Figure | Working principle of polaritonic neuromorphic computing. (IMAGE)
Caption
Figure | Working principle of polaritonic neuromorphic computing. a, The operating process of polariton neuromorphic computing on image recognition task. The image information is first encoded by an SLM into an intensity-modulated laser beam. This laser beam then non-resonantly excites the exciton polaritons, leading to polariton condensation, which serves as the output information. Finally, a linear regression scheme is applied in the output layer to obtain the desired results. b, A simplified schematic of the experimental setup used for loading experimental datasets and measuring the polariton response. In this figure, all optical elements, except for the lenses, mirrors, and filters, are labeled. Specifically, BS refers to the beam splitter. SLM refers to spatial light modulator. BBO is used as a frequency-doubling crystal. The Kerr medium used is carbon disulfide, and the axes of the polarizers are perpendicular to each other.
Credit
Yusong Gan, Ying Shi et al.
Usage Restrictions
Credit must be given to the creator.
License
CC BY