This letter introduces a novel physics-informed approach for neural network-based 3-D electromagnetic modeling. The proposed method combines standard leap-frog time-stepping with neural network-driven ...