Transient breathing dynamics during extinction of dissipative solitons in mode‑locked fiber lasers

The utilization of the dispersive Fourier transformation approach has enabled comprehensive observation of the birth process of dissipative solitons in fiber lasers.However,there is still a dearth of deep understanding regarding the extinction process of dissipative solitons.In this study,we have utilized a combination of experimental and numerical techniques to thoroughly examine the breathing dynamics of dissipative solitons during the extinction process in an Er-doped mode-locked fiber laser.The results demonstrate that the transient breathing dynamics have a substantial impact on the extinction stage of both steady-state and breathing-state dissipative solitons.The duration of transient breathing exhibits a high degree of sensitivity to variations in pump power.Numerical simulations are utilized to produce analogous breathing dynamics within the framework of a model that integrates equations characterizing the population inversion in a mode-locked laser.These results corroborate the role of Q-switching instability in the onset of breathing oscillations.Furthermore,these findings offer new possibilities for the advancement of various operational frameworks for ultrafast lasers.

基于纳米散射结构的可集成光学神经网络及其逆向设计

基于集成光学和硅光子学的光学神经网络硬件有很多优势:集成度高、速度快并且与CMOS工艺兼容.然而,目前的集成光学神经网络尺寸较大,很难扩展到大量神经元(>1000),实现大规模计算.本文提出了一种基于光学散射单元的神经网络硬件架构,除了具备一般光学神经网络的优势外,突出的优势是尺寸小,易于大规模扩展.光学散射单元允许光在一个小区域中发生散射,通过逆向设计散射区域结构,实现目标的计算功能.光学散射单元在一个很小的尺寸下,提供了很大的优化自由度,研究表明要实现一个4 4的矩阵乘法,计算单元尺寸只需要4 4μm^2.基于光学散射单元,本文设计了光学神经网络,在经典图像识别测试集MNIST上实现了97.1%的准确度.此外,这种光学散射单元还可以适用于相干光和非相干光.本研究提供了一个新的光学神经网络架构,能在不影响效率和功能下减小神经网络硬件尺寸.