Researchers present Green-ELM, a non-iterative neural architecture that uses a closed-form analytic solution to optimize the output layer, reducing training time and increasing accuracy on MNIST and Fashion-MNIST datasets. This approach has potential applications in real-time Edge AI.