Data Driven Modeling
Data-driven modeling uses machine learning to create models of complex systems directly from observational data, bypassing the need for explicit mathematical formulations. Current research emphasizes improving model accuracy and robustness, particularly using deep learning architectures like recurrent neural networks (RNNs), physics-informed neural networks (PINNs), and generative adversarial networks (GANs), along with techniques like Koopman operator learning and sparse system identification. This approach is transforming various fields, enabling more accurate predictions and efficient control in applications ranging from urban infrastructure management and energy systems to materials science and autonomous systems.
Papers
Combining physics-based and data-driven techniques for reliable hybrid analysis and modeling using the corrective source term approach
Sindre Stenen Blakseth, Adil Rasheed, Trond Kvamsdal, Omer San
On the balance between the training time and interpretability of neural ODE for time series modelling
Yakov Golovanev, Alexander Hvatov