This paper examines the use of interdependencies of parameter classes in transformation-based speaker adaptation algorithms such as maximum likelihood linear regression (MLLR). In transformation-based adaptation, increasing the number of transformation classes can provide more detailed information for adaptation, but at the expense of greater estimation error with small amounts of data. In this...

In this paper, we introduce the HMM-state sequence confusion characteristics as prior knowledge into the framework of MLLR to relax the transformation and reduce the risks of over-training when adaptation data size is small. There are two issues to be addressed as follows: first, how to estimate such confusion information reliably; second how to use the information in refining the estimation of...

The purpose of this work is to show how recent developments in cepstral-based systems for speaker recognition can be leveraged for the use of Maximum Likelihood Linear Regression (MLLR) transforms. Speaker recognition systems based on MLLR transforms have shown to be greatly beneficial in combination with standard systems, but most of the advances in speaker modeling techniques have been implem...

Transformation Sharing Strategies for MLLR Speaker Adaptation Arindam Mandal Chair of the Supervisory Committee: Professor Mari Ostendorf Electrical Engineering Maximum Likelihood Linear Regression (MLLR) estimates linear transformations of automatic speech recognition (ASR) parameters and has achieved significant performance improvements in speaker-independent ASR systems by adapting to target...

In this paper, we propose two techniques to extend the recently introduced global Maximum Likelihood Linear Regression (MLLR) transformation (i.e. super-vector) based m-vector system for speaker verification into a multi-class MLLR mvector system in the Universal Background Model (UBM) framework. In the first method, Gaussian mean vectors of the UBM are first grouped into several classes using ...

We studied the effect of MLLR adaptation with Spanishaccented English to understand the strengths and weaknesses of MLLR with unseen foreign accents. We trained a global MLLR transform on 10 adaptation sentences per speaker, giving a 3.4% absolute decrease in phone error rate. We then studied the pattern of improvements across phones and phone classes. Phones that improved the least tended to b...

This paper presents an eigenspace-based fast speaker adaptation approach which can improve the modeling accuracy of the conventional maximum likelihood linear regression (MLLR) techniques when only very limited adaptation data is available. The proposed eigenspace-based MLLR approach was developed by introducing a priori knowledge analysis on the training speakers via PCA, so as to construct an...

In this paper a novel speech feature generationbased acoustic model training method is proposed. For decades, speaker adaptation methods have been widely used. All existing adaptation methods need adaptation data. However, our proposed method creates speaker-independent acoustic models that cover not only known but also unknown speakers. We do this by adopting inverse maximum likelihood linear ...

We introduce a strategy for modeling speaker variability in speaker adaptation based on maximum likelihood linear regression (MLLR). The approach uses a speaker clustering procedure that models speaker variability by partitioning a large corpus of speakers in the eigenspace of their MLLR transformations and learning clusterspecific regression class tree structures. We present experiments showin...

One of the most popular approaches to parameter adaptation in hidden Markov model (HMM) based systems is the maximum likelihood linear regression (MLLR) technique. In our previous work, we proposed factored MLLR (FMLLR) where an MLLR parameter is defined as a function of a control parameter vector. We presented a method to train the FMLLR parameters based on a general framework of the expectati...