Arbitrary-scale super-resolution (ASSR) aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs with arbitrary upsampling factors using a single model, addressing the limitations of traditional SR methods constrained to fixed-scale factors (\textit{e.g.}, $\times$ 2). Recent advances leveraging implicit neural representation (INR) have achieved great progress by modeling coordinate-to-pixel mappings. However, the efficiency of these methods may suffer from repeated upsampling and decoding, while their reconstruction fidelity and quality are constrained by the intrinsic representational limitations of coordinate-based functions. To address these challenges, we propose a novel ContinuousSR framework with a Pixel-to-Gaussian paradigm, which explicitly reconstructs 2D continuous HR signals from LR images using Gaussian Splatting. This approach eliminates the need for time-consuming upsampling and decoding, enabling extremely fast ASSR. Once the Gaussian field is built in a single pass, ContinuousSR can perform arbitrary-scale rendering in just 1ms per scale. Our method introduces several key innovations. Through statistical analysis, we uncover the Deep Gaussian Prior (DGP) and propose DGP-Driven Covariance Weighting, which dynamically optimizes covariance via adaptive weighting. Additionally, we present Adaptive Position Drifting, which refines the positional distribution of the Gaussian space based on image content, further enhancing reconstruction quality. Extensive experiments on seven benchmarks demonstrate that our ContinuousSR delivers significant improvements in SR quality across all scales, with an impressive 19.5× speedup when continuously upsampling an image across forty scales.
Cyanobacteria is an indicator for freshwater ecosystem health. This study aims to (a) simulate the spatial distribution of cyanobacteria dynamics; and (b) optimize machine learning (ML) models with explainable artificial intelligence for cyanobacteria prediction in tropical freshwater lake. The water quality parameters (e.g. cyanobacteria concentration (μg/L), chlorophyll concentration (μg/L), turbidity, and cell count) were analysed from eleven sampling stations in Tasik Kenyir. The cyanobacteria dynamics was modelled via GeoPhyton analysis. Linear Regression (LR), Artificial Neural Network (ANN), and Extreme Gradient Boosting (XGBoost), were optimized to assess predictive performance. Strong temporal variability observed major peak in August and October 2024. High predictive capabilities were obtained by LR (training R2 = 0.9435; testing R2 = 0.9902), ANN (training R2 = 0.9660; testing R2 = 0.9739), and XGBoost (training R2 = 1.000; testing R2 = 0.7468). Feature importance analysis using SHapley Additive exPlanations (SHAP) identified cell count and chlorophyll as the dominant predictors.
In this study, the attenuation performance of graded-index multimode optical fibers (GIMFs) is investigated and optimized using the Taguchi design of experiment (DoE) methodology. Six key control factors including core diameter, cladding diameter, minimum wavelength, maximum wavelength, number of wavelengths, and number of supported modes were examined to determine their relative influence on fiber attenuation. Analysis of both the signal-to-noise (S/N) ratio and data mean shown that the control factors affect attenuation in descending order of significance as follows: number of wavelengths, minimum wavelength, maximum wavelength, number of modes, core diameter, and cladding diameter. This ranking provided a systematic basis for parameter tuning and enabled the design of a GIMF structure achieving a minimized attenuation value of 0.2080 dB/Km, without relying on conventional trial and error optimization. A tuning process was carried out using the Taguchi-derived factor rankings, and the resulting optimized design was validated via COMSOL Multiphysics simulations, demonstrating stable mode confinement, controlled dispersion behavior, and negligible material absorption loss. The finding confirm that the Taguchi method offers an efficient, reliable, and structured optimization strategy for GIMF design, reducing development complexity and computational effort. The methodology presented can be extended to optimize additional performance metrics and applied to next generation’s multimode and mode division multiplexed fiber systems.
This investigation examines quasi-periodic oscillations (QPOs) in two quantum-corrected black hole (BH) space-times that preserve general covariance while incorporating quantum gravitational effects through a dimensionless parameter zeta. We combine analytical derivations of epicyclic frequencies with comprehensive numerical simulations of Bondi-Hoyle-Lyttleton (BHL) accretion to explore how quantum corrections manifest in observable astrophysical phenomena. Using a fiducial BH mass of M = 10M ae representative of stellar-mass X-ray binaries, we demonstrate that the two models exhibit fundamentally different behaviors: Model-I modifies both temporal and radial metric components, leading to innermost stable circular orbit migration proportional to zeta 4and dramatic stagnation point evolution from 27M to 5M as quantum corrections strengthen. Model-II preserves the classical temporal component while altering only spatial geometry, maintaining constant stagnation points and stable cavity structures throughout the parameter range. Our numerical simulations reveal distinct QPO generation mechanisms, with Model-I showing systematic frequency evolution and cavity shrinkage that suppresses oscillations for zeta >= 3M, while Model-II maintains stable low-frequency modes up to zeta >= 5M. Power spectral density analyzes demonstrate characteristic frequency ratios (3 : 2, 2 : 1, 5 : 3) consistent with observations from X-ray binaries, providing specific targets for discriminating between quantum correction scenarios. The hydrodynamically derived constraints ( 4M) show remarkable agreement with independent Event Horizon Telescope limits for M87* and Sgr A*, validating our theoretical framework through multiple observational channels. These results establish QPO frequency analysis as a probe for detecting quantum gravitational effects in astrophysical BHs and demonstrate the complementary nature of timing and imaging observations in constraining fundamental physics.
In a recent study [1], authors introduced a new class of exact space-times in Einstein's gravity, which are Kerr black holes immersed in an external uniform magnetic field that is oriented along the rotational axis. Motivated by this work, we investigate a Kerr-like black hole solution with a cloud of strings surrounded by a uniform magnetic field. For the zero rotation case, the space-time reduces to the Schwarzschild-Bertotti-Robinson black hole with a cloud of strings. Moreover, for zero magnetic field, the metrics simplify to a Kerr-like black hole surrounded by a cloud of strings, and its static counterpart reduces to the Schwarzschild black hole with a cloud of strings.
