1616 papers
This paper proposes a secure and scalable Decentralised Federated Learning framework integrated with a hybrid TCN-BiLSTM architecture to achieve high-accuracy load forecasting in dynamic microgrids while overcoming the privacy, reliability, and resource constraints of traditional centralized approaches.
This paper demonstrates that low-complexity state-of-polarization monitoring on live, buried fiber-to-the-home networks in the seismically active Campi Flegrei area can successfully detect and characterize multiple earthquakes, validating the potential of existing telecommunication infrastructure for complementary seismic sensing.
This study demonstrates that helically winding hemp rope around geogrid ribs significantly enhances interface shear resistance in expansive soils, with a two-turn configuration (HWG-2) achieving a 14.4% increase in peak shear force at 90 kPa compared to unmodified geogrids.
This study integrates in-situ bearing plate tests and 2D discrete-element simulations at the Gushui Hydropower Station to reveal that while weathering-based grouping provides baseline deformation parameters, local structural discontinuities—specifically the E2-2 fault—are the dominant controls on irreversible deformation and energy evolution in basalt rock masses.
This study presents a computationally efficient model based on effective medium theory that accurately predicts gas diffusivity in fibrous gas diffusion layers by transforming standard porosimetry data into a representative pore structure and deriving a compact correlation validated against experimental and literature data.
This study utilizes Association Rules Mining on 2016–2023 FARS data to identify distinct multifactorial patterns linking land use, lighting, and roadway geometry to fatal U-turn crashes in rural and urban settings, thereby informing targeted safety countermeasures.
This study validates the hybrid model NeuralGCM by demonstrating that it intensifies unprecedented African heat extremes through regionally appropriate physical mechanisms, proving that differentiability can serve as a robust instrument for verifying the physical realism of learned components in extreme weather forecasting.
This study introduces on-chip nanoplasma switches (NPMS) based on gallium nitride and silicon carbide as a superior alternative to semiconductor diodes for adaptive electromagnetic protection, offering enhanced thermal tolerance, broader bandwidth, and robust shielding against high-power microwave threats for next-generation RF front ends.
This study employs an explainable XGBoost-SHAP framework to analyze how urban three-dimensional morphology influences seismic vulnerability in Jiangyou City, revealing that spatial crowding is the most critical factor and that specific morphological configurations exhibit distinct nonlinear effects and spatial heterogeneity that can guide resilient urban planning.
This study reframes critical heat flux prediction as an experimental evidence transport problem by demonstrating that whole-campaign separation reveals significant generalization limits in existing models, thereby establishing an offline, campaign-initialized adaptive surrogate framework that leverages conformal uncertainty, transfer learning, and sparse Bayesian updating to determine when historical data is transferable and when limited new measurements are required for reliable predictions.