessian matrix. Furthermore, the framework incorporates the FedProx optimization objective to mitigate client drift under heterogeneous non-IID data distributions. The proposed framework is evaluated on the HAM10000 and ISIC datasets under realistic non-IID federated settings. Results: Experimental results demonstrate stable convergence, effective knowledge aggregation, and robust diagnostic discrimination across distributed clients. By adaptively pruning approximately 60% of Stage-1 tokens, the proposed framework substantially reduces the computational burden of local transformer processing while maintaining high multiclass classification performance, achieving an accuracy of up to 96.1% on the evaluated datasets. Conclusions: These results highlight the potential of FedSwin-LHTP as a practical, privacy-preserving, and resource-efficient solution for collaborative healthcare intelligence. Full article

Federated continual learning (FCL) must preserve useful historical knowledge while learning from evolving and statistically heterogeneous client streams. However, indiscriminate replay can retain client-specific bias, noise, or task conflicts and thereby cause negative transfer. In this paper, we propose Adaptive Multi-Factor Accurate Forgetting for Heterogeneous Federated Continual Learning (AMAF-FCL), a selective memory-management framework that jointly assesses replay reliability and adapts the influence of generated historical features. AMAF-FCL achieves this goal by (1) modeling historical knowledge in feature space with a conditional real-valued non-volume-preserving (RealNVP) normalizing flow; (2) combining class-conditional likelihood, predictive uncertainty, and global distribution consistency; and (3) adjusting replay weights according to client–global heterogeneity. The local objective combines a classification loss for learning the current task, a reliability-weighted replay loss for retaining useful historical knowledge, and a feature-distillation loss for limiting drift in the feature representation. In the EMNIST long-task-pool (EMNIST-LTP) benchmark, each client learns six two-class tasks drawn from a client-specific set of handwritten letters. AMAF-FCL achieves 51.2% average accuracy and 7.6% average forgetting, improving over the likelihood-based AF-FCL baseline by 3.7 and 1.5 percentage points, respectively. In the cross-domain digit-and-fashion setting, it reports 71.2% average accuracy and 6.8% average forgetting; with four noisy clients, it obtains 56.1% average accuracy and 9.8% average forgetting. These results indicate that multi-factor reliability assessment and heterogeneity-aware adaptive forgetting improve the balance between useful knowledge retention and harmful-knowledge suppression in heterogeneous FCL. Full article

Federated learning enables privacy-preserving distributed training but suffers from client drift and slow convergence under statistical data heterogeneity. Most existing federated optimization methods address client drift via parameter-space constraints or aggregation-level corrections, while fewer works directly shape the gradient covariance spectral structure of the optimization landscape. This paper analyzes the convergence problem from a spectral perspective, revealing that non-IID data causes spectral diffusion in the gradient covariance matrix and degrades convergence. Guided by random matrix theory, we propose federated spectral regularization (Fed-SR), a computationally efficient method that indirectly constrains spectral spread via gradient norm regularization. Although computing the regularizer gradient requires Hessian vector products, our optimized auto-differentiation implementation avoids storing full Hessian matrices and restricts extra computational overhead to a negligible level. Experiments on CIFAR-10, CIFAR-100, and other benchmarks show that Fed-SR outperforms baselines including FedAvg, FedProx, and SCAFFOLD in non-IID scenarios, reducing communication rounds and improving accuracy and stability. Ablation studies, spectral analysis, and controlled spectral feature manipulation experiments provide consistent empirical evidence showing a strong empirical association between the “spectral concentration” effect and performance gains, offering mechanistic interpretability consistent with our proposed theoretical framework within the tested experimental settings. Full article

Federated learning provides a feasible way to support collaborative vehicle part recognition without sharing raw data across distributed vehicle service nodes. However, real vehicle scenarios exhibit spatiotemporal heterogeneity, where client distributions vary and new vehicle models, parts, and viewpoints emerge over time. Such heterogeneity causes catastrophic forgetting and degrades knowledge retention in federated continual learning. To address this problem, this paper proposes FedACA (Federated Continual Learning with Adaptive Class Anchors), a class-anchor-based federated continual learning method for vehicle part image classification. The method uses a frozen Vision Transformer backbone and introduces scene-aware prompt adaptation, adaptive class anchors, non-parametric global prototype selection, and class-aware prompt fusion. These modules stabilize class semantics, reduce feature drift, and improve cross-client knowledge alignment in a data-local federated setting with raw-data non-sharing; however, FedACA does not provide formal privacy guarantees. Experiments on a self-constructed 67-class multi-view vehicle-part dataset and CIFAR-100 show that, under the task-aware class-masking protocol, FedACA achieves final-stage average accuracies of 97.34% ± 0.72% and 95.71% ± 0.63%, respectively, over three independent random seeds, outperforming the evaluated federated continual learning baselines. These results demonstrate that FedACA improves recognition performance and historical knowledge retention under spatiotemporally heterogeneous federated continual learning settings. Full article

Federated learning enables collaborative model training without centralizing private data, yet its deployment on edge devices remains constrained by communication overhead, heterogeneous data distributions, and privacy leakage from shared model updates. Spiking neural networks offer an energy-efficient alternative to conventional artificial neural networks by transmitting sparse binary events rather than dense activations, but existing federated spiking learning methods still suffer from inefficient gradient exchange and insufficient privacy protection under non-independent and identically distributed data. This paper proposes PriSparK, a privacy-preserving and communication-efficient federated spiking neural learning framework that jointly exploits temporal event sparsity, adaptive client-side spike-gradient compression, and privacy-calibrated aggregation. The core of PriSparK is a novel Event-Sparse Differentially Private Federated Spiking Optimization algorithm, which converts local surrogate gradients into spike-saliency-aware sparse updates, dynamically allocates communication budgets across layers and time steps, and injects calibrated Gaussian noise after clipping in a low-dimensional event subspace. To mitigate accuracy degradation caused by aggressive compression and privacy perturbation, PriSparK further introduces a membrane-aware error-feedback mechanism and a heterogeneity-adaptive server aggregation rule that weights client updates according to spike activity stability and local distribution drift. Experiments on neuromorphic and vision benchmarks, including N-MNIST, DVS128 Gesture, CIFAR-10, and Fashion-MNIST, show that PriSparK achieves competitive or superior accuracy compared with federated artificial neural and spiking baselines while substantially reducing uplink communication. Under strong privacy constraints, PriSparK maintains stable convergence and improves the accuracy–communication–privacy trade-off, demonstrating its potential for privacy-sensitive edge intelligence with event-driven neural computation. Full article

Federated learning (FL) is a promising approach for training distributed machine learning models while preserving clients’ data privacy. However, in real-world FL systems, data are often not independent and identically distributed (non-IID). This heterogeneity can slow convergence, degrade model performance, and increase client drift. To address these challenges, numerous methods have been proposed to mitigate non-IID data effects by optimizing client selection, local training, and model aggregation strategies. Despite their effectiveness in improving performance and efficiency, these methods rarely consider fairness across clients. Improving global accuracy does not guarantee balanced participation, influence, or outcomes, which may lead to biased model behavior across clients. In this survey, we review existing non-IID mitigation methods in FL from a fairness perspective and provide a systematic analysis of their implicit impact on client participation and influence. Unlike prior surveys that treat fairness as a separate research direction, this work analyzes how these methods designed for non-IID mitigation implicitly shape fairness outcomes across clients. Our taxonomy classifies existing methods into three categories—fairness-aware, semi-fairness-aware, and fairness-unaware—based on their design strategies for client selection and model aggregation. Using this taxonomy, we analyze the advantages, trade-offs, and limitations of each category and highlight that mitigating non-IID data does not guarantee fairness across clients. Finally, we identify open challenges and outline future directions, including system-level FL design that jointly considers non-IID mitigation and fairness and the development of standardized fairness evaluation metrics. Overall, this survey aims to provide a structured perspective on the relationship between non-IID mitigation and fairness and support the development of more balanced and scalable FL systems under non-IID conditions. Full article

Federated object detection allows distributed IoT cameras to learn a shared detector without exposing raw images. Its performance is limited by non-IID scenes, unequal device resources, intermittent links, and selfish participation. We propose GO-FedDet, a game-theoretic optimized federated detection framework for heterogeneous IoT application. Client participation is formulated as a Stackelberg game, selection is formulated as an exact-potential resource game, and the equilibrium is embedded into a proximal detection objective. A utility-aligned aggregation balances detection contribution, communication/energy cost, and long-term fairness. We prove the existence of a Stackelberg equilibrium and the finite-improvement convergence of the selection game, and we derive a non-convex convergence bound under client drift and partial participation. Experiments on heterogeneous edge object detection show that GO-FedDet improves accuracy, lowers communication cost, and stabilizes fairness compared with representative federated baselines. Full article

Post-login threats such as device sharing and session takeover motivate continuous authentication with behavioral signals. This paper studies a lightweight keystroke-dynamics pipeline designed for strict cross-session evaluation and browser-side scoring. Using the fixed-text and free-text tracks of the public KeyRecs dataset, we extract compact repetition-level and sliding-window digraph-timing features and train per-user one-vs-rest Logistic Regression verifiers on Session 1 (S1). Thresholds are selected only on S1 and transferred unchanged to Session 2 (S2), preventing test-set tuning and exposing operating-point instability under session drift. Fixed-text achieves S2 AUC mean/median 0.895/0.918 with a half total error rate (HTER) around 0.19, while free-text reaches AUC mean/median 0.884/0.899 with a similar transferred-threshold HTER. Personal thresholds and a pooled-S1 global threshold perform similarly on average, suggesting that global thresholding can simplify deployment without replacing per-user scoring models. A scaler-only warm-up update yields limited and inconsistent gains, showing that mean/variance adaptation alone is insufficient. Finally, compact JSON artifacts and replay-based browser benchmarks demonstrate deterministic client-side scoring with very small per-sample latency. Overall, the results show that useful threshold-free separability does not by itself guarantee stable operating-point transfer under cross-session drift. Full article

Federated Learning (FL) enables privacy-preserving model training in edge and IoT environments. However, in adversarial settings, FL suffers from two key challenges: robustness degradation due to data heterogeneity and poisoning attacks, and runtime instability on resource-constrained devices. Existing work mainly focuses on robustness while overlooking system-level stability. To address this, we propose FedCASKD, a robustness- and stability-aware FL framework. It employs a score-based soft aggregation mechanism to suppress unreliable client updates without requiring a trusted dataset, and introduces a selection-aware bidirectional knowledge distillation protocol to mitigate model drift under Non-IID data. The novelty lies in integrating aggregation and distillation into a unified feedback framework that enhances robustness and stability. Experiments on AGNews and SogouNews show that FedCASKD outperforms baselines under label-flipping attacks and heterogeneous settings. Memory and Out-of-Memory (OOM) tests further demonstrate its superior runtime stability in edge environments. Full article

Rolling bearing fault diagnosis is essential for ensuring the safety and reliability of rotating machinery. Although deep learning-based methods have achieved promising performance, they usually require sufficient labeled data, which is difficult to obtain in practical industrial scenarios where fault samples are scarce and data sharing across sites is restricted by privacy and confidentiality constraints. Federated learning enables collaborative model training without transmitting raw data, but existing federated fault diagnosis methods often degrade under few-shot conditions. Moreover, current federated meta-learning approaches mainly focus on model-level adaptation and lack explicit class-level representation alignment, leading to prototype drift across heterogeneous operating conditions. To address these challenges, this paper proposes ProtoFed, a prototype-enhanced federated meta-learning framework for few-shot rolling bearing fault diagnosis. ProtoFed converts raw vibration signals into time–frequency representations using continuous wavelet transform and performs local episodic learning with prototypical networks. A Global Prototype Calibration mechanism aggregates local class prototypes into stable global prototypes with exponential moving average smoothing, while a Prototype-Distance Aware Aggregation strategy adaptively adjusts client aggregation weights according to local–global prototype divergence. Experiments on the CWRU and Paderborn University bearing datasets under non-IID 5-shot and 10-shot settings show that ProtoFed consistently outperforms standard federated learning, prototype-based federated learning, and federated meta-learning baselines. Under the 5-shot setting, ProtoFed achieves 95.63% and 91.35% accuracy on CWRU and PU, respectively, approaching centralized few-shot upper-bound performance while preserving the federated training paradigm. Full article

Cooperative localization enables teams of robots to maintain better positioning in GNSS-denied environments by sharing state estimates and inter-robot range measurements to reduce the rate of proprioceptive odometry drift. In scenarios without fixed navigation beacons or pre-surveyed reference nodes, each robot functions as both a positioning client and a mobile ranging peer. A critical vulnerability in this architecture is silent fault propagation. A robot with a degraded localization solution may broadcast an incorrect, often overconfident position estimate, corrupting its peers’ localization. Classical Global Navigation Satellite System (GNSS) Receiver Autonomous Integrity Monitoring (RAIM) methods are ineffective in this context because meter-scale inter-robot separations introduce strong geometric nonlinearity and unstable Geometric Dilution of Precision (GDOP), resulting in scattered subset solutions rather than the coherent, biased clusters that RAIM is designed to detect. This paper addresses this vulnerability by proposing a two-stage distributed Fault Detection and Identification (FDI) architecture for peer-to-peer ranging-based cooperative localization. The first stage applies a global chi-square test on Weighted Least-Squares trilateration residuals to detect the presence of a fault. The second stage identifies the faulty robot by computing Leave-One-Out and Leave-Two-Out subset solutions, which are then partitioned using a clustering algorithm. The cluster that exempts measurements from the faulty robot is identified using either a maximum-cardinality or a minimum-variance criterion. A decentralized voting protocol that requires at least two independent corroborations is then employed for network-wide fault declaration. Monte Carlo simulations show that the clustering-based identification method outperforms classical residual-based methods across multiple fault types, with results reported for the planar (2D) case. No single clustering configuration dominates in terms of identification performance across all tested fault conditions, as performance varies with the fault profile. The proposed architecture operates fully in a distributed manner, requiring only the exchange of position estimates, covariances, and binary votes. Full article