Interstate Commission for Water Coordination of Central Asia

Plant diseases represent a persistent challenge to global food security, with annual crop yield losses reaching as high as 40%. Timely and reliable identification of these diseases is therefore essential for reducing economic losses and ensuring sustainable farming practices. Although deep learning has achieved remarkable success in image-based disease classification, much of the existing research emphasizes complex convolutional neural networks that are unsuitable for resource-constrained environments. In this work, we present a systematic evaluation of two widely adopted transfer learning models ResNet50 and MobileNetV2 applied to the New Plant Disease dataset (~87,000 images spanning 38 categories). Our experiments examine the role of augmentation, learning rates, batch sizes, and fine-tuning strategies in model performance. Results demonstrate that MobileNetV2 not only achieved superior accuracy (98.33%) with strong precision, recall, and AUC values, but also required seven times fewer parameters and significantly reduced training time compared with ResNet50. Error analysis further revealed MobileNetV2’s ability to differentiate between diseases with overlapping symptoms. Importantly, its lightweight architecture supports real-time implementation on mobile devices, drones, and IoT systems, offering clear advantages for field deployment. Unlike prior studies that emphasize raw accuracy, this research highlights efficiency, robustness, and deployability. Overall, our findings establish MobileNetV2 as a practical and scalable solution for next-generation precision agriculture and low-cost disease monitoring systems.

We investigate the gravitational lensing effects on neutrino oscillations within the framework of Modified Gravity (MOG). By modeling neutrinos as both plane waves and Gaussian wave packets, we analyze how the presence of a massive lensing object between the neutrino source and detector modifies the flavor transition probabilities. In the plane-wave approximation, we derive analytical expressions for lensing-induced interference patterns, showing explicit dependence on the MOG parameter α , the impact parameters, and the neutrino mass hierarchy. Extending the analysis to a wave-packet framework, we account for decoherence effects and quantify the damping of oscillation amplitudes due to finite energy and spatial spreads of the wave packets. We compute the decoherence length and damping factor for a realistic Sun–Earth system and demonstrate how both the absolute neutrino mass scale and the MOG parameter influence coherence loss. Our results indicate that gravitational lensing in alternative gravity theories can leave observable imprints on neutrino oscillation patterns, offering a novel probe of both neutrino properties and modifications of General Relativity.

This study presents an innovative approach for unveiling the hidden relationships between natural frequency patterns and structural parameters in grid-form frames. By analyzing vibrational characteristics, we determine key features, namely the number of vertical beams, boundary conditions, and aspect ratios. Extensive finite element analysis generates a dataset, mapping the natural frequencies as features against structural parameters as labels reveals distinct, streamlined clusters in the feature hyperspace, highlighting an underlying order in the system’s dynamics. An advanced classification and interpolation model navigates these spectral trajectories to predict structural parameters accurately, even in the presence of damage or different materials. This study offers new insights into the intrinsic dynamics of complex structures, inviting further exploration into the subtle interplay between vibrational characteristics and structural identity. These findings open new avenues for research, potentially transforming the understanding of structural behavior in practical engineering applications.

ABSTRACT In Central Asia, by the end of the Bronze Age, a fundamental change in funeral tradition occurred, from traditional kurgans or burial mounds to Zoroastrian belief in dakhmas or Towers of Silence, for ritualized exposure of the dead. The Loylagan burial ground, discovered in 1973, contains numerous barrows which have been mapped but never geophysically investigated. This study presents an integrated geophysical investigation of mound No. 47 in order to detail and characterize its internal structure. Methods included magnetometry, electrical resistivity tomography (ERT), seismic refraction and multichannel analysis of surface waves (MASW). Results showed that the magnetometric surveys imaged a central circular anomaly, interpreted as the structure's core. ERT, seismic and MASW surveys identified a high‐resistivity/density upper layer, interpreted as stone fill, above an oval‐shaped, low‐resistivity/velocity zone at depths of 1.5–3 m, likely corresponding to a burial pit or chamber. While little remains of these dakhmas above ground, below ground still remains relict archaeology. This study highlights the value of multi‐method site investigations for non‐invasive heritage studies in Central Asia.

This study presents a comparative framework for seismic signal classification based on adaptive decomposition and Hilbert spectral analysis. A self-recorded vibration dataset containing approximately 4,500 one-second segments of pedestrian, bicycle, and vehicle activities was acquired at a 1 kHz sampling rate using a buried geophone sensor. The raw signals were preprocessed through direct-current removal, fourth-order Butterworth band-pass filtering (1–50 Hz), and min–max normalization to mitigate environmental and instrumental noise. Three decomposition methods—Variational Mode Decomposition (VMD), Empirical Mode Decomposition (EMD), and Matching Pursuit Decomposition (MPD) were independently applied to obtain narrow-band oscillatory components. The Hilbert Transform (HT) was subsequently employed to derive marginal-spectral features, including spectral energy, spectral entropy, and dominant frequency, preserving a clear physical correspondence to source–soil coupling characteristics. These features were used to train multiple classifiers, including Support Vector Machine, Logistic Regression, k-Nearest Neighbors, Gradient Boosting, Random Forest, and Extreme Gradient Boosting (XGBoost), under group-aware cross-validation to avoid trial-level data leakage. Experimental evaluation demonstrates that VMD-HT features provide the highest discriminative performance, achieving a test accuracy of 91.4% and a macro-F1 score of 0.89 using the Random Forest classifier, followed by XGBoost with 90.3% accuracy. The EMD-HT framework attained 87.5% accuracy (macro-F1 = 0.84), while MPD-HT achieved 84.2% accuracy (macro-F1 = 0.81). Ensemble classifiers consistently outperformed linear models, confirming the advantage of nonlinear decision boundaries for decomposed spectral representations. The observed results indicate that VMD yields superior mode separation and spectral compactness compared with EMD and MPD, leading to improved feature stability and reduced overlap across classes. The proposed framework thus offers a robust and physically interpretable approach for ground-vibration classification, with potential applications in intrusion detection, structural health monitoring, and environmental sensing.

In this study, a novel GO/Fe@TANG composite catalyst was designed by functionalizing graphene oxide (GO) with iron (Fe) and incorporating a hydrogen (H 2 ) trapping TANG COF.