Fast Surrogate

Fast surrogates are data-driven models designed to rapidly approximate the outputs of computationally expensive high-fidelity simulations, primarily focusing on partial differential equations (PDEs) and other complex physical systems. Current research emphasizes the use of neural networks, particularly graph neural networks (GNNs) and transformers, often incorporating multi-scale time-stepping or physics-informed approaches to enhance accuracy and efficiency. These advancements enable faster design optimization, real-time decision-making in applications like treatment planning, and accelerate scientific discovery across diverse fields including fluid dynamics, material science, and urban planning.