Abstract: Explored the Hierarchical Mixture of Experts(HME) architecture for a substantial problem, that of text-dependent speaker identification. For a specific multiway classification, we propose a generalized Bernoulli density instead of the multinomial logit density. Using the proposed density and the HME along with the EM algorithm, we show that the system yields satisfactory performance and significantly fast training.

Key words: speaker identification, hierarchical mixture of experts, multiway classification

摘要： 探讨了层次混合专家(HME)模型在说话人辨认中的应用。对于一个多路分类问题，提出了一个推广的贝努利概率分布密度函数，取代早先用于HME中的多项式概率分布密度函数。利用提出的概率分布密度函数和HME模型，用EM学习算法对模块网络进行训练所得到的说话人辨认系统不仅具有良好的性能，而且具有非常快的训练速度。

关键词: 说话人辨认, 层次混合专家模型, 多路分类