Fast training of accurate physics-informed neural networks without gradient descent
ICLROral2026
TL;DR
Our approach - Frozen-PINNs addresses longstanding training and accuracy bottlenecks of Physics-Informed Neural Networks (PINNs) and makes PINNs highly realize high-precision, temporal causality, and extremely fast training.
Opening excerpt from the authors’ abstract. source
Read the paper
Topics
causality physics causal