Constrained maximum likelihood linear regression for speaker adaptation
نویسندگان
چکیده
This paper proposes a new structure for use in MLLR adaptation aiming at constraining the transform for potentially better parameter estimation from sparse adaptation data. Motivations for the use of the new structure, and EM based parameter estimation are presented. Experimental results on Spoke3 of the Wall Street Journal task revealed that the proposed transformations outperform a full matrix for a small amount of adaptation data and performs equally well for large adaptation set. They also outperform diagonal transformations for all amounts of adaptation data.
منابع مشابه
Maximum Likelihood Linear Regression (MLLR) for ASR Severity Based Adaptation to Help Dysarthric Speakers
Automatic speech recognition (ASR) for dysarthric speakers is one of the most challenging research areas. The lack of corpus for dysarthric speakers makes it even more difficult. The speaker adaptation (SA) is an alternative solution to overcome the lack of dysarthric speech and enhance the performance of ASR. This paper introduces the Severity-based adaptation, using small amount of speech dat...
متن کاملThe Speaker Adaptation of an Acoustic Model
This paper deals with several adaptation techniques, which are of the importance in cases when the identity of a speaker is known and we want to recognize his speech. We are using three different methods, namely Maximum Apriori Probability adaptation, Maximum Likelihood Linear Regression and Constrained Maximum Likelihood Linear Regression. Each of the methods yields various benefits, therefore...
متن کاملImprovement of MLLR Speaker Adaptation Using a Novel Method
This paper presents a technical speaker adaptation method called WMLLR, which is based on maximum likelihood linear regression (MLLR). In MLLR, a linear regression-based transform which adapted the HMM mean vectors was calculated to maximize the likelihood of adaptation data. In this paper, the prior knowledge of the initial model is adequately incorporated into the adaptation. A series of spea...
متن کاملDiscounted likelihood linear regression for rapid speaker adaptation
The widely used maximum likelihood linear regression speaker adaptation procedure suffers from overtraining when used for rapid adaptation tasks in which the amount of adaptation data is severely limited. This is a well known difficulty associated with the expectation maximization algorithm. We use an information geometric analysis of the expectation maximization algorithm as an alternating min...
متن کاملFast speaker adaptation using eigenspace-based maximum likelihood linear regression
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...
متن کاملLinear feature space projections for speaker adaptation
We extend the well-known technique of constrained Maximum Likelihood Linear Regression (MLLR) to compute a projection (instead of a full rank transformation) on the feature vectors of the adaptation data. We model the projected features with phone-dependent Gaussian distributions and also model the complement of the projected space with a single class-independent, speaker-specific Gaussian dist...
متن کامل