From HMM's to segment models: a unified view of stochastic modeling for speech recognition

By Mari Ostendorf, Vassilios V. Digalakis, Owen Kimball. 1996

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

Pushing the envelope - aside [speech recognition] (2005, methodology): “One approach that has been proposed is to add feature dependencies or explicitly model the dynamics of frame-based features in various extensions of HMMs [5], [ 24 ].”
Bayesian reasoning and machine learning (2012, unknown): “The hidden semi-Markov model generalises the explicit duration model in that once a new duration ct is sampled, the model emits a distribution p(vt:t+ct−1|ht) defined on a segment of the next ct observations[216].”
Hidden semi-Markov models (2010, unknown): “Therefore, there exits This implies that Q ( j ) is a queue that only older ones can be discarded. newer items ( t , i ) , i ∈ S , are pushed into the queue after δ t ( j ) for all j ∈ S are determined by for 1 (cid”
A bidirectional target-filtering model of speech coarticulation and reduction: two-stage implementation for phonetic recognition (2006, unknown): “Earlier work directly represented speech dynamics in the observed acoustic domain [5], [16], [19], [24].”
Target-directed mixture dynamic models for spontaneous speech recognition (2004, result): “The mixture concept at the segmental level discussed here is similar to that in [11], [13], [14], and [16] (all summarized in [20]).”
A dynamic, feature-based approach to the interface between phonology and phonetics for speech modeling and recognition (1998, methodology): “[15,56]), speaking rate and the associated spectral variations would not be satisfactorily integrated into a common model structure.”
Speech trajectory discrimination using the minimum classification error learning (1998, methodology): “T HE formulation of the trended hidden Markov model (HMM), also called the parametric nonstationary-state HMM or parametric stochastic trajectory model, has been used in speech recognition applications for the past few years by a number of research groups [2], [4], [6]–[8], [13].”
Machine Learning Paradigms for Speech Recognition: An Overview (2013, methodology): “This class of beyond-HMM models have been variably called stochastic segment model [31], [32], trended or nonstationary-state HMM [33], [34], trajectory segmental model [32], [35], trajectory HMMs [36], [37], stochastic trajectory models [38], hidden dynamic models [39]–[45], buriedMarkov models [46”
Hierarchical Aligned Cluster Analysis for Temporal Clustering of Human Motion (2013, background): “In particular, there is a substantial amount of work in the field of data mining [27], [28], speech processing [29], [30], animal behavior analysis [12], [31], and signal processing [32], [33].”
MVA Processing of Speech Features (2007, methodology): “An account of the capacity/limitation of hidden Markov models is given in [2] and a general framework of segment modeling is given in [ 3 ].”
Deep segmental neural networks for speech recognition (2013, methodology): “It has been known for decades that modeling speech using the conventional HMM has several limitations as analyzed in [9, 10, 11].”