【摘要】 以典型的前馈网络设计为例,对网络学习中神经元激活函数类型优化的重要性做了进一步的探讨。理论分析和实际算例显示:神经元激活函数类型相当于信号分解中的基函数,如果不对它进行优化,就不能获得理想的泛化性能,特别是外推泛化性能。更多还原

【Abstract】 Perfect artificial neural network (ANN) learning should include the optimization of neural activation function types, and the tradition of optimizing the network weights only in ANN learning is not consistent with biology. As an instance with feed-forward ANN, this paper has further discussed the importance of optimizing neuronal activation function types. With several theoretical analyses and practical examples, it has been shown that like the basis function in signal decomposition, the optimization of it is very important to ANN generalization performance.更多还原