Maximum likelihood estimation for multivariate mixture observations of markov chains

By Biing-Hwang Fred Juang, S. Levinson, M. Sondhi. 1986

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

An ICA Mixture Hidden Markov Model for Video Content Analysis (2008, methodology): “To apply maximum likelihood estimation, we define the auxiliary function (the -function) with similar form as [9], [ 10 ] :”
A Unified Framework for Model-based Clustering (2003, methodology): “The singularity problem is often dealt with in one of the following three ways: (a) restarting the whole clustering algorithm with a different initialization (Juang et al., 1986); (b) using maximum a posteriori estimation with an appropriate prior (Gauvain and Lee, 1994); (c) using constrained ML es”
Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups (2012, unknown): “…always had the same size, but even with this constraint it was impossible to train all possible combinations of number of hidden layers [1] [2] [3] [4] [5] [6] [7] [8], number of units per layer [512] [1] [024] [2] [048] [3] [072], and number of frames of acoustic data in the input layer [7]…”
Deep Neural Networks for Acoustic Modeling in Speech Recognition (2012, methodology): “With the EM algorithm, it became poss ible to develop speech recognition systems for real world tasks using the richness of Gaussian mixture mode ls (GMM) [3] to represent the relationship between HMM states and the acoustic input.”
Automatic Speech Recognition - A Brief History of the Technology Development (2004, methodology): “In the ly 1980’s at Bell the theory of HMM was extended to mixture [ 30-31 ] wh ich have since proven vitally important in”
Hermitian Polynomial for Speaker Adaptation of Connectionist Speech Recognition Systems (2013, methodology): “Although there is a concern of the deep neural network runtime efficiency compared to GMMs since a deep neural network has much more parameters to evaluate, researchers at Microsoft have recently conducted studies on how to reduce this runtime cost by restructuring the deep neural network with less ”