Search Results (1,317)

This paper introduces AdamN, a nested-momentum adaptive optimizer that replaces the single Exponential Moving Average (EMA) numerator in Adam/AdamW with a compounded EMA of gradients plus an EMA of that EMA, paired with an exact double-EMA bias correction. This yields a smoother, curvature-aware search direction at essentially first-order cost, with longer, more faithful gradient-history memory and a stable, warmup-free start. Under comparable wall-clock time per epoch, AdamN matches AdamW’s final accuracy on ResNet-18/CIFAR-100, while reaching 80% and 90% training-accuracy milestones ~127 s and ~165 s earlier, respectively. On pre-benchmarking workloads (toy problems and CIFAR-10), AdamN shows the same pattern: faster early-phase convergence with similar or slightly better final accuracy. On language modeling with token-frequency imbalance—Wikitext-2-style data with training-only token corruption and a 10% low-resource variant—AdamN lowers rare-token perplexity versus AdamW without warmup while matching head and mid-frequency performance. In full fine-tuning of Llama 3.1–8B on a small dataset, AdamN reaches AdamW’s final perplexity in roughly half the steps (≈ 2.25 xfaster time-to-quality). Finally, on a ViT-Base/16 transferred to CIFAR-
100 (batch size 256), AdamN achieves 88.8% test accuracy vs. 84.2% for AdamW and reaches
40–80% validation-accuracy milestones in the first epoch (AdamW reaches 80% by epoch 59),
reducing epochs, energy use, and cost. Full article

Background/Objectives: Safe disposal of unused medicines represents an increasing public health and environmental concern. Until 2024, Romanian community pharmacies collected expired medicines from the public, though implementation was inconsistent. Using a knowledge–attitude–practice (KAP) framework, this study assessed community pharmacists’ self-reported involvement in pharmaceutical waste prevention in Bihor County, Romania, one year after new legislation transferred collection responsibilities to hospital-based centers. Methods: A cross-sectional survey was conducted in May 2025 using a self-administered questionnaire comprising 22 items covering socio-demographics, professional practices, knowledge, and attitudes. Eligible participants were community pharmacists (N = 285). Results: Respondents reported high awareness and favourable attitudes toward pharmaceutical waste management: 98.2% indicated awareness of current legislation, 94.4% reported receiving training on the new regulations, 99.6% acknowledged health and environmental risks, and 98.9% expressed agreement that patient education is important. However, 55.4% reported providing disposal information only when patients requested it, while 89.8% indicated that patients rarely asked about medicine disposal. Self-reported proactive counseling increased with patient request frequency (χ²(3) = 7.914, p = 0.048), with pharmacists in the high-request group reporting substantially higher proactive counseling than those in the low-request group (83.3% vs. 42.9%). In an adjusted logistic regression, low request frequency was associated with lower odds of proactive counseling (aOR = 0.21, 95% CI: 0.05–0.98, p = 0.047). Most respondents (94.6%) perceived waste-related responsibilities, though these perceptions were only weakly related to reported counseling behaviors. Conclusions: Pharmacists reported high awareness and positive attitudes toward pharmaceutical waste management, but counseling remained reactive. Patient demand was a key correlate of counseling proactivity, underscoring the need for structured education within Romania’s hospital-based take-back system. Full article

In video scene graph prediction, the aim is to capture structured object interactions that occur over time in dynamic visual content. While recent spatiotemporal attention-based models have improved performance, they often suffer from high computational costs and limited structural consistency across long sequences. Therefore, we developed a Fourier transformer mixer network (FTM-Net), a modular, frequency-aware architecture that integrates spatial and temporal modeling via spectral operations. It incorporates a resolution-invariant Fourier Mixer for global spatial encoding and a Fast Fourier Transform (FFT)-Net-based temporal encoder that efficiently represents long-range dependencies with less complexity. To improve structural integrity, we introduce a spectral consistency loss function that synchronizes high-frequency relational patterns between frames. Experiments conducted utilizing the Action Genome dataset demonstrate that FTM-Net surpasses previous methodologies in terms of both Recall@K and mean Recall@K while markedly decreasing parameter count and inference duration, providing an efficient, interpretable, and generalizable approach for structured video comprehension. Full article

Background: The aim of this study is to investigate the potential link between the use of specific medications and oral adverse drug reactions. Methods: The 100 most frequently prescribed drugs in Germany in 2023 were compiled using the “PharMaAnalyst” database. According to the descriptions of adverse drug reactions (ADRs) in the patient information leaflets the ADRs were selected, analyzed and weighted with scores according to a classification system that distinguishes four groups of ADRs by frequency: ‘very common’ (4), ‘common’ (3), ‘uncommon’ (2) and ‘rare’ (1). The objective was to summarize the scores of the oral ADRs and define the ‘oral side effect score’ (OSES). Results: After accounting for duplication due to various brand names, 49 medications were reviewed. A total of 65% of the medications exhibited oral ADRs. The number of oral ADRs per medication ranged from one to seven. Xerostomia and dysgeusia were the most prevalent oral side effects, accounting for 37% of cases. Overall, 34% of side effects were classified as either ‘very common’ or ‘common’. The medication groups with the highest OSES were antidepressants, antibiotics and analgesics. Of the individual medications, azithromycin, gabapentin and pregabalin exhibited the highest OSES. Conclusions: This study provides a comprehensive overview of oral side effects associated with the 100 most frequently prescribed drugs. Patients with polypharmacy are particularly likely to experience oral side effects such as xerostomia and dysgeusia. Due to their high OSES combinations, antibiotics, analgesics or antidepressants may trigger multiple oral ADRs. It is essential that the medical community is continuously updated on pharmacological knowledge to raise awareness of oral ADRs. Full article

Neutral atomic hydrogen (H I) regulates galaxy growth and quenching, but direct 21 cm measurements remain observationally expensive and affected by selection biases. We develop Bayesian neural networks (BNNs)—a type of neural model that returns both a prediction and an associated uncertainty—to infer the H I mass, ${log}_{10}\left(M_{\mathrm{HI}}\right)$, from widely available optical properties (e.g., stellar mass, apparent magnitudes, and diagnostic colors) and simple structural parameters. For continuity with the photometric gas fraction (PGF) literature, we also report the gas-to-stellar-mass ratio, ${log}_{10}(G/S)$, where explicitly noted. Our dataset is a reproducible cross-match of SDSS DR12, the MPA–JHU value-added catalogs, and the 100% ALFALFA release, resulting in 31,501 galaxies after quality controls. To ensure fair evaluation, we adopt fixed train/validation/test partitions and an additional sky-holdout region to probe domain shift, i.e., how well the model extrapolates to sky regions that were not used for training. We also audit features to avoid information leakage and benchmark the BNNs against deterministic models, including a feed-forward neural network baseline and gradient-boosted trees (GBTs, a standard tree-based ensemble method in machine learning). Performance is assessed using mean absolute error (MAE), root-mean-square error (RMSE), and probabilistic diagnostics such as the negative log-likelihood (NLL, a loss that rewards models that assign high probability to the observed H I masses), reliability diagrams (plots comparing predicted probabilities to observed frequencies), and empirical 68%/95% coverage. The Bayesian models achieve point accuracy comparable to the deterministic baselines while additionally providing calibrated prediction intervals that adapt to stellar mass, surface density, and color. This enables galaxy-by-galaxy uncertainty estimation and prioritization for 21 cm follow-up that explicitly accounts for predicted uncertainties (“risk-aware” target selection). Overall, the results demonstrate that uncertainty-aware machine-learning methods offer a scalable and reproducible route to inferring galactic H I content from widely available optical data. Full article

Space–air–ground systems employing free-space optical (FSO) communication leverage high-altitude platform stations (HAPS) to deliver seamless and ubiquitous connectivity. Although FSO links offer high capacity, they are highly susceptible to cloud extinction, which severely degrades link availability. Hybrid FSO/radio-frequency (RF) transmission and cloud-aware HAPS trajectory optimization can enhance resilience. However, the conventional cloud-aware hybrid FSO/RF transmission system based on hard-switching (HS) between the FSO and RF links leads to frequent link transitions and unstable throughput. To address these challenges, we propose a joint optimization framework that integrates soft-switch between FSO and RF links with deep reinforcement learning (DRL) for HAPS trajectory optimization. Soft-switching based on rateless codes (RCs) enables simultaneous transmission over both links, where the receiver accumulates packets until successful decoding with a single feedback. The feedback frequency of RC is sparse, which avoids feedback storms but also poses challenges to HAPS trajectory optimization. The DRL agent proactively optimizes HAPS trajectories to avoid cloud cover and maintain link availability. To address the sparse feedback of RCs for DRL training, a reward-shaped proximal policy optimization (PPO)-based agent is developed to jointly optimize throughput and trajectory smoothness. Simulations using realistic ERA5 data show that RC-PPO achieves higher throughput and smoother trajectories compared to the HS-PPO baseline. Full article

Building extraction from remote sensing imagery plays a pivotal role in applications such as smart cities, urban planning, and disaster assessment. Although deep learning has significantly advanced this task, existing methods still struggle to strike an effective balance among global semantic understanding, local detail recovery, and multi-scale contextual awareness—particularly when confronted with challenges including extreme scale variations, complex spatial distributions, occlusions, and ambiguous boundaries. To address these issues, we propose TriadFlow-Net, an efficient end-to-end network architecture. First, we introduce the Multi-scale Attention Feature Enhancement Module (MAFEM), which employs parallel attention branches with varying neighborhood radii to adaptively capture multi-scale contextual information, thereby alleviating the problem of imbalanced receptive field coverage. Second, to enhance robustness under severe occlusion scenarios, we innovatively integrate a Non-Causal State Space Model (NC-SSD) with a Densely Connected Dynamic Fusion (DCDF) mechanism, enabling linear-complexity modeling of global long-range dependencies. Finally, we incorporate a Multi-scale High-Frequency Detail Extractor (MHFE) along with a channel–spatial attention mechanism to precisely refine boundary details while suppressing noise. Extensive experiments conducted on three publicly available building segmentation benchmarks demonstrate that the proposed TriadFlow-Net achieves state-of-the-art performance across multiple evaluation metrics, while maintaining computational efficiency—offering a novel and effective solution for high-resolution remote sensing building extraction. Full article

The sustainable management of marine resources requires accurate knowledge of fishing activity patterns and their interaction with coastal infrastructure. Intelligent Transportation Systems (ITS) are increasingly applied in the maritime domain, where data-driven approaches enhance safety, efficiency, and sustainability. In this context, Automatic Identification System (AIS) data provide valuable insights into vessel behavior and fisheries management. This study employs the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to identify fishing grounds, and a density map-based approach to recognize port locations. By integrating AIS data with machine learning techniques, the study detects and analyzes fishing vessel activities, providing deeper insights into behaviors such as fishing ground visit times, durations, and transitions between fishing grounds and ports. A case study in the Aalesund area of Norway demonstrates that DBSCAN effectively reveals fishing activity patterns relevant to regulatory oversight and spatial planning, while density mapping accurately identifies fishing ports. The findings highlight the potential of AIS-based analytics and clustering methods within maritime ITS frameworks to enhance situational awareness, support compliance with fisheries regulations, and contribute to sustainable marine resource management. Using 2023 AIS data from the Aalesund region, 6 recurrent fishing grounds and 15 port locations are identified, and size-stratified visit frequency and residence-time distributions are quantified together with monthly seasonality in ground usage. Full article

Physical-layer covert communication is increasingly challenged by powerful detectors that exploit the fine-grained statistical structure of received signals. In realistic Radio Frequency (RF) front ends, signal-dependent impairments such as power amplifier (PA) nonlinearity and In-phase and Quadrature (I/Q) imbalance induce transmitter-specific, non-Gaussian emission statistics under which conventional Gaussian embedding rules cause detectable distribution drift. We propose ResDiff, a two-stage learn-then-embed framework that first trains a symbol-conditional diffusion prior to capture a hardware-consistent emission manifold, then embeds covert information through bounded, variance-adaptive residuals spread over a K-symbol block with coherent block decoding at the legitimate receiver. Simulations under a severe impairment profile in an Additive White Gaussian Noise (AWGN) channel show that ResDiff improves stealthiness while maintaining reliable covert recovery and that increasing K reduces detectability by lowering the per symbol embedding pressure. Overall, the results indicate that hardware-aware generative priors, combined with rate-controlled block embedding, provide a practical path to covert-in-cover-traffic communication under modern detection capabilities. Full article

Graph anomaly detection (GAD) is pivotal for security-critical applications like cybersecurity and financial fraud detection. While reconstruction-based Graph Neural Networks (GNNs) are prevalent, their efficacy is often compromised by two phenomena: (1) anomaly overfitting, where expressive models capture anomalous patterns, and (2) homophily-induced attenuation, where message passing smooths localized anomaly cues. This paper proposes CLEAN-GAD, a contamination-aware framework that mitigates anomaly influence during training through multi-view robust learning. Specifically, we develop a contrastive augmentation module that utilizes local inconsistency scores to identify and suppress pseudo-anomalous nodes and edges, thereby yielding a purified augmented view. To capture diverse anomaly signals, a frequency-adaptive encoder with dual-pass channels is designed to integrate low- and high-frequency information. Furthermore, we introduce a distribution-separation regularizer and cross-view alignment to stabilize learning and resolve view shifts. Theoretical analysis confirms that reducing the contamination ratio $\rho$ expands the reconstruction-risk gap between normal and anomalous nodes, inherently boosting detection performance. Extensive experiments on multiple benchmark datasets from various domains demonstrate the superior anomaly detection performance of CLEAN-GAD. Full article

Background: Pharmacovigilance is a critical component of patient safety, particularly among older adults with chronic diseases who are frequently exposed to polypharmacy. In Kosovo, adverse drug reactions (ADRs) reported by patients remain insufficiently recognized within the healthcare system. Polypharmacy, limited access to pharmaceutical counseling, and self-medication practices may contribute to increased medication-related harm. Capturing ADRs directly from patients provides valuable insight into medication safety challenges and communication gaps in clinical care. Objective: To assess the frequency, characteristics, and reporting behavior of adverse drug reactions among adults aged 60–75 years with chronic diseases in Kosovo, and to identify factors associated with awareness and reporting practices. Methods: A multicenter cross-sectional study was conducted between January and September 2025 in four major cities in Kosovo (Prishtina, Prizren, Peja, and Gjilan). A total of 1024 patients receiving continuous therapy for at least one chronic condition were surveyed using a structured questionnaire covering demographic characteristics, drug exposure, ADR experience, and reporting behavior. Statistical analyses included descriptive statistics, chi-square testing, and multivariable logistic regression to identify predictors of ADR reporting. Results: Overall, 47.3% of participants reported experiencing at least one ADR in the preceding 12 months. Among those, 39.5% reported the event to a healthcare professional, whereas 60.5% did not seek professional advice. The most frequently implicated drug classes were antihypertensives (32.8%), analgesics and non-steroidal anti-inflammatory drugs (27.4%), and antirheumatic agents (14.6%), with mainly gastrointestinal (24.1%) and cardiovascular (18.9%) manifestations. Approximately 19.8% of participants reported discontinuing medication due to adverse effects. Female patients were more likely to report ADRs compared to males (p < 0.01). Lack of prior counseling about potential side effects was independently associated with lower reporting (OR = 2.17; 95% CI: 1.41–3.33). Patients using more than six medications had a higher prevalence of ADRs (61.2%). Conclusion: Adverse drug reactions were frequently reported by older patients, while formal reporting to healthcare professionals remained limited. Strengthening patient education, improving patient–provider communication, and integrating clinical pharmacists into primary care may enhance pharmacovigilance practices and medication safety. Full article

In clinical CT imaging, high-density metallic implants often induce severe metal artifacts that obscure critical anatomical structures and degrade image quality, thereby hindering accurate diagnosis. Although deep learning has advanced CT metal artifact reduction (CT-MAR), many methods do not effectively use frequency information, which can limit the recovery of both fine details and overall image structure. To address this limitation, we propose a Hybrid-Frequency-Aware Mixture-of-Experts (HFMoE) network for CT-MAR. The proposed method synergizes the spatial-frequency localization of the wavelet transform with the global spectral representation of the Fourier transform to achieve precise multi-scale modeling of artifact characteristics. Specifically, we design a hybrid-frequency interaction encoder with three specialized branches, incorporating wavelet-domain, Fourier-domain, and cascaded wavelet–Fourier modulation, to distinctively refine local details, global structures, and complex cross-domain features. Then, they are fused via channel attention to yield a comprehensive representation. Furthermore, a Frequency-Aware Mixture-of-Experts (MoE) mechanism is introduced to dynamically route features to specific frequency experts based on the degradation severity, thereby adaptively assigning appropriate receptive fields to handle varying metal artifacts. Evaluations on synthetic (DeepLesion) and clinical (SpineWeb, CLINIC-metal) datasets show that HFMoE outperforms existing methods in both quantitative metrics and visual quality. Our method demonstrates the value of explicit frequency-domain adaptation for CT-MAR and could inform the design of other image restoration tasks. Full article

Unmanned aerial vehicles (UAVs) and birds are typical low-altitude small targets in optical remote sensing, often occupying only a few pixels and exhibiting highly similar appearances, which limits the effectiveness of appearance-based discrimination at long distances and low resolutions. To overcome this, we propose a non-appearance-based classification framework using kinematic and time–frequency features. At the trajectory level, kinematic features—including the coefficient of variation of velocity and acceleration, the Spatiotemporal Box-counting Fractal Dimension (SBFD), and the Local Higuchi Fractal Dimension (LHFD)—quantify multi-scale trajectory complexity. At the scale-variation level, time–frequency features, specifically the Time-Frequency Aware Singular Value Entropy (TF-SVE) derived from bounding-box area sequences, capture non-stationary oscillations from bird wing flapping, reflecting behavioral differences from rigid UAV motion. Experiments on a complex real-world dataset show that stacking these features achieves 99.47% classification accuracy, demonstrating a robust, resolution-invariant, and practically effective approach for non-appearance-based recognition of low-altitude targets. Full article

Precise medical segmentation of oral ulcers is mandatory and crucial for early diagnosis, but it remains a very challenging task due to rich backgrounds, overexposed or underexposed lesions, and the complex surrounding areas. Therefore, in order to address this challenge, this paper introduces HF-EdgeFormer, a novel hybrid model for oral ulcer segmentation on the AutoOral dataset. This U-shaped transformer-like architecture is, based on publicly available models, the second documented solution for oral ulcer segmentation and it explicitly integrates high-order frequency interactions by using multi-dimensional edge cues. At the encoding stage, a HFConv (High-order Focus Convolution) module divides the feature channels into local streams and global streams, performing learnable filtering via FFT and depth-wise convolutions. After that, it fuses them through stacks of focal transformers and attention gates. In addition to the HFConv block, there are two edge-aware units: the EdgeAware Localization module (that uses eight-direction Sobel filters) and a new Precision EdgeEnhance module (channel-wise Sobel fusion), both used in order to reinforce the boundaries. Skip connections imply Multi-dilated Attention Gates, accompanied by a Spacial-Channel Attention Bridge to accentuate lesion-consistent activations. Moreover, the novel architecture employs an innovative lightweight vision transformer-based bottleneck. It consists of four SegFormerBlock modules localized at the network’s deepest point, so we can achieve global relational modeling exactly where the largest receptive field is present. The model is trained on the AutoOral dataset (introduced by the same team that developed the HF-Unet arhitecture), but due to the limited available images, it needed to be extended by using extensive geometric and photometric augmentations (like RandomAffine, flips, and rotations). This novel architecture achieves a test Dice score of almost 82% and a little over 85% sensitivity while maintaining high precision and specificity, highly valuable in medical segmentation. These results surpass prior HF-UNet baselines while maintaining the model light, with minimal inference memory gains. Full article

In low-resource automatic speech recognition (ASR) scenarios, parameter-efficient fine-tuning (PEFT) has become a crucial approach for adapting large pre-trained speech models. Although low-rank adaptation (LoRA) offers clear advantages in efficiency, stability, and deployment friendliness, its performance remains constrained because random initialization fails to capture the time–frequency structural characteristics of speech signals. To address this limitation, this work proposes a structured initialization mechanism that integrates LoRA with the discrete wavelet transform (DWT). By combining wavelet-based initialization, a multi-scale fusion mechanism, and a residual strategy, the proposed method constructs a low-rank adaptation subspace that better aligns with the local time–frequency properties of speech signals. Discrete Wavelet Transform-Based LoRA Initialization (DWTLoRA) enables LoRA modules to incorporate prior modeling of speech dynamics at the start of fine-tuning, substantially reducing the search space of ineffective directions during early training and improving convergence speed, training stability, and recognition accuracy under low-resource conditions. Experimental results on Sichuan dialect speech recognition based on the Whisper architecture demonstrate that the proposed DWTLoRA initialization outperforms standard LoRA and several PEFT baseline methods in terms of character error rate (CER) and training efficiency, confirming the critical role of signal-structure-aware initialization in low-resource ASR. Full article