Deep Learning and Inverse Design in Plasmonic

Laser pulses can colour noble metals by inducing nanoparticles on their surface. The colours are linked to laser parameters and nanoparticles geometry. We apply deep learning to the direct prediction of colours from a laser parameter set or a nanoparticle particle distribution. A new method for inverse design via deep learning is also proposed to retrieve the appropriate laser parameters or nanoparticle distribution given the desired colour.