Enhancing Stability of Physics-Informed Neural Network Training Through Saddle-Point Reformulation

Open in new window