An Alternative Model for Mixtures of Experts
An Alternative Model for Mixtures of Experts
复制标题
DOI：
--
复制
发表时间：
1994
期刊：
影响因子：
--
通讯作者：
L. Xu;Michael I. Jordan;Geoffrey E. Hinton
中科院分区：
文献类型：
--
作者：
L. Xu;Michael I. Jordan;Geoffrey E. Hinton
我们提出了一种专家混合的替代模型，该模型使用不同的参数形式来表示门控网络。利用EM算法对改进后的模型进行训练。与早期的模型--通过EM或梯度上升训练--相比，不需要选择学习步长。我们报告了仿真实验，结果表明新的结构产生了更快的收敛速度。我们还将新模型应用于两个问题领域：分段非线性函数逼近和多个先前训练的分类器的组合。
We propose an alternative model for mixtures of experts which uses a different parametric form for the gating network. The modified model is trained by the EM algorithm. In comparison with earlier models--trained by either EM or gradient ascent--there is no need to select a learning stepsize. We report simulation experiments which show that the new architecture yields faster convergence. We also apply the new model to two problem domains: piecewise nonlinear function approximation and the combination of multiple previously trained classifiers.