P‐252: Late‐News Poster: Deep Learning based Inverse Design Technology for White Organic Light‐emitting Diode Thin‐film Structure

Abstract

We propose a reverse design technique for WOLEDs using an artificial neural network. The inverse neural network is trained by combining forward and inverse neural networks and predicts the thin‐film structure from the out‐coupling spectrum. The predicted structures have a low error rate of less than 9 percent, demonstrating effectiveness of the inverse neural network.