{"id":"arxiv/2603.14029","url":"https://arxiv.org/abs/2603.14029","tags":["math.NA"],"score":0,"title":"Energy Dissipation Preserving Feature-based DNN Galerkin Methods for Gradient Flows","domain":"math.NA","source":"arxiv","authors":["Tao Tang","Jiang Yang","Yuxiang Zhao","Quanhui Zhu"],"reviews":0,"subject":"math.NA","summary":"In recent years, deep learning methods, exemplified by Physics-Informed Neural Networks (PINNs), have been widely applied to the numerical solution of differential equations. However, these methods may suffer from limited accuracy, high training costs, and lack of robustness, particularly their inability to preserve the intrinsic physical structures of continuous PDE models, such as the energy dissipation property in gradient flow systems. To address these challenges, we propose a feature-based Deep Neural Network Galerkin (DNN-G) framework designed for structure-preserving simulations of…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-14","reactions":{}}