中国邮电高校学报(英文) ›› 2009, Vol. 16 ›› Issue (3): 126-130.doi: 10.1016/S1005-8885(08)60238-1

• Others • 上一篇    

Improved discriminative training for generative model

吴娅辉,郭军,刘刚   

  1. Laboratory of Pattern Recognition and Intelligent System, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2009-06-30
  • 通讯作者: 吴娅辉

Improved discriminative training for generative model

WU Ya-hui, GUO Jun, LIU Gang   

  1. Laboratory of Pattern Recognition and Intelligent System, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:1900-01-01 Revised:1900-01-01 Online:2009-06-30
  • Contact: WU Ya-hui

摘要:

This article proposes a model combination method to enhance the discriminability of the generative model. Generative and discriminative models have different optimization objectives and have their own advantages and drawbacks. The method proposed in this article intends to strike a balance between the two models mentioned above. It extracts the discriminative parameter from the generative model and generates a new model based on a multi-model combination. The weight for combining is determined by the ratio of the inter-variance to the intra-variance of the classes. The higher the ratio is, the greater the weight is, and the more discriminative the model will be. Experiments on speech recognition demonstrate that the performance of the new model outperforms the model trained with the traditional generative method.

关键词:

model;combination,;discriminability,;generative;model,;training,;speech;recognition

Abstract:

This article proposes a model combination method to enhance the discriminability of the generative model. Generative and discriminative models have different optimization objectives and have their own advantages and drawbacks. The method proposed in this article intends to strike a balance between the two models mentioned above. It extracts the discriminative parameter from the generative model and generates a new model based on a multi-model combination. The weight for combining is determined by the ratio of the inter-variance to the intra-variance of the classes. The higher the ratio is, the greater the weight is, and the more discriminative the model will be. Experiments on speech recognition demonstrate that the performance of the new model outperforms the model trained with the traditional generative method.

Key words:

model combination;discriminability;generative model;training;speech recognition