MXene-Based Broadband Ultrafast Nonlinear Activator for Optical Computing

摘要

Optical neural networks (ONNs) are particularly advantageous owing totheir inherent parallelism and low energy consumption. However, one of theobstacles to the implementation of ONNs is the lack of optical nonlinearity. Inthis study, optical nonlinear activators for ONNs are prepared by combiningTi_3C_2T_x Mxene with microfibers and their principles are verified. Activationfunctions obtained from experimental measurements are used to simulatemulticlassification and super-resolution reconstruction tasks with performancecomparable to that of activation functions commonly used in computers. Fournecessary criteria are proposed and validated for evaluating the performanceof the nonlinear activator: recovery time, deviation from linearity, the activationfunction close to identity mapping, and reconfigurability of the configuration.Theoretically, the nonlinear activator can compute 100 times faster than commonlyused electronic computers and can be used as a nonlinear activationunit for ONNs to help the integration of ONNs with artificial intelligence.