Browse Articles

A novel distributed gradient algorithm for composite constrained optimization over directed network

Combining parameter fragmentation and group shuffling to defend against the untrustworthy server in federated learning

Hydrothermal synthesis and optimization of hierarchical copper molybdate nanostructures for photocatalytic degradation of crystal violet and antimicrobial applications

Abstract In this work, the Hierarchical Copper Molybdate Nanostructures (HCM) photocatalyst was synthesized via a hydrothermal method and characterized using XRD, FTIR, UV-vis spectroscopy, SEM, TEM, and EDX analyses, confirming the formation of hierarchical nanocrystalline structures with a direct band gap of ~ 2.4 eV. The photocatalytic performance of HCM was systematically evaluated for the degradation of crystal violet (CV) dye under UV irradiation. Experimental design and process optimization were conducted using response surface method (RSM) and ANOVA, which demonstrated that pH, irradiation time, and catalyst dosage were the most influential variables, whereas dye concentration exerted a minor effect. Under optimized conditions (pH 10, 60 min, 15 mg/L CV, and 20 mg catalyst dosage), more than 99% CV dye degradation was achieved. Kinetic analysis demonstrated that the degradation followed a pseudo-first-order model, while thermodynamic studies indicated that the process is spontaneous, endothermic, and entropy-driven. Mechanistic evaluation confirmed that reactive oxygen species (·OH, O 2· , HOO·) generated through electron–hole separation played a dominant role in CV mineralization. In addition, HCM exhibited significant antimicrobial activity against bacterial and fungal strains, supporting its multifunctional potential. Overall, the findings highlight HCM as a highly efficient, low-cost, and environmentally friendly photocatalyst with promising applications in wastewater treatment and environmental remediation.

Factors influencing older adults’ participation in community volunteering service in China: a logistic regression and fsQCA perspective

Diagnostic significance of rhythmicity in postural hand tremor

Abstract Rhythmicity is an important feature of tremor that is widely viewed as having diagnostic significance. We hypothesized that rhythmicity might be a generic function of tremor severity, reflecting greater oscillatory entrainment of motor pathways. Postural hand tremor and forearm electromyograms were recorded from 49 controls, 78 Parkinson patients, and 133 essential tremor patients. Rhythmicity was quantified in terms of approximate entropy and three measures of cycle-to-cycle frequency variability: tremor stability index, cycle-to-cycle variability, and power spectral bandwidth. Physiological tremor was less rhythmic than Parkinson tremor and essential tremor, but the two pathological tremors did not differ significantly. Hand tremor amplitude and forearm electromyogram-hand tremor coherence were moderate statistical predictors of rhythmicity in both pathological tremors. Adding a 1-kg weight to the hand had little effect on the rhythmicity metrics, except for a moderate reduction in the approximate entropy of physiological tremor. We conclude that these four rhythmicity metrics are not helpful in distinguishing postural tremors in Parkinson disease and essential tremor. The moderate correlations of rhythmicity with tremor amplitude and electromyogram coherence suggest that rhythmicity of pathological tremors is largely a generic reflection of oscillatory neuronal entrainment. Comparisons of tremor rhythmicity in pathological conditions must control for tremor amplitude and electromyogram coherence.

A nine-level switched-capacitor multilevel inverter based on ANPC topology with optimized component count and suppressed charging current

Abstract This paper presents a nine-level switched-capacitor (SC) based on ANPC inverter structure, where efficiency, compactness, and reliability are essential. Unlike conventional ANPC-based and previously reported SC topologies, the proposed structure suppresses the hard charging issue in SCs by inserting an inductor in the charging path, effectively reducing peak charging currents and associated switch stress. The inverter achieves natural self-starting and self-voltage-balancing of capacitors without additional circuitry, while requiring fewer active switches in the conduction path. These features lead to reduced component count, and minimized power losses. A peak efficiency of 96.9% is obtained at a rated power of 0.3 kW. Experimental verification confirms the theoretical predictions, demonstrating stable capacitor voltage under varying load conditions and modulation indices. Comparative analysis highlights the superiority of the proposed topology over existing SCMLIs in terms of efficiency, charging current suppression, and reduced component requirements.

Handcrafted MRI radiomics of enlarged perivascular spaces and machine learning predict cognitive impairment and sleep disturbance in young adults

Personalized skill transfer optimization in swimming training through multi-agent reinforcement learning driven digital twin environments

Abstract Traditional swimming training methodologies face inherent limitations in providing personalized, adaptive, and scalable training solutions that accommodate diverse learning patterns and individual athlete characteristics. This research introduces a novel framework integrating multi-agent reinforcement learning with digital twin technology to create an intelligent swimming training environment capable of delivering personalized skill transfer optimization through meta-learning strategies. The proposed system addresses conventional training limitations by providing adaptive, data-driven training recommendations that evolve based on individual swimmer characteristics and performance dynamics. The multi-agent architecture enables simulation of complex training scenarios while incorporating real-time feedback mechanisms that continuously refine training strategies. Key contributions include: (1) development of a comprehensive digital twin swimming environment modeling biomechanical and hydrodynamic processes, (2) implementation of multi-agent reinforcement learning algorithms for personalized sports training, (3) integration of meta-learning based skill transfer optimization enabling efficient knowledge transfer across swimmers and contexts, and (4) experimental validation demonstrating improved training efficiency and performance outcomes. Experimental results show 34% faster convergence rates and 22% higher final performance scores compared to baseline methods, with 2.7× faster skill acquisition rates and 89% retention rates over extended periods. The framework demonstrates robust adaptation capabilities across diverse swimmer populations while maintaining computational efficiency and system stability.

The arrangement of anisotropic spin couplings can optimize sensitivity of the cryptochrome radical pair to the direction of geomagnetic field

Abstract Sensing of the geomagnetic field direction by many living organisms is commonly thought to involve radical pairs, such as those formed photochemically between the flavin and tryptophan radicals in the cryptochrome proteins. Previous theoretical studies have shown that strongly axial hyperfine couplings in the cryptochrome radicals greatly enhance the formation of a signaling state of the protein when the magnetic field is directed perpendicular to the hyperfine axis of either of the radicals. However, further analysis led to the conclusion that sharpness of detecting those magnetic directions is strongly suppressed by the inter-radical electron spin coupling. Here, we perform theoretical simulations of the compass function for a set of arrangements of the intra- and inter-radical spin couplings in the idealized cryptochrome radical pair, and find certain arrangements that preserve the sharpness in detecting the direction of the geomagnetic field. One particular arrangement, with the hyperfine axes of the radicals orthogonal to the symmetry axis of inter-radical coupling, provides even sharper field-direction sensitivity than that contributed solely by the anisotropy of the hyperfine coupling.

Lightweight deep learning model with spatial attention for accurate and efficient breast cancer prediction

Automated identification of contextually relevant biomedical entities with grounded LLMs

Abstract This study investigates the effectiveness of different large language models for automated biomedical entity annotation in research articles with a focus on contextualized and grounded results. A 4-step generative workflow iteratively generates and refines entity candidates by considering a metadata schema for context and agentic tool use for validation with the PubTator 3 data base. The precision of this flow was assessed with a random effects meta-analysis after face-to-face interviews with authors of six papers from the Collaborative Research Center (CRC) 1453 “NephGen”. With an overall precision of 91.3%, the selected models provide qualitatively valuable annotations, with models GPT-4.1, GPT-4o Mini, and Gemini 2.0 Flash showing the highest precision. While GPT-4.1 and Gemini 2.0 Flash excelled in the total number of correct annotations, GPT-4o Mini and Gemini 2.0 Flash were fastest and most cost-effective. Large variations in annotation count and the conflation of publication and dataset-specific annotations highlight that human review ("human-in-the-loop") is still important. The results further highlight the trade-offs between precision, total number of correct annotations, cost, and speed. While quality is paramount in collaborative research settings, cost-effectiveness could be more critical in public implementations.

Transformer-augmented dual-branch siamese tracker with confidence-aware regression and adaptive template updating

Abstract Visual object tracking using Siamese networks has proven effective by matching a reference target with candidate regions. However, their performance is limited by static templates, insufficient context modeling, and weak multi-level feature integration, especially under occlusion, background clutter, and appearance variation. To address these limitations, we propose TSDTrack, a transformer-augmented Siamese tracker designed for quality-aware and robust tracking. Our framework employs a ResNet backbone to extract multi-scale hierarchical features, which are fused using a transformer-based module that applies global attention to enhance semantic and spatial consistency. The prediction head consists of two branches: a confidence aware branch (CAB) that assesses the confidence of classification responses, and a regression distribution learning (RDL) branch that models bounding box localization as discrete probability distributions, improving precision under uncertainty. Furthermore, we introduce a confidence-gated template update strategy that selectively refreshes the target representation based on the CAB score, enabling adaptive appearance modeling while avoiding drift. Experiments on LaSOT, GOT-10k, OTB100, and UAV123 demonstrate that TSDTrack achieves state-of-the-art performance in both accuracy and robustness, achieving 55.5% success on LaSOT, 67.5% AO on GOT-10k, 71.6% AUC on OTB100, and 66.4% success on UAV123, outperforming recent transformer-based and Siamese trackers.

Scalable privacy-preserving data analytics for IoMT via FHE and zk-SNARK-enabled edge aggregation

Implementation and validation of the F4aT laboratory for flow in rough fractures

Abstract Accurate characterization of fracture hydraulics is crucial for optimizing subsurface systems, notably geothermal energy extraction where high flow rates are essential for efficient energy production. The precise transition from linear to nonlinear fracture hydraulics at already moderate flow rates is still undefined, due to the complexity of fracture roughness, where the influence of various roughness parameters and the comparability of individual rough fractures are still unclear. Here, we introduce the Forced Fluid Fracture Flow and Transport Laboratory (F 4 aT-Hydraulic Laboratory), a novel experimental laboratory designed to address this knowledge gap. It focuses on a comprehensive workflow encompassing high-resolution measurements of the rock surface roughness and the experimental investigation of fracture hydraulics at a large range of flow rates ( $$0.05 < Re < 100$$ ). A unique feature of the F 4 aT-Hydraulic Laboratory is its ability to conduct systematic and stochastic investigations of roughness-hydraulic interactions through 3D printed fracture replicas with defined, statistically varied fracture roughness. In this study, we present the developed workflow in detail, provide benchmarking experiments against analytical solutions, and demonstrate the ability to measure roughness effects on the transition from linear to nonlinear hydraulic regimes at already moderate flow rates ( $$Re \approx 10$$ ).

New experimental configuration for investigation of debris accumulation effect on local scour at bridge pier and abutment

Process optimization of microwave drying for rice based on response surface methodology

To optimize the microwave drying process of paddy rice and improve its quality, the effects of hot air temperature, microwave power, and grain layer thickness on the post-drying potassium content, calcium content, vitamin B₁ content, and free fatty acid content of rice were investigated. Using response surface methodology, an experimental scheme based on the Box-Behnken design was constructed to analyze the influence of these three factors on the four nutritional and biochemical indicators. A multi-index optimization model was established and validated. The results showed that all response indicator models were statistically significant. The optimal process parameters were determined as follows: hot air temperature of 52.47°C, microwave power of 20 kW, and grain layer thickness of 2.78 cm. The corresponding predicted values were potassium content of 3724 mg/kg, calcium content of 113.7 mg/kg, vitamin B₁ content of 0.290 mg/100g, and free fatty acid content of 21.5 mg/100g. Validation experiments demonstrated that the dried rice under these conditions exhibited excellent quality, with relative errors between predicted and measured values below 3%, indicating the reliability of the model and the significant improvement in rice drying quality achieved by the optimized process. Furthermore, visual process reference charts were developed to provide a theoretical basis for adjusting process parameters in practical production.

EvoThy-Net: an evolutionary encoder-decoder network for thyroid nodule segmentation in ultrasound imaging

A physics-inspired memory-augmented deep learning framework for magnetic core loss prediction

Accurate prediction of magnetic core loss is a key challenge for improving the efficiency and reliability of power electronic systems. Traditional empirical models such as the Steinmetz equation are only applicable to sinusoidal steady-state conditions and struggle with the complex non-sinusoidal waveforms and variable operating conditions in modern power electronics. While existing deep learning methods have shown improvements, they still face fundamental limitations in handling the nonlinear mismatch between B(t) and H(t) waveforms, coupling of multi-scale loss mechanisms, and generalization under extreme operating conditions. This paper proposes an Enhanced Memory Augmented Mamba (EMA-Mamba) model that achieves breakthrough progress in magnetic core loss prediction. It utilizes a state-space memory augmentation mechanism that stores and retrieves typical magnetization patterns through a trainable external memory matrix, endowing the model with a capability similar to the “magnetic memory” of magnetic materials, effectively solving the gradient vanishing problem in long sequence modeling. Combined with an attention-guided intelligent feature selection mechanism, it adaptively identifies critical turning points in hysteresis curves through a Top-K strategy, fundamentally solving the temporal mismatch problem between B(t) and H(t) waveforms. Finally, through a physics-constrained multi-objective optimization framework, it achieves decoupled modeling of hysteresis loss, eddy current loss, and residual loss through loss function combination, overcoming the optimization difficulties caused by data spanning six orders of magnitude. Experiments on the MagNet dataset containing 10 materials and over 150,000 data points show that EMA-Mamba achieves an average prediction error of 4.50% and a coefficient of determination of 99.9947%, reducing error by 34.2% compared to state-of-the-art baseline methods, with a 36.2% reduction in 95th percentile error under extreme conditions. The model demonstrates excellent temperature robustness and cross-material generalization capability, providing a reliable theoretical tool for intelligent design and optimization of magnetic components.

Tribological characteristics of composite brake pads under variable load and speed

Abstract This study investigates the tribological performance of fiber-reinforced composite brake pad materials fabricated using a compression molding technique. The work focuses on evaluating the influence of applied load (10–30 N) and rotational speed (200–1000 rpm) on the coefficient of friction (COF) and wear rate of the developed samples. Experimental tests were conducted using a pin-on-disc tribometer under controlled laboratory conditions to simulate braking contact. The results revealed that both parameters significantly affect the friction and wear behavior of the composites. At lower speeds (200–400 rpm) and loads (10–20 N), the COF remained relatively stable, ranging from 0.63 to 0.72, with a low wear rate below 0.85 mg/N, due to the formation of a compact tribo-film that protected the surface from severe abrasion. As the load and speed increased to 30 N and 800–1000 rpm, the COF increase to 0.795, and the wear rate increased to 1.065 mg/N, indicating the breakdown of the protective layer and the predominance of abrasive and adhesive wear mechanisms. Microscopic analysis using FESEM and EDS confirmed fiber pull-out, particle fragmentation, and localized matrix softening as the main surface features under severe conditions. These findings demonstrate a direct correlation between frictional stability and wear resistance, highlighting that the balance between operating load and sliding speed plays a crucial role in the durability and performance of composite brake pads.

Melatonin alleviates chronic intermittent hypoxia-induced gastric mucosal injury via attenuation of oxidative stress and JNK-mediated apoptotic signaling in rats

Background To investigate the mechanism of chronic intermittent hypoxia on gastric injury in rats and the intervening effect and possible mechanism of melatonin. Methods Forty-eight male Wistar rats were randomly divided into normal control, intermittent hypoxia, and melatonin treatment groups. Subgroups (n = 4 per time point) were treated for 2, 4, 6, and 8 weeks. Gastric tissue morphology, gastric juice pH, pepsin levels, oxidative stress markers (MDA and SOD), and the expression of JNK and apoptosis-related genes (Bax, Bcl-2) were assessed. Results The intermittent hypoxia group exhibited significant gastric mucosal damage, decreased pH, increased pepsin, elevated MDA, reduced SOD, and upregulation of JNK and Bax/Bcl-2 mRNA ratio. Melatonin treatment markedly alleviated these pathological and molecular changes compared to the intermittent hypoxia group ( P  < 0.05). Conclusion Chronic intermittent hypoxia induces gastric mucosal injury, which is associated with oxidative stress imbalance and activation of JNK-mediated apoptotic signaling. Melatonin exerts a protective effect by enhancing antioxidant capacity and suppressing the JNK-Bax/Bcl-2 pathway.