Efficient optical computing with exciton-polaritons
Michał Matuszewski, Institute of Physics, Polish Academy of Sciences
For decades, optical computing has been considered a promising alternative to electronics for high performance computing. One of the major obstacles on the way to practical applications of all-optical computing has been the weakness of optical nonlinearity, or interactions between photons. We propose that exciton-polaritons, hybrid particles of light and matter existing in semiconductor microcavities, can be used to construct optical neural networks characterized by ultrastrong nonlinearity. In result, they can provide the missing element necessary for the implementation of optical systems outperforming electronics by orders of magnitude.