H∞ Filtering in Neural Network Training and Pruning with Application to System Identification
H∞ Filtering in Neural Network Training and Pruning with Application to System Identification
复制标题
DOI：
10.1061/(asce)0887-3801(2007)21:1(47
复制
发表时间：
2007
影响因子：
6.9
通讯作者：
He-sheng Tang;S. Xue;Tadanobu Sato
中科院分区：
文献类型：
--
作者：
He-sheng Tang;S. Xue;Tadanobu Sato
提出了一种基于H-∞滤波算法的人工神经网络训练和剪枝方法。首先用H-∞滤波算法对神经网络进行训练，然后在训练的基础上去除一些不重要的权值。实验结果表明，该方法在不损失网络泛化能力的前提下，获得了更好的剪枝效果。对于给定的任意网络结构，它还提供了一种健壮的训练算法。通过建立滞回结构系统的人工神经网络模型，验证了该方法的实用性和有效性。
An efficient training and pruning methodology based on the H∞ filtering algorithm is proposed for artificial neural networks (ANNs). ANNs are first trained by the H∞ filtering algorithm and then some unimportant weights are removed based on the training. The results presented in the paper show that the proposed method provides better pruning results of the network without losing its generalization capacity. It also provides a robust training algorithm for given arbitrary network structures. The usefulness and effectiveness of the proposed methodology are demonstrated in developing an ANN model of a hysteretic structural system.