The field of physics-informed neural networks is rapidly advancing, with a focus on developing innovative methods for solving complex systems governed by partial differential equations. Recent developments have centered around improving the accuracy, stability, and physical consistency of these models, particularly in areas such as cardiac electrophysiology, radiative transfer, and multiphysics simulations. Notable advancements include the incorporation of physical constraints into neural network architectures, the development of asymptotic-preserving neural networks, and the use of graph neural networks to capture multi-scale features. These innovations have the potential to significantly improve the efficiency and scalability of simulations in various fields. Noteworthy papers include: Physics-Informed Neural Operators for Cardiac Electrophysiology, which proposes a novel approach to solve PDE problems in cardiac electrophysiology. PEGNet: A Physics-Embedded Graph Network for Long-Term Stable Multiphysics Simulation, which introduces a redesigned graph neural network architecture that incorporates PDE-guided message passing.