In this work,we develop a stochastic gradient descent method for the computational optimal design of random rough surfaces in thin-film solar cells.We formulate the design problems as random PDE-constrained optimizati...In this work,we develop a stochastic gradient descent method for the computational optimal design of random rough surfaces in thin-film solar cells.We formulate the design problems as random PDE-constrained optimization problems and seek the optimal statistical parameters for the random surfaces.The optimizations at fixed frequency as well as at multiple frequencies and multiple incident angles are investigated.To evaluate the gradient of the objective function,we derive the shape derivatives for the interfaces and apply the adjoint state method to perform the computation.The stochastic gradient descent method evaluates the gradient of the objective function only at a few samples for each iteration,which reduces the computational cost significantly.Various numerical experiments are conducted to illustrate the efficiency of the method and significant increases of the absorptance for the optimal random structures.We also examine the convergence of the stochastic gradient descent algorithm theoretically and prove that the numerical method is convergent under certain assumptions for the random interfaces.展开更多
In this paper,we propose an accelerated stochastic variance reduction gradient method with a trust-region-like framework,referred as the NMSVRG-TR method.Based on NMSVRG,we incorporate a Katyusha-like acceleration ste...In this paper,we propose an accelerated stochastic variance reduction gradient method with a trust-region-like framework,referred as the NMSVRG-TR method.Based on NMSVRG,we incorporate a Katyusha-like acceleration step into the stochastic trust region scheme,which improves the convergence rate of the SVRG methods.Under appropriate assumptions,the linear convergence of the algorithm is provided for strongly convex objective functions.Numerical experiment results show that our algorithm is generally superior to some existing stochastic gradient methods.展开更多
光电混合人工智能计算芯片在人工智能应用中通过人工智能算法实现高速和高效的计算,其中光学神经网络(Optical Neural Networks,ONNs)算法在实现大量矩阵运算方面尤为重要.通过使用由马赫曾德尔干涉仪(Mach-Zehnder interferometers,MZI...光电混合人工智能计算芯片在人工智能应用中通过人工智能算法实现高速和高效的计算,其中光学神经网络(Optical Neural Networks,ONNs)算法在实现大量矩阵运算方面尤为重要.通过使用由马赫曾德尔干涉仪(Mach-Zehnder interferometers,MZI)搭建的快速傅里叶变换(Fast Fourier transform,FFT)型光学神经网络来实现手写数字的高精确度识别.在模型构建方面,利用奇异值分解将神经网络的线性层进行分解,从而实现数据降维,主要特征提取.在对该ONN的训练中,分别采用了带动量的随机梯度下降算法(Stochastic Gradient Descent with momentum,SGD with momentum)和均方根传递(Root Mean Square propagation,RMSprop)算法,分析了在不同训练算法下该ONN对手写数字的识别精度.此外,还深入剖析了两种训练算法背后的数学理论,探究造成两种训练算法实验结果差异的本质原因.最后,通过实验对比,发现RMSprop算法在FFT型光学神经网络上具有较高的识别精确度,达到97.4%;并且采用RMSprop算法的ONN计算速度优于SGD with momentum算法.展开更多
基金partially supported by the DOE grant DE-SC0022253the work of JL was partially supported by the NSF grant DMS-1719851 and DMS-2011148.
文摘In this work,we develop a stochastic gradient descent method for the computational optimal design of random rough surfaces in thin-film solar cells.We formulate the design problems as random PDE-constrained optimization problems and seek the optimal statistical parameters for the random surfaces.The optimizations at fixed frequency as well as at multiple frequencies and multiple incident angles are investigated.To evaluate the gradient of the objective function,we derive the shape derivatives for the interfaces and apply the adjoint state method to perform the computation.The stochastic gradient descent method evaluates the gradient of the objective function only at a few samples for each iteration,which reduces the computational cost significantly.Various numerical experiments are conducted to illustrate the efficiency of the method and significant increases of the absorptance for the optimal random structures.We also examine the convergence of the stochastic gradient descent algorithm theoretically and prove that the numerical method is convergent under certain assumptions for the random interfaces.
文摘In this paper,we propose an accelerated stochastic variance reduction gradient method with a trust-region-like framework,referred as the NMSVRG-TR method.Based on NMSVRG,we incorporate a Katyusha-like acceleration step into the stochastic trust region scheme,which improves the convergence rate of the SVRG methods.Under appropriate assumptions,the linear convergence of the algorithm is provided for strongly convex objective functions.Numerical experiment results show that our algorithm is generally superior to some existing stochastic gradient methods.
文摘光电混合人工智能计算芯片在人工智能应用中通过人工智能算法实现高速和高效的计算,其中光学神经网络(Optical Neural Networks,ONNs)算法在实现大量矩阵运算方面尤为重要.通过使用由马赫曾德尔干涉仪(Mach-Zehnder interferometers,MZI)搭建的快速傅里叶变换(Fast Fourier transform,FFT)型光学神经网络来实现手写数字的高精确度识别.在模型构建方面,利用奇异值分解将神经网络的线性层进行分解,从而实现数据降维,主要特征提取.在对该ONN的训练中,分别采用了带动量的随机梯度下降算法(Stochastic Gradient Descent with momentum,SGD with momentum)和均方根传递(Root Mean Square propagation,RMSprop)算法,分析了在不同训练算法下该ONN对手写数字的识别精度.此外,还深入剖析了两种训练算法背后的数学理论,探究造成两种训练算法实验结果差异的本质原因.最后,通过实验对比,发现RMSprop算法在FFT型光学神经网络上具有较高的识别精确度,达到97.4%;并且采用RMSprop算法的ONN计算速度优于SGD with momentum算法.
