Search Results (771)

Bacteriophages are a key ecological factor in the legume rhizosphere, controlling bacterial populations and affecting introduced inoculant strains. Despite their importance, rhizobiophage genomic diversity remains poorly characterized. We report the complete genome of a novel predicted temperate Sinorhizobium phage, AP-202, isolated from agricultural Chernozem. This siphovirus infects the symbiont Sinorhizobium meliloti. Its 121,599 bp dsDNA genome has a strikingly low GC content (27.1%), likely reflecting adaptive evolution and a strategy to evade host defenses. The linear genome is flanked by 240 bp direct terminal repeats (DTRs), and its DNA packaging follows a T7-like strategy. Annotation predicted 178 protein-coding genes and one tRNA. Functional analysis revealed a complete lysogeny module and a divergent, two-pronged codon-usage strategy for translational control. A significant part of the proteome (74.2%) comprises hypothetical proteins, with 50 CDSs having no database homologs, underscoring its genetic novelty. Complete-genome comparison shows minimal similarity to known rhizobiophages, defining AP-202 as a distinct lineage. Phenotypic analysis indicates AP-202 acts as a selective ecological filter, with host resistance being more prevalent in agricultural than in natural soils. The AP-202 genome provides a unique model for studying phage–host coevolution in the rhizosphere and is a valuable resource for comparative genomics and soil virome research. Full article

Accurate measurement of luminous flux from solid-state light sources typically requires spectroradiometric equipment or integrating spheres. This work investigates a compact alternative based on a fiber-coupled RGB photodiode system and develops the optical, spectral, and geometric foundations required to obtain traceable flux estimates from reduced-channel measurements. The system under study comprises an LED with known spectral power distribution (SPD), optical head, optical fiber, a protective sensor window, and a photodiode matrix type sensor. A complete end-to-end analysis of the optical path is presented, including geometric coupling efficiency, fiber transmission and angular redistribution, Fresnel losses in the sensor window, and the mosaic structure of the sensor. Additional effects such as fiber–sensor alignment, fiber-facet tilt, air gaps, and LED placement tolerances are quantified and incorporated into a formal uncertainty budget. Using the manufacturer-supplied SPD of the reference LED together with the measured R, G, and B channel responsivity functions of the sensor, a calibration-based mapping is established to reconstruct photopic luminous flux from the three-channel outputs. These results demonstrate that, with appropriate modeling and calibration of all optical stages, a fiber-coupled RGB photodiode mosaic can provide practical and scientifically meaningful luminous-flux estimation for white LEDs, offering a portable and cost-effective alternative to conventional photometric instrumentation in mid-accuracy applications. Further optimization of computation speed can enable fully integrated measurement systems in resource-constrained environments. Full article

Fetal growth restriction (FGR) remains a significant problem in obstetrics and is a key risk factor for perinatal brain injury. The fetal neuronal vesicles (FNVs) isolated from maternal blood represent an innovative approach—a “fetal brain liquid biopsy”—enabling early diagnostics of neuronal dysfunction in FGR. Western blotting was used to evaluate the protein pattern expression of FNVs isolated from the blood of pregnant women with FGR and uncomplicated pregnancy. Significant changes in the neurotrophic proteins levels (pro-BDNF, pro-NGF) and presynaptic neurotransmission proteins (SYN1, SYP, SYNPO) were identified. New data were obtained on changes in the expression of proteins of sumoylation (SUMO2/3/4) and neddylation (NAE1, UBC12), which differs in early-onset and late-onset FGR. Moreover, increased SUMO2/3/4 levels can be considered as an endogenous neuroprotective response to cerebral hemodynamic reaction in fetuses with late-onset growth restriction. An association has been established between changes in the expression of the studied proteins and intraventricular hemorrhage (IVH) in newborns with late-onset growth restriction. Full article

Homomorphic encryption provides a powerful cryptographic solution for privacy-preserving deep neural network inference, enabling computation on encrypted data. However, the practical application of homomorphic encryption is fundamentally constrained by the noise budget, a core component of homomorphic encryption schemes. The substantial multiplicative depth of modern deep neural networks rapidly consumes this budget, necessitating frequent, computationally expensive bootstrapping operations to refresh the noise. This bootstrapping process has emerged as the primary performance bottleneck. Current noise management strategies are predominantly static, triggering bootstrapping at pre-defined, fixed intervals. This approach is sub-optimal for deep, complex architectures, leading to excessive computational overhead and potential accuracy degradation due to cumulative precision loss. To address this challenge, we propose a Deep Network-aware Adaptive Noise-budget Management mechanism, a novel mechanism that formulates noise budget allocation as a sequential decision problem optimized via reinforcement learning. The core of the proposed mechanism comprises two components. First, we construct a layer-aware noise consumption prediction model to accurately estimate the heterogeneous computational costs and noise accumulation across different network layers. Second, we design a Deep Q-Network-driven optimization algorithm. This Deep Q-Network agent is trained to derive a globally optimal policy, dynamically determining the optimal timing and network location for executing bootstrapping operations, based on the real-time output of the noise predictor and the current network state. This approach shifts from a static, pre-defined strategy to an adaptive, globally optimized one. Experimental validation on several typical deep neural network architectures demonstrates that the proposed mechanism significantly outperforms state-of-the-art fixed strategies, markedly reducing redundant bootstrapping overhead while maintaining model performance. Full article

This study investigates the determinants of medium-term revenue forecast errors across eight developed countries: the United States, the United Kingdom, Germany, Ireland, Hong Kong, New Zealand, Australia, and Canada. By examining two- and three-year revenue forecasts, this study applies the Kenneth Holden–David Peel test to identify forecast biases and employs panel regression models to assess the economic factors influencing forecast accuracy. The findings indicate that inflation, GDP growth, and budget balance rules are positively associated with forecast errors, whereas unemployment, population growth, and the number of fiscal rules mitigate these errors. Panel estimates reveal that fiscal structure-related variables are not only statistically significant but also economically meaningful determinants of medium-term revenue forecast errors. The results underscore the persistent challenges in achieving accurate revenue forecasts and highlight the necessity for improved forecasting methodologies to enhance fiscal policy effectiveness and resource allocation. Strengthening forecasting frameworks can contribute to more reliable revenue projections, reducing fiscal uncertainty and supporting sound economic decision making. Full article

Dairy herd decisions about replacement, herd size, reproduction, and capital investments have long-lasting consequences for herd structure and farm cash flow. Yet most planning tools emphasize static budgets rather than the dynamic evolution of animal numbers and cash availability. The Dairy Herd Structure Simulation and Cash Flow tool is a web-based decision-support system, available through the Dairy Management Decision Support Tools website, designed to simulate these dynamics under alternative management strategies. The model operates in monthly time steps using a Markov–chain framework in which transition probabilities among animal states are driven by user-specified parameters such as culling, reproduction, and heifer management. Calves, heifers, and cows are tracked by age and lactation group, and starting conditions can be entered as herd-level summaries or via individual-animal spreadsheets. Economic components include milk income, variable costs, cull-cow income, heifer purchases or sales, miscellaneous costs, and loan amortization. For each scenario, the tool projects monthly cash flow and income over variable cost per cow, together with graphical summaries of herd structure. An example application compares a baseline steady-state herd with a heifer-driven herd growth scenario, illustrating how replacement strategies influence herd composition and net cash flow, supporting more informed dairy herd planning and risk management. Full article

Data-driven quality control (QC) systems for the hot forming of steel parts increasingly rely on deep learning models deployed at the network edge, making multivariate sensor time series a critical asset for both local decisions and management information system (MIS) reporting. However, these models are vulnerable to adversarial perturbations and realistic signal disturbances, which can induce misclassification and distort key performance indicators (KPIs) such as first-pass yield (FPY), scrap-related losses, and latency service-level objectives (SLOs). To address this risk, this study introduces a Digital-Twin-Conditioned Diffusion Purification (DTCDP) framework that constrains latent diffusion-based denoising using process states from a lightweight digital twin of the hot-forming line. At each reverse-denoising step, the twin provides physics residuals that are converted into a scalar penalty, and the diffusion latent is updated with a guidance term. This directly bends the sampling trajectory toward reconstructions that adhere to process constraints while removing adversarial perturbations. DTCDP operates as an edge-side preprocessing module that purifies sensor sequences before they are consumed by existing long short-term memory (LSTM)-based QC models, while exposing purification metadata and physics-guidance diagnostics to the plant MIS. In a four-week production dataset comprising more than 40,000 bars, with white-box ℓ∞ attacks crafted on multivariate sensor time series using Fast Gradient Sign Method and Projected Gradient Descent at perturbation budgets of 1–3% of the physical range, combined with additional realistic disturbances, DTCDP improves the robust classification performance of an LSTM-based QC model from 61.0% to 81.5% robust accuracy, while keeping clean accuracy (≈93%) and FPY on clean data (≈97%) essentially unchanged. These results indicate that physics-aware, digital-twin-guided diffusion purification can enhance the adversarial robustness of edge QC in hot forming without compromising operational KPIs. Full article

Fiscal capacity is a core dimension of state capacity. Effective oversight of public expenditure is therefore essential for fiscal sustainability, a foundational element of sustainable development. As local government debt has steadily increased in China, participatory budgeting has emerged as an innovative mechanism for citizens to exercise such oversight and influence fiscal decisions. Our paper examines the effect of participatory budgeting on local government debt in China. Using a panel dataset covering 242 Chinese cities from 2013 to 2022, we examine the effect of participatory budgeting adoption on the scale of explicit government debt. Our results show that adopting participatory budgeting moderately reduces local government debt levels. Further mechanism analysis indicates that participatory budgeting operates through two channels. First, by enhancing budgetary transparency, it strengthens public scrutiny, which in turn disciplines government borrowing. Second, it redirects public spending toward welfare sectors like education and health, thereby crowding out large, debt-financed investment projects. Our findings contribute to the literature on participatory budgeting, fiscal democracy, and bottom-up accountability in public finance. The results suggest that participatory budgeting can be an effective policy tool for improving fiscal discipline and curbing government debt risks, ultimately fostering more sustainable and equitable local governance. Full article

Sporadic-E (Es) layers can strongly perturb HF/VHF propagation and create intermittent interference, motivating higher-revisit monitoring at the frequencies most affected. EsTRACE (Es-layer TRAnsient Cloud Explorer) is a PlanarSat mission concept that transmits sequential beacons in the 28/50 MHz amateur bands using FT4 (weak-signal digital) and CW (continuous wave) waveforms and leverages distributed amateur receiver networks for near-real-time SNR mapping. This paper documents the early-phase spacecraft design from the Bid/proposal phase (Bid), through the Conceptual Design Review (CoDR), to the Preliminary Design Review (PDR), using a power-first sizing loop that couples link-budget closure to duty cycle and solar-array area under a free-tumbling, batteryless constraint. The analysis supports conceptual feasibility of the architecture under stated antenna and ground-segment assumptions; on-orbit demonstration and measured RF/antenna characterization are identified as required future validation steps. Full article

We present LazyNet, a compact one-step neural-ODE model for single-cell CRISPR activation/interference (A/I) that operates directly on two-snapshot (“pre → post”) measurements and yields parameters with clear mechanistic meaning. The core log–linear–exp residual block exactly represents multiplicative effects, so synergistic multi-locus responses appear as explicit components rather than opaque composites. On a 53k-cell × 18k-gene neuronal Perturb-seq matrix, a three-replica LazyNet ensemble trained under a matched 1 h budget achieved strong threshold-free ranking and competitive error (genome-wide r ≈ 0.67) while running on CPUs. For comparison, we instantiated transformer (scGPT-style) and state-space (RetNet/CellFM-style) architectures from random initialization and trained them from scratch on the same dataset and within the same 1 h cap on a GPU platform, without any large-scale pretraining or external data. Under these strictly controlled, low-data conditions, LazyNet matched or exceeded their predictive performance while using far fewer parameters and resources. A T-cell screen included only for generalization showed the same ranking advantage under the identical evaluation pipeline. Beyond prediction, LazyNet exposes directed, local elasticities; averaging Jacobians across replicas produces a consensus interaction matrix from which compact subgraphs are extracted and evaluated at the module level. The resulting networks show coherent enrichment against authoritative resources (large-scale co-expression and curated functional associations) and concordance with orthogonal GPX4-knockout proteomes, recovering known ferroptosis regulators and nominating testable links in a lysosomal–mitochondrial–immune module. These results position LazyNet as a practical option for from-scratch, low-data CRISPR A/I studies where large-scale pretraining of foundation models is not feasible. Full article

We formalize an inverse, data-conditioned variant of the Variational Quantum Eigensolver (VQE) for clinical biomarker discovery. Given patient-encoded quantum states, we construct a task-specific Hamiltonian whose coefficients are inferred from clinical associations and interpret its expectation value as a calibrated energy score for prognosis and treatment monitoring. The method integrates coefficient estimation, ansatz specification with basis rotations, commuting-group measurements, and a practical shot budget analysis. Evaluated on public infectious disease datasets under severe class imbalance, the approach yields consistent gains in balanced accuracy and precision–recall over strong classical baselines, with stability across random seeds and feature ablations. This variational energy scoring framework bridges Hamiltonian learning and clinical risk modeling, offering a compact, interpretable, and reproducible route to biomarker prioritization and decision support. Full article

While most prison healthcare litigation seeks individual relief, some cases lead to broader structural reform via consent decrees—court-approved “legally binding performance improvement plans” designed to improve conditions. This study systematically analyzes 121 such settlements from 1970 to 2022 to assess their policy goals and implementation strategies. We identify the substantive areas targeted—general medical care, mental health, dental services, and treatment for specialized conditions like HIV, Hepatitis C, and COVID-19—and trace trends across time and geography. These agreements span 39 states and the federal system, with most states subject to multiple cases. They frequently mandate changes to budgets, staffing, facility infrastructure, training, and patient rights, alongside monitoring for quality improvement. Our findings suggest that consent decrees function not only as judicial remedies but as tools of policy development and institutional reform, shedding light on the role of courts in shaping correctional healthcare delivery. These findings also show how institutional responses to healthcare failures in prisons shape the conditions under which serious harm—and in some cases, preventable death—occur behind bars. Full article

The quality of human capital is pivotal to the promotion of economic growth and development. In this regard, the quality of education and of academic, practice-based research systems plays a crucial role. The authors conducted an analysis of the systemic framework for financing higher education and science in Poland, with a particular focus on the economic and financial situation of the 21 largest Polish universities from 2019 to 2024. In order to assess whether the current financial basis of these entities facilitates conducting research, bridges the gap between academia and industry, and thus supports their sustainable development, a taxonomic composite indicator and cluster analysis were employed. The results obtained indicated unfavourable trends in the domain of higher education and research and development (R&D) activity among Polish universities. These include, in particular, the exceeding of operating costs over operating revenues, the deterioration of financial results, insufficient funding from the state budget in relation to the scale of operations, and relatively high volatility of the economic and financial situation in subsequent years. Full article

This paper evaluates neural-network approaches for lithium-ion battery state-of-charge (SoC) estimation under a unified pipeline, fixed data partitions, and identical preprocessing. We study FNNs trained with Levenberg–Marquardt (LM), Bayesian Regularization (BR), and Scaled Conjugate Gradient (SCG) across three hidden sizes (10, 20, 30) and three topologies: Fitting, Nonlinear Input–Output (Nonlinear I/O), and time-series NAR/NARX. Models are assessed using test MSE and RMSE, correlation (R), generalization gap, convergence indicators (final gradient, damping factor), wall time per epoch, and a relative compute-cost index. On the Fitting task, BR-Fitting-FNN with 20 neurons provides the best accuracy-efficiency balance, while LM-Fitting-FNN with 30 neurons reaches slightly lower error at a higher cost. For Nonlinear I/O, BR-Nonlinear I/O-FNN with 30 neurons achieves the lowest test MSE with clear evidence of effective weight shrinkage; LM-Nonlinear I/O-FNN with 20 neurons is a close alternative. In time-series settings, LM-NAR-FNN with 10 neurons attains the lowest test error and fastest epochs but shows a very negative gap that suggests test-split favorability; BR-NAR-FNN with 30 neurons is more costly yet consistently strong. For NARX, LM-NARX-FNN with 20 neurons yields the best test accuracy and robust convergence. Overall, BR delivers the most reliable accuracy–robustness trade-off as networks widen, LM often achieves the best raw accuracy with careful split validation, and SCG offers the lowest training cost when resources are limited. These results provide practical guidance for selecting SoC estimators to match accuracy targets, computing budgets, and deployment constraints in battery management systems. Full article

Although Convolutional Neural Networks (CNNs) have delivered state-of-the-art accuracy in plant disease classification, their deployment is still hindered by data scarcity, computational cost, and architectural heterogeneity. Transfer learning from large-scale pre-trained datasets alleviates these issues, yet generic feature extraction suffers from domain shift, while indiscriminate fine-tuning risks over-fitting and elevated training budgets. To address the identified limitations, PlantClassiNet is implemented as a unified framework. This framework facilitates systematic comparative analysis of six CNN architectures, AlexNet, ResNet50, InceptionV3, MobileNetV3Small, DenseNet121 and EfficientNetB0, across three publicly available datasets: PlantVillage, PlantLeaves and Eggplant. Two alternative fine-tuning approaches are proposed: Adaptive Fine-tuning (AdapFitu), which adaptively determines the depth of unfrozen layers, learning rates, and reinitializes selected layers, and a fixed-parameter baseline, which trains only the newly added classifier while keeping the convolutional backbone frozen and unfreezes a fixed number of network layers for retraining. Extensive experiments demonstrate that large models AlexNet, ResNet50, and Inceptionv3 achieve test accuracy exceeding 98.74% on the sizable PlantVillage dataset, whereas lightweight counterparts MobileNetV3Small, DenseNet121, and EfficientNetB0 achieve high accuracy of 99.79% ± 0.21% on the smaller Eggplant collection after fine-tuning. Full article