Non Equilibrium Steady State

Non-equilibrium steady states (NESS) describe systems continuously driven out of equilibrium yet maintaining statistically stationary properties. Current research focuses on developing accurate and efficient methods for characterizing NESS, including novel deep learning architectures like physics-informed neural networks and variational force projection methods, as well as optimization-based approaches that leverage stochastic processes. These advancements are crucial for understanding complex systems across diverse fields, from biophysics (modeling protein dynamics) to machine learning (analyzing neural network behavior), enabling more accurate modeling and prediction of their long-term behavior.