Tag
PhysAttNet is a physics-informed attention framework that augments lightweight CNN forecasters with domain-guided regularization to improve accuracy and generalization in industrial and astrophysical time series forecasting.