Neural Networks for Optimization and Synthesis

This paper shows how neural networks are ideally suited for solving the inverse electromagnetic field problems quickly. However, they need to rely on classical deterministic techniques for training.