A Hidden Trajectory Model with Bi-directional Target-Filtering: Cascaded vs. Integrated Implementation for Phonetic Recognition

Deng, Li; Li, Xiang; Yu, Dong; Acero, Alex

doi:10.1109/icassp.2005.1415119

Cited by 8 publications

(8 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, in our future work we expect greater advantages of the HTM for these difficult databases than the already demonstrated superiority for TIMIT as demonstrated in this paper. In our earlier work on TIMIT [4,5], we found that the oracle error rate for the N-best lists (N as large as 2000) produced by the HMM is as high as 18%. This accounts for the large difference between the N-best rescoring accuracies with and without including the reference hypotheses in the N-best lists.…”

Section: Discussionmentioning

confidence: 97%

“…The observation takes the form of LPC cepstra or LPCC (and their frequencywarped version) in this paper. An analytical form of the nonlinear prediction function F[z(k)] was presented in [4] and summarized as:…”

Section: Generating Acoustic Observationsmentioning

confidence: 99%

“…Further, the current paper presents more comprehensive evaluation results on the HTM than in the earlier work. The work in [4,5] showed the effectiveness of the HTM for phonetic recognition mainly when the reference transcripts are included in the N-best lists where N = 1000. This success was more recently extended to lattice search with an equivalent of much larger lists with N in the order of billions to trillions [10].…”

Section: Introductionmentioning

confidence: 99%

“…Several aspects of the HTM and its preliminary evaluation in phonetic recognition have been presented in our earlier publications [4,5,10], including the training algorithms for a partial set of model parameters. In this paper, we provide the training algorithms for the entire set of model parameters, and in particular, we present details of gradient descent techniques for learning the covariance matrices for both the residuals and targets.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A long-contextual-span model of resonance dynamics for speech recognition: parameter learning and recognizer evaluation

Deng

et al. 2005

IEEE Workshop on Automatic Speech Recognition and Understanding, 2005.

Self Cite

View full text Add to dashboard Cite

We present a structured speech model that is equipped with the capability of jointly representing incomplete articulation and long-span co-articulation in natural human speech. Central to this model is compact statistical parameterization of the highly regular dynamic patterns (exhibited in the hidden vocal-tract-resonance domain) that are driven by the stochastic segmental targets. We provide a rigorous mathematical description of this model, and present novel algorithms for learning the full set of model parameters using the cepstral data of speech. In particular, the gradient ascend techniques for learning variance parameters (for both resonance targets and cepstral prediction residuals) are described in detail. Phonetic recognition experiments are carried out using two paradigms of N-best rescoring and lattice search. Both sets of results demonstrate higher recognition accuracy achieved by the new model compared with the best HMM system. The higher accuracy is consistently observed, with and without combining HMM scores, and with and without including the references in the N-best lists and lattices. Further, the new model with rich parameter-free structure uses only the context-independent, singlemodal Gaussian parameters, which are fewer than one percent of the parameters in the context-dependent HMM system with mixture distributions.

show abstract

Section: Discussionmentioning

confidence: 97%

Section: Generating Acoustic Observationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A long-contextual-span model of resonance dynamics for speech recognition: parameter learning and recognizer evaluation

Deng

et al. 2005

IEEE Workshop on Automatic Speech Recognition and Understanding, 2005.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This motivates many research works on dynamic statistics modeling by introducing more complicated probabilistic models to relax such limitation. Some typical works are switching linear dynamical system (SLDS) [3], stochastic segment model (SSM) [4], [5], factor analyzed HMM (FAHMM) [6], polynomial segment model [7], hidden trajectory model (HTM) [8], buried Markov model (BMM) [9] and trajectory HMM [10], [11]. However, these techniques have shown little success for large vocabulary continuous speech recognition (LVCSR) [12] using maximum likelihood training.…”

Section: Introductionmentioning

confidence: 99%

Temporally Varying Weight Regression: A Semi-Parametric Trajectory Model for Automatic Speech Recognition

Liu

Sim

2014

IEEE/ACM Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Standard Hidden Markov Model (HMM) assumes that successive observations are independent to one another given the state sequence. This leads to a poor trajectory model for speech. Many explicit trajectory modeling techniques have been studied in the past to improve trajectory modeling for HMM. However, these techniques do not yield promising improvements over conventional HMM systems where differential parameters and Gaussian Mixture Model have been used implicitly to circumvent the poor trajectory modeling issue of HMM. Recently, semi-parametric trajectory modeling techniques based on temporally varying model parameters such as fMPE and pMPE have been shown to yield promising improvements over state-of-the-art systems on large vocabulary continuous speech recognition tasks. These techniques use high dimensional posterior features derived from a long span of acoustic features to model temporally varying attributes of the speech signal. Bases corresponding to these posterior features are then discriminatively estimated to yield temporally varying mean (fMPE) and precision matrix (pMPE) parameters. Motivated by the success of fMPE and pMPE, Temporally Varying Weight Regression (TVWR) was recently proposed to model HMM trajectory implicitly using time-varying Gaussian weights. In this paper, a complete formulation of TVWR is given based on a probabilistic modeling framework. Parameter estimation formulae using both maximum likelihood (ML) and minimum phone error (MPE) criteria are derived. Experimental results based on the Wall Street Journal (CSR-WSJ WSJ ) and Aurora 4 corpora show that consistent promising improvements over the standard HMM systems can be obtained in both the 20 k open vocabulary recognition task (NIST Nov'92 WSJ0) and 5 k closed vocabulary noisy speech recognition for both ML and MPE criteria.

show abstract

Discriminative speaker adaptation using articulatory features

Metze¹

2007

Speech Communication

View full text Add to dashboard Cite

This paper presents an automatic speech recognition system using acoustic models based on both sub-phonetic units and broad, phonological features such as Voiced and Round as output densities in a hidden Markov model framework. The aim of this work is to improve speech recognition performance particularly on conversational speech by using units other than phones as a basis for discrimination between words. We explore the idea that phones are more of a shorthand notation for a bundle of phonological features, which can also be used directly to distinguish competing word hypotheses. Acoustic models for different features are integrated with phone models using a multi-stream approach and log-linear interpolation. This paper presents a new lattice based discriminative training algorithm using the maximum mutual information criterion to train stream weights. This algorithm allows us to automatically learn stream weights from training or adaptation data and can also be applied to other tasks. Decoding experiments conducted in comparison to a non-feature baseline system on the large vocabulary English Spontaneous Scheduling Task show reductions in word error rate of about 20% for discriminative model adaptation based on articulatory features, slightly outperforming other adaptation algorithms.

show abstract

A Hidden Trajectory Model with Bi-directional Target-Filtering: Cascaded vs. Integrated Implementation for Phonetic Recognition

Cited by 8 publications

References 5 publications

A long-contextual-span model of resonance dynamics for speech recognition: parameter learning and recognizer evaluation

A long-contextual-span model of resonance dynamics for speech recognition: parameter learning and recognizer evaluation

Temporally Varying Weight Regression: A Semi-Parametric Trajectory Model for Automatic Speech Recognition

Discriminative speaker adaptation using articulatory features

Contact Info

Product

Resources

About