Get Started
Home
Topics
Search
Library
Research questionHow can PINN transfer learning recover physical parameters when source and target PDEs differ?Transferred PINNs may fit the target solution field accurately while compensating for errors in the underlying physical parameters. Differences between source and target physics make it difficult to reuse learned representations without undermining parameter identifiability.
AI
Machine Learning
Neural and Evolutionary Computing
Research Paper
Latest papersRecent research connected to this question, newest first.Target-Guided Selective Reweighting for Physics-Informed Neural Network Inverse Problems: A Transfer Learning ApproachThe evidence covers zero-source high-Péclet inflow–outflow, Allen–Cahn-to-Burgers cross-PDE transfer, and 5% noisy reaction–diffusion inverse problems. The proposed TGSR-PINN transfers source network weights and biases, initializes target physical parameters independently, and uses short target adaptation to selectively adjust neurons; reported results show improved parameter recovery with low field error in these settings.research paper · Sep 3, 2026
Related questions
How can periodic PINNs reduce the cost of high-order spatial derivatives without sacrificing solution accuracy?How can parametric PDE solution operators extrapolate to unseen regimes without failing silently?How can federated PDE learning define and measure client heterogeneity across different equations?How can physics-informed neural networks handle stiff, multiscale ion-electronic transport while supporting inverse estimation?