Understanding and mitigating gradient pathologies in physics-informed neural networks

Open in new window