In this paper we describe various fast and convenient implementations of Speaker Adaptive Training (SAT) for use in training when Maximum Likelihood Linear Regression (MLLR) is to be used in test time to adapt Gaussian means. The memory and disk requirements for most of these are similar to those for normal ML training; the computation in all cases is dominated by the need to compute the MLLR t...

This paper describes an extension of maximum likelihood linear regression (MLLR) to hidden semi-Markov model (HSMM) and presents an adaptation technique of phoneme/state duration for an HMM-based speech synthesis system using HSMMs. The HSMM-based MLLR technique can realize the simultaneous adaptation of output distributions and state duration distributions. We focus on describing mathematical ...

This paper presents three forms of linear transform based speaker adaptation that can give better performance than standard maximum likelihood linear regression (MLLR) adaptation. For unsupervised adaptation, a lattice-based technique is introduced which is compared to MLLR using confidence scores. For supervised adaptation, estimation of the adaptation matrices using the maximum mutual informa...

 Inspired by the success of least absolute shrinkage and selection operator (LASSO) in statistical learning, we propose an regularized maximum likelihood linear regression (MLLR) to estimate models with only a limited set of adaptation data to improve accuracy for automatic speech recognition, by regularizing the standard MLLR objective function with an constraint. The so-called LASSO MLLR is ...

Maximum-Likelihood Linear Regression (MLLR) and Constrained MLLR (CMLLR) have been recently used for feature extraction in speaker recognition. These systems use (C)MLLR transforms as features that are modeled with Support Vector Machines (SVM). This paper evaluates and compares several of these approaches for the NIST Speaker Recognition task. Single CMLLR and up to 4-phonetic-class MLLR trans...

This paper presents a novel target-driven MLLR adaptation algorithm with multiply layer structure, which is based on the thorough analysis of MLLR using the generation of regression class trees. The new algorithm is constructed on the targetdriven principal. It generates the regression class dynamically, basing on the outcome of the former MLLR transformation. The regression classes is defined ...

In this paper we explore the use of lattice-based information for unsupervised speaker adaptation. As initially formulated, maximum likelihood linear regression (MLLR) aims to linearly transform the means of the gaussian models in order to maximize the likelihood of the adaptation data given the correct hypothesis (supervised MLLR) or the decoded hypothesis (unsupervised MLLR). For the latter, ...

In this paper, we propose a speaker-verification system based on maximum likelihood linear regression (MLLR) super-vectors, for which speakers are characterized by m-vectors. These vectors are obtained by a uniform segmentation of the speaker MLLR super-vector using an overlapped sliding window. We consider three approaches for MLLR transformation, based on the conventional 1-best automatic tra...

The need for close to real time speech recognition has recently driven interest in fast LVCSR systems. Due to the time constraint, such systems often discard, where possible, sub-processes of the entire recognition process which demand relatively large amounts of computation and yield relatively small accuracy gains. This report focusses on such speed-accuracy tradeoffs with regard to speaker a...