摘要
基本粒子滤波算法已被成功用于训练神经网络,但该算法在建议分布的选择上并没有考虑当前时刻观测值的影响,针对该问题提出在神经网络训练中,使用迭代扩展卡尔曼滤波器来生成建议分布.由于迭代扩展卡尔曼滤波器在传递近似建议分布的均值和协方差的过程中,充分利用了观测值信息,从而可以更好地描述神经网络权值的后验概率分布.实验结果表明,在训练神经网络时,迭代扩展卡尔曼滤波器作为建议分布的粒子滤波算法训练性能明显优于基本粒子滤波算法及扩展卡尔曼粒子滤波算法(EKPF).
The generic particle filter has been applied with success to neural network training, but the proposal distribution chosen by the generic particle filter doesn＇t incorporate the latest observations which can deteriorate the performance of the algorithm. In this paper, we propose to use the iterated extended Kalman filter to generate proposal distribution in particle filtering framework. The iterated extended Kalman filter can make efficient use of the latest observation, and the generated proposal distribution can approximate the posterior distribution of neural network weights much better and improve the performance of particle filter. The experimental results show that the proposed particle filter outperforms the generic particle filter and the EKPF.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第8期103-107,共5页
Microelectronics & Computer
