Here we create a user-friendly interface (UI) for a well-trained AI model to help the user quickly find out the optical constants of the DRC materials according to the desired equilibrium temperature. This AI-based Deep Generative Model performs the inverse design process in a probabilistic way. By employing an encoder-decoder configuration, our deep generative model compresses the thin film structure, equilibrium temperature, and cooling power into a latent space. Next, the AI will sample the latent space with certain prior distribution, and function the stochastic latent variables as codes, from which the designed candidates are generated upon required equilibrium temperature in the decoding process.

AI PERFORMANCE FOR DRC MATERIAL INVERSE DESIGN

Inverse Design of Extraordinary Radiative Cooling Materials Using Deep Generative Model