In this webinar hosted by the OSA Photonic Metamaterials Technical Group, Dr. Willie Padilla of Duke University will provide an overview on the emergence of machine learning/deep learning applied to the study of metasurfaces, including inverse design. Dr. Padilla’s talk will be motivated by illustrating the challenges and opportunities of all-dielectric metasurfaces contrasted to those based on metallic metasurfaces. Four different neural network architectures will be explored and the performance of each will be detailed, with results from the highest performing model shown explicitly. A solution to the inverse model will be presented during the webinar, which offers significant opportunities for design of advanced structured materials for challenging applications.
What You Will Learn:
• Challenges and opportunities in all-dielectric metasurfaces
• Overview of machine learning for metasurface design
• Efficient inverse design solution
Who Should Attend:
• Research scientists and engineers in university, government and industrial laboratories
• Master and PhD students