Browse Articles
Discover research articles across all indexed journals
Design of displacement-based viscous damper damping structures
Development and validation of interpretable machine learning models to predict distant metastasis and prognosis of muscle-invasive bladder cancer patients
Early transcriptomic changes in cucumber and maize roots in response to FePO4 nanoparticles as a source of P and Fe
Real-time fear emotion recognition in mice based on multimodal data fusion
An explainable hybrid feature aggregation network with residual inception positional encoding attention and EfficientNet for cassava leaf disease classification
Abstract Cassava is a tuberous edible plant native to the American tropics and is essential for its versatile applications including cassava flour, bread, tapioca, and laundry starch. Cassava leaf diseases reduce crop yields, elevate production costs, and disrupt market stability. This places significant burdens on farmers and economies while highlighting the need for effective management strategies. Traditional methods of manual disease diagnosis are costly, labor-intensive, and time-consuming. This research aims to address the challenge of accurate disease classification by overcoming the limitations of existing methods, which encounter difficulties with the complexity and variability of leaf disease symptoms. To the best of our knowledge, this is the first study to propose a novel dual-track feature aggregation architecture that integrates the Residual Inception Positional Encoding Attention (RIPEA) Network with EfficientNet for the classification of cassava leaf diseases. The proposed model employs a dual-track feature aggregation architecture which integrates the RIPEA Network with EfficientNet. The RIPEA track extracts significant features by leveraging residual connections for preserving gradients and uses multi-scale feature fusion for combining fine-grained details with broader patterns. It also incorporates Coordinate and Mixed Attention mechanisms which focus on cross-channel and long-range dependencies. The extracted features from both tracks are aggregated for classification. Furthermore, it incorporates an image augmentation method and a cosine decay learning rate schedule to improve model training. This improves the ability of the model to accurately differentiate between Cassava Bacterial Blight (CBB), Brown Streak Disease (CBSD), Green Mottle (CGM), Mosaic Disease (CMD), and healthy leaves, addressing both local textures and global structures. Additionally, to enhance the interpretability of the model, we apply Grad-CAM to provide visual explanations for the model’s decision-making process, helping to understand which regions of the leaf images contribute to the classification results. The proposed network achieved a classification accuracy of 93.06%.
Developing a standardized framework for evaluating health apps using natural language processing
Multi-scale adversarial diffusion network for image super-resolution
Predator crow search optimization with explainable AI for cardiac vascular disease classification
Abstract The proposed framework optimizes Explainable AI parameters, combining Predator crow search optimization to refine the predictive model’s performance. To prevent overfitting and enhance feature selection, an information acquisition-based technique is introduced, improving the model’s robustness and reliability. An enhanced U-Net model employing context-based partitioning is proposed for precise and automatic left ventricular segmentation, facilitating quantitative assessment. The methodology was validated using two datasets: the publicly available ACDC challenge dataset and the imATFIB dataset from internal clinical research, demonstrating significant improvements. The comparative analysis confirms the superiority of the proposed framework over existing cardiovascular disease prediction methods, achieving remarkable results of 99.72% accuracy, 96.47% precision, 98.6% recall, and 94.6% F1 measure. Additionally, qualitative analysis was performed to evaluate the interpretability and clinical relevance of the model’s predictions, ensuring that the outputs align with expert medical insights. This comprehensive approach not only advances the accuracy of CVD predictions but also provides a robust tool for medical professionals, potentially improving patient outcomes through early and precise diagnosis.
Control of cucumber downy mildew disease under greenhouse conditions using biocide and organic compounds via induction of the antioxidant defense machinery
Study on the two-phase coupling migration mechanism of deceleration aggregate and water in coal mine water inrush channel
Prediction of methane hydrate equilibrium in saline water solutions based on support vector machine and decision tree techniques
Optimization of low-temperature nitrogen plasma in reducing fungi and aflatoxin human exposure through maize
Abstract Globally, aflatoxin contamination in maize remains a huge burden despite many interventions put in place. The use of low-temperature plasma to decontaminate the maize is a potential solution for ensuring the safety and extended shelf life of the grain. This study optimized the parameters and investigated the efficacy of low-temperature nitrogen plasma (LTNP) in destroying fungi and reducing exposure to aflatoxins in naturally contaminated maize from an endemic region. The study generated 17 experimental runs using the Response Surface Methodology (RSM) of the Box Behnken Design (BBD) with exposure time, pressure, and ionization density as independent variables. Quantitative exposure assessment was conducted using Monte Carlo simulations followed by sensitivity and scenario analysis to study factors influencing exposure and best aflatoxin-reducing plasma parameters. The best-fitting RSM model, the linear model, indicated that increased exposure time but not pressure and power led to a corresponding statistically significant decrease in the fungal load and aflatoxin content. LTNP reduced aflatoxin contamination to levels below all the main global regulatory limits. Numerical optimization of the percent reduction in aflatoxin and fungal load indicated that an exposure time of 1793.4 s, pressure of 0.98 pascal and ionization power of 189.8 W are required to achieve an optimal reduction of aflatoxin content of 82.6% and fungal load of 96.9%. Exposure assessment indicated high exposure especially for populations with lower body weight with ρ = -0.46 between body weight and exposure. The best LTNP combinations achieved aflatoxin exposure reduction results comparable to but with markedly less variation than existing practically used decontamination methods. Further optimization studies during upscaling are recommended, incorporating independent factors such as temperature and processing volume and outcomes such as organoleptic, physical, and chemical changes in the food matrices after treatment.
Implications of intrinsic capacity subtypes for post-PCI patients older than 45
Uncertainty quantification with graph neural networks for efficient molecular design
Cytogenetic signatures favoring metastatic organotropism in colorectal cancer
Abstract Colorectal carcinoma (CRC) exhibits metastatic organotropism, primarily targeting liver, lung, and rarely the brain. Here, we study chromosomal imbalances (CIs) in cohorts of primary CRCs and metastases. Brain metastases show the highest burden of CIs, including aneuploidies and focal CIs, with enrichment of +12p encoding KRAS. Compared to liver and lung metastases, brain metastases present with increased co-occurrence of KRAS mutation and amplification. CRCs with concurrent KRAS mutation and amplification display significant metabolic reprogramming with upregulation of glycolysis, alongside upregulation of cell cycle pathways, including copy number gains of MDM2 and CDK4. Evolutionary modeling suggests early acquisition of many organotropic CIs enriched in both liver and brain metastases, while brain-enriched CIs preferentially emerge later. Collectively, this study supports a model where cytogenetic events in CRCs favor site-specific metastatic colonization. These site-enriched CI patterns may serve as biomarkers for metastatic potential in precision oncology.
Redox disruption using electroactive liposome coated gold nanoparticles for cancer therapy
Universal law of hierarchical dynamics in gels arising from confluence of local physically dynamic bonds
Bright upconversion over extended temperatures enabled by an organic surface layer
Cooperative role of PACT and ADAR1 in preventing aberrant PKR activation by self-derived double-stranded RNA
Abstract Double-stranded RNAs (dsRNAs) produced during viral infections are recognized by the innate immune sensor protein kinase R (PKR), triggering a host translation shutoff that inhibits viral replication and propagation. Given the harmful effects of uncontrolled PKR activation, cells must tightly regulate PKR to ensure that its activation occurs only in response to viral infections, not endogenous dsRNAs. Here, we use CRISPR-Translate, a FACS-based genome-wide CRISPR-Cas9 knockout screening method that exploits translation levels as a readout and identifies PACT as a key inhibitor of PKR during viral infection. We find that PACT-deficient cells hyperactivate PKR in response to different RNA viruses, raising the question of why cells need to limit PKR activity. Our results demonstrate that PACT cooperates with ADAR1 to suppress PKR activation from self-dsRNAs in uninfected cells. The simultaneous deletion of PACT and ADAR1 results in synthetic lethality, which can be fully rescued in PKR-deficient cells. We propose that both PACT and ADAR1 act as essential barriers against PKR, creating a threshold of tolerable levels to endogenous dsRNA in cells without activating PKR-mediated translation shutdown and cell death.