Outperforming Self-Attention Mechanisms in Solar Irradiance Forecasting via Physics-Guided Neural Networks
This research challenges the dominance of complex Transformer models by introducing a lightweight, physics-guided CNN-BiLSTM framework for solar irradiance f...