The artificial neural network-spiking neural network(ANN-SNN)conversion,as an efficient algorithm for deep SNNs training,promotes the performance of shallow SNNs,and expands the application in various tasks.However,th...The artificial neural network-spiking neural network(ANN-SNN)conversion,as an efficient algorithm for deep SNNs training,promotes the performance of shallow SNNs,and expands the application in various tasks.However,the existing conversion methods still face the problem of large conversion error within low conversion time steps.In this paper,a heuristic symmetric-threshold rectified linear unit(stReLU)activation function for ANNs is proposed,based on the intrinsically different responses between the integrate-and-fire(IF)neurons in SNNs and the activation functions in ANNs.The negative threshold in stReLU can guarantee the conversion of negative activations,and the symmetric thresholds enable positive error to offset negative error between activation value and spike firing rate,thus reducing the conversion error from ANNs to SNNs.The lossless conversion from ANNs with stReLU to SNNs is demonstrated by theoretical formulation.By contrasting stReLU with asymmetric-threshold LeakyReLU and threshold ReLU,the effectiveness of symmetric thresholds is further explored.The results show that ANNs with stReLU can decrease the conversion error and achieve nearly lossless conversion based on the MNIST,Fashion-MNIST,and CIFAR10 datasets,with 6×to 250 speedup compared with other methods.Moreover,the comparison of energy consumption between ANNs and SNNs indicates that this novel conversion algorithm can also significantly reduce energy consumption.展开更多
The modeling of hydrocarbon selectivity and CO conversion of the Fischer-Tropsch synthesis over Fe-Ni/Al2O3 catalyst by using coupled artificial neural networks (ANN) and design of experiment (DOE) approaches were inv...The modeling of hydrocarbon selectivity and CO conversion of the Fischer-Tropsch synthesis over Fe-Ni/Al2O3 catalyst by using coupled artificial neural networks (ANN) and design of experiment (DOE) approaches were investigated. The variable parameters for modeling consisted of the pressure range between 2 and 10 bar and the temperature range of 523-573 K. After training of data by ANN and determination of DOE points by central composite design (CCD), the results were compiled together for producing simulated data used in the response surface method (RSM). The RSM was used as an applied mathematics model to dem on strate the CO conversi on and selectivity of hydrocarbons depende nee on the CO hydrogenation conditions. The results indicated that CO conversion and Cg selectivity increased with rising both temperature and pressure. The methane selectivity showed upward trend as the temperature in creased. It also in creased by decreasing pressure. Finally, the optimization of the catalytic process was carried out and conditions with maximum desired product were obtained. A comparison of experimental values and RSM values show that the RSM equations are able to predict the behavior of experimental data.展开更多
基金the National Key Research and Development Program of China(No.2020AAA0105900)National Natural Science Foundation of China(No.62236007)Zhejiang Lab,China(No.2021KC0AC01).
文摘The artificial neural network-spiking neural network(ANN-SNN)conversion,as an efficient algorithm for deep SNNs training,promotes the performance of shallow SNNs,and expands the application in various tasks.However,the existing conversion methods still face the problem of large conversion error within low conversion time steps.In this paper,a heuristic symmetric-threshold rectified linear unit(stReLU)activation function for ANNs is proposed,based on the intrinsically different responses between the integrate-and-fire(IF)neurons in SNNs and the activation functions in ANNs.The negative threshold in stReLU can guarantee the conversion of negative activations,and the symmetric thresholds enable positive error to offset negative error between activation value and spike firing rate,thus reducing the conversion error from ANNs to SNNs.The lossless conversion from ANNs with stReLU to SNNs is demonstrated by theoretical formulation.By contrasting stReLU with asymmetric-threshold LeakyReLU and threshold ReLU,the effectiveness of symmetric thresholds is further explored.The results show that ANNs with stReLU can decrease the conversion error and achieve nearly lossless conversion based on the MNIST,Fashion-MNIST,and CIFAR10 datasets,with 6×to 250 speedup compared with other methods.Moreover,the comparison of energy consumption between ANNs and SNNs indicates that this novel conversion algorithm can also significantly reduce energy consumption.
文摘The modeling of hydrocarbon selectivity and CO conversion of the Fischer-Tropsch synthesis over Fe-Ni/Al2O3 catalyst by using coupled artificial neural networks (ANN) and design of experiment (DOE) approaches were investigated. The variable parameters for modeling consisted of the pressure range between 2 and 10 bar and the temperature range of 523-573 K. After training of data by ANN and determination of DOE points by central composite design (CCD), the results were compiled together for producing simulated data used in the response surface method (RSM). The RSM was used as an applied mathematics model to dem on strate the CO conversi on and selectivity of hydrocarbons depende nee on the CO hydrogenation conditions. The results indicated that CO conversion and Cg selectivity increased with rising both temperature and pressure. The methane selectivity showed upward trend as the temperature in creased. It also in creased by decreasing pressure. Finally, the optimization of the catalytic process was carried out and conditions with maximum desired product were obtained. A comparison of experimental values and RSM values show that the RSM equations are able to predict the behavior of experimental data.
