Speaker adaptation with autonomous model complexity control by MDL principle
نویسندگان
چکیده
A speaker adaptation method for continuous density HMMs, which performs well for any amount of data for adaptation, is proposed. This method estimates shift parameters for the means of Gaussian mixture components in the HMM. Each shift parameter is shared by more than one Gaussian components. Many sets of shift parameters with various degree of sharing are prepared, and the set with the appropriate complexity for the given amount of data is selected using Minimum Description Length (MDL) principle. Unlike previous similar works, the proposed method needs no control parameters for selecting models. A series of 5000-word recognition experiments have demonstrated the e ectiveness of this new method.
منابع مشابه
Optimal on-line Bayesian model selection for speaker adaptation
In this paper, we show how to accomodate a Bayesian variant of Rissanen’s MDL into on-line Bayesian adaptation to control both model structural complexity and parameterization complexity to best fit an available amount of adaptation data, the goal being minimization of resulting recognition error. An efficient bottom-up dynamic programming based pruning algorithm is developed for selecting mode...
متن کاملOnline Bayesian tree-structured transformation of HMMs with optimal model selection for speaker adaptation
This paper presents a new recursive Bayesian learning approach for transformation parameter estimation in speaker adaptation. Our goal is to incrementally transform or adapt a set of hidden Markov model (HMM) parameters for a new speaker and gain large performance improvement from a small amount of adaptation data. By constructing a clustering tree of HMM Gaussian mixture components, the linear...
متن کاملMDL-Based Cluster Number Decision Methods for Speaker Clustering and MLLR Adaptation
Speaker clustering is one of the major methods for speaker adaptation. MLLR (Maximum Likelihood Linear Regression) adaptation using transformation matrices corresponding to phone classes/clusters is another useful method especially when the length of utterances for adaptation is limited. In these methods, how to decide the most appropriate number of clusters is an important research issue. This...
متن کاملA New Minimum Description Length
The minimum description length(MDL) method is one of the pioneer methods of parametric order estimation with a wide range of applications. We investigate the definition of two-stage MDL for parametric linear model sets and exhibit some drawbacks of the theory behind the existing MDL. We introduce a new description length which is inspired by the Kolmogorov complexity principle.
متن کاملInferencing in Database Semantics
As a computational model of natural language communication, Database Semantics1 (DBS) includes a hearer mode and a speaker mode. For the content to be mapped into language expressions, the speaker mode requires an autonomous control. The control is driven by the overall task of maintaining the agent in a state of balance by connecting the interfaces for recognition with those for action. This p...
متن کامل