Lightweight Federated Learning Cuts Power Grid Edge Computing Redundancy by 77 Percent
Engineers developed a lightweight multi-task federated learning framework that trims model parameter redundancy by over 77 percent while maintaining accuracy ...
Engineers developed a lightweight multi-task federated learning framework that trims model parameter redundancy by over 77 percent while maintaining accuracy ...
A comparative evaluation of electricity demand in Panama reveals that a simpler linear regression model surpassed a multi-layer neural network ...
Researchers developed a hybrid access control architecture combining neural collaborative filtering with deterministic compliance gating to govern artificial intelligence agent ...
A review of cryptographic architectures for smart grid cyber-physical systems reveals that balancing robust security with the computational constraints of ...
A newly developed personalized federated learning framework enables accurate energy appliance identification across diverse households while maintaining consumer data privacy.
This evidence-based review synthesizes current AI capabilities for smart-grid fault detection while identifying maturity gaps and future research priorities.
A hybrid deep learning model combining recurrent neural networks and long short-term memory achieved a 0.963 assessment accuracy and a ...
A new multi-input LSTM framework improves day-ahead residential load forecasting by independently encoding daily and weekly temporal patterns, achieving the ...
A study of daily market data from 2010 to 2024 reveals that European Union carbon-pricing signals predict global clean energy ...
A comprehensive review examines how multi-objective reinforcement learning can guide artificial intelligence agents toward explicit ethical alignment through five foundational ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine