Fast clustering of Gaussians and the virtue of representing Gaussians in exponential model format
نویسندگان
چکیده
This paper aims to show the power and versatility of exponential models by focusing on exponential model representations of Gaussian Mixture Models (GMMs). In a recent series of papers by several authors, GMMs of varying structure and complexity have been considered. These GMMs can all be readily represented as exponential models and oftentimes favorably so. This paper shows how the exponential model representation can offer useful insight even in the case of diagonal and full covariance GMMs! The power of the exponential model is illustrated by proving the concavity of the log det function and also by discovering how to speed up diagonal covariance gaussian clustering.
منابع مشابه
An efficient approximate method for solution of the heat equation using Laguerre-Gaussians radial functions
In the present paper, a numerical method is considered for solving one-dimensional heat equation subject to both Neumann and Dirichlet initial boundary conditions. This method is a combination of collocation method and radial basis functions (RBFs). The operational matrix of derivative for Laguerre-Gaussians (LG) radial basis functions is used to reduce the problem to a set of algebraic equatio...
متن کاملA sparse Mixture of Gaussians model for blind separation of more source than sensors
We explore the use of mixtures of Gaussians for noisy and overcomplete ICA. In particular we introduce a simplifying approximation for the case of sparse sources that avoids the exponential growth of mixture components and results in a modified clustering algorithm. This can be extended to also deal with complex data which is important for audio applications. Experiments show that the algorithm...
متن کاملRapid speaker adaptation using MLLR and subspace regression classes
In recent years, various adaptation techniques for hidden Markov modeling with mixture Gaussians have been proposed, most notably MAP estimation and MLLR transformation. When the amount of adaptation data is limited, adaptation can be done by grouping similar Gaussians together to form regression classes and then transforming the Gaussians in groups. The grouping of Gaussians is often determine...
متن کاملDesign of compact acoustic models through clustering of tied-covariance Gaussians
We propose a new approach for designing compact acoustic models particularly suited to large systems that combine multiple model sets to represent distinct acoustic conditions or languages. We show that Gaussians based on mixtures of inverse covariances (MIC) with shared parameters can be clustered using an efficient Lloyd algorithm. As a result, more compact acoustic models can be built by clu...
متن کاملA Comparative Study of Gauss in Large Vocabulary Continuou
Gaussian mixture models are the most popular probability density used in automatic speech recognition. During decoding, often many Gaussians are evaluated. Only a small number of Gaussians contributes significantly to probability. Several promising methods to select relevant Gaussians are known. These methods have different properties in terms of required memory, overhead and quality of selecte...
متن کامل