Large vocabulary speech recognition using a hidden Markov model for acoustic/phonetic classification

Levinson, Stephen E.; Ljolje, Andrej; Miller, L.G.

doi:10.1109/icassp.1988.196630

Cited by 18 publications

(10 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…, i n , is given, then only the optimal segmentation of state durations needs to be determined. This is accomplished by simply rewriting (11) as [109,110] δ…”

Section: Map and Mle Estimate Of Statesmentioning

confidence: 99%

Hidden semi-Markov models

Shu

2010

Artificial Intelligence

604

416

View full text Add to dashboard Cite

As an extension to the popular hidden Markov model (HMM), a hidden semi-Markov model (HSMM) allows the underlying stochastic process to be a semi-Markov chain. Each state has variable duration and a number of observations being produced while in the state. This makes it suitable for use in a wider range of applications. Its forwardbackward algorithms can be used to estimate/update the model parameters, determine the predicted, filtered and smoothed probabilities, evaluate goodness of an observation sequence fitting to the model, and find the best state sequence of the underlying stochastic process. Since the HSMM was initially introduced in 1980 for machine recognition of speech, it has been applied in thirty scientific and engineering areas, such as speech recognition/synthesis, human activity recognition/prediction, handwriting recognition, functional MRI brain mapping, and network anomaly detection. There are about three hundred papers published in the literature. An overview of HSMMs is presented in this paper, including modelling, inference, estimation, implementation and applications. It first provides a unified description of various HSMMs and discusses the general issues behind them. The boundary conditions of HSMM are extended. Then the conventional models, including the explicit duration, variable transition, and residential time of HSMM, are discussed. Various duration distributions and observation models are presented. Finally, the paper draws an outline of the applications.

show abstract

“…, i n , is given, then only the optimal segmentation of state durations needs to be determined. This is accomplished by simply rewriting (11) as [109,110] δ…”

Section: Map and Mle Estimate Of Statesmentioning

confidence: 99%

Hidden semi-Markov models

Shu

2010

Artificial Intelligence

604

416

View full text Add to dashboard Cite

show abstract

“…Inputs: We illustrate the different ways of modeling play and talk activities in a BOSCC session in Figure 2. Under the assumption of a left-right HMM (as is the case of this work), we can improve the computational efficiency at (2) by constraining i to be the previous state of j [24,25]. We provide the algorithm for finding the optimum state sequence using EDHMM in Algorithm 1.…”

Section: Algorithm 1: Estimating Optimum State Sequence For Edhmmmentioning

confidence: 99%

A Knowledge Driven Structural Segmentation Approach for Play-Talk Classification During Autism Assessment

Kumar

Chebolu²,

Kim

et al. 2018

Interspeech 2018

View full text Add to dashboard Cite

Automatically segmenting conversational audio into semantically relevant components has both computational and analytical significance. In this paper, we segment play activities and conversational portions interspersed during clinicallyadministered interactions between a psychologist and a child with autism spectrum disorder (ASD). We show that various acoustic-prosodic and turn-taking features commonly used in the literature differ between these segments, and hence can possibly influence further inference tasks. We adopt a two-step approach for the segmentation problem by taking advantage of the structural relation between the two segments. First, we use a supervised machine learning algorithm to estimate class posteriors at frame-level. Next, we use an explicit-duration hidden Markov model (EDHMM) to align the states using the posteriors from the previous step. The durational distributions for both play and talk regions are learnt from training data and modeled using the EDHMM. Our results show that speech features can be used to successfully discriminate between play and talk activities, each providing important insights into the child's condition.

show abstract

“…The latter situation will allow the duration of the segment analyzed by one instantiation of the selected MLN to be greater than the duration of the signal analyzed in the former case. Recent publications confirm that better coding of speech and better recognition performance are obtained if more than a single time frame are considered for extracting speech parameters (Levinson et al 1988; Rabiner et al 1988), although they suggest the use of a time window with fixed duration. Specialized MLNs have been introduced.…”

Section: Generalization Of the Modelmentioning

confidence: 99%

Use of multilayer networks for the recognition of phonetic features and phonemes

Bengio

Mori

1989

Computational Intelligence

View full text Add to dashboard Cite

Artificial neural networks capable of doing hard learning offer a new way to undertake automatic speech recognition. The Boltzmann machine algorithm and the error back-propagation algorithm have been used to perform speaker normalization. Spectral segments are represented by spectral lines. Speaker-independent recognition of place of articulation for vowels is performed on lines. Performance of the networks is shown to depend on the coding of the input data. Samples were extracted from continuous speech of 38 speakers. The error rate obtained (4.2% error on test set of 72 samples with the Boltzmann machine and 6.9% error with error back-propagation) is better than that of previous experiments, using the same data, with continuous Hidden Markov Models (7.3% error on test set and 3% error on training set). These experiments are part of an attempt to construct a data-driven speech recognition system with multiple neural networks specialized to different tasks. Results are also reported on the recognition performance of other trained networks, such as one trained on the E-set consonants.Key words: automatic speech recognition, neural networks, Boltzmann machine algorithm, error back-propagation algorithm, spectral lines, speaker-independent speech recognition.Les riseaux neuraux artificiels en mesure d'effectuer des apprentissages difficiles offrent une nouvelle facon d'effectuer la reconnaissance automatique de la parole. L'algorithme de la machine de Boltzmann et I'algorithme de propagation arrike d'erreurs ont it6 utilisb pour effectuer la normalisation du locuteur. Les segments spectraux sont representis par des lignes spectrales. La reconnaissance independante de I'endroit de I'articulation des voyelles par le locuteur est effectuke sur des lignes. La performance des rCseaux s'est rivelee dipendante de I'encodage des donnies d'entrte. Des echantillons ont ete tires du discours continu de 38 locuteurs. Le taux d'erreur obtenu (4,2010 d'erreur dans les 72 echantillons avec la machine de Boltzmann, 6,9070 d'erreur avec I'algorithme de propagation arrikre d'erreurs) est inferieur a celui des experiences precedentes, effectuees a I'aide des mCmes donnies et avec les modkles continus de Markov (7,3% d'erreur avec la serie d'essai et 3% d'erreur avec la serie de formation). Ces experiences font partie d'un programme en vue d'tlaborer un systkme de reconnaissance de la parole guidCe par les donnees qui serait dote de reseaux neuraux multiples specialisis dans diverses tlches. Cet article prksente Cgalement des donnies sur la performance d'autres reseaux formts.Mots clPs : reconnaisance automatique de la parole, reseaux neuraux, algorithme de la machine de Boltzmann, algorithme de propagation arriere d'erreurs, lignes spectrales, reconnaissance de la parole indkpendante du locuteur.[Traduit par la revue] Cornput. Intell. 5, 134-141 (1989)

show abstract

Large vocabulary speech recognition using a hidden Markov model for acoustic/phonetic classification

Cited by 18 publications

References 4 publications

Hidden semi-Markov models

Hidden semi-Markov models

A Knowledge Driven Structural Segmentation Approach for Play-Talk Classification During Autism Assessment

Use of multilayer networks for the recognition of phonetic features and phonemes

Contact Info

Product

Resources

About