Stochastic Segment Model Research Articles

Promoter prediction on a genomic scale--the Adh experience.

We describe our statistical system for promoter recognition in genomic DNA with which we took part in the Genome Annotation Assessment Project (GASP1). We applied two versions of the system: the first uses a region-based approach toward transcription start site identification, namely, interpolated Markov chains; the second was a hybrid approach combining regions and signals within a stochastic segment model. We compare the results of both versions with each other and examine how well the application on a genomic scale compares with the results we previously obtained on smaller data sets.

Speech recognition using stochastic phonemic segment model based on phoneme segmentation

This paper discusses speech recognition based on a new statistical phoneme segment model which is trained by phoneme parameters derived from automatically extracted phoneme segments. The proposed system operates as follows. In preprocessing before recognition, the phoneme boundaries are detected by segmentation. The phonemes are discriminated using a stochastic phoneme segment model, and a phoneme segment lattice with scores is constructed. Next the speech recognition is performed by matching of symbol sequences to dictionary items. The segmentation system that is employed can infer phoneme boundaries with high accuracy. This helps to eliminate unnecessary parameters, leaving the feature parameters which are effective in separating phonemes. In other words, the phoneme recognition problem in continuous speech can be reduced to a discrimination problem and thus a speaker-independent model can be constructed from a relatively small number of training data. The stochastic phoneme segment model is trained with training samples extracted from a phoneme-balanced word set of 4920 words uttered by 10 speakers. In a recognition experiment with 6709 words uttered by 63 nontraining speakers, a recognition rate of 92.6% was obtained as the average for all speakers, using a word dictionary of 212 words. © 2000 Scripta Technica, Syst Comp Jpn, 31(10): 89–98, 2000

Stochastic segment models of eukaryotic promoter regions.

We present a new statistical approach for eukaryotic polymerase II promoter recognition. We apply stochastic segment models in which each state represents a functional part of the promoter. The segments are trained in an unsupervised way. We compare segment models with three and five states with our previous system which modeled the promoters as a whole, i.e. as a single state. Results on the classification of a representative collection of human and D. melanogaster promoter and non-promoter sequences show great improvements. The practical importance is demonstrated on the mining of large contiguous sequences.

Recent advances in speech recognition for spontaneous speech translation

Speech recognition research for speech translation at ATR will be reviewed. For spontaneous speech recognition, new methods for parameter extraction will be shown that yield acoustic measures to allow discriminative acoustic model training. A new HMM state-sharing algorithm (ML-SSS), fine-acoustic modeling using stochastic segment models, and recurrent neural networks will also be described. The MAP-VFS speaker adaptation techniques will be presented in the context of speaker clustered acoustic models. As for language modeling, variable order N-gram models and task adaptation of language modeling will be described. In addition, further research efforts on pronunciation modeling and training using error characteristics will also be reviewed. Finally, an outline of the continuous speech recognition system (ATR-SPREC) for spontaneous speech will be presented, along with a review of recent recognition experiment results.

Use of explicit duration models in speech recognition

Although researchers have observed that automatic speech recognition errors are often associated with improbable duration patterns, current recognition systems take little advantage of known durational cues. In addition to using simplistic models of duration, recognizers put very little weight on durational probabilities (because of the high dimensionality of the feature vectors typically used). This work explores the benefits that can be gained in speaker‐independent continuous speech recognition from more accurate duration modeling and automatically estimating a duration score weight. In the context of the stochastic segment model, segment durations are modeled using gamma distributions conditioned on phonetic, lexical, and phrasal context, as well as speaking rate. Robust parameter estimation, a potential problem with the large number of different conditioning factors, is achieved using automatic clustering techniques with a maximum likelihood criterion. A by‐product of the clustering procedure is a better understanding of the relative importance of the conditioning factors. Speech recognition experiments on the resource management task show that simple context‐independent gamma distributions give similar results to nonparametric relative frequency models, and that error can be reduced by more than 10% by conditioning models on such factors as prepausal position, speaking rate, and lexical stress.

Context modeling with the stochastic segment model

An approach for context modeling in continuous speech recognition for models based on multivariate Gaussian distributions, specifically, the stochastic segment model, is described. Robust context models are obtained by typing distribution parameters across different classes of context. Experimental results in phoneme and word recognition are comparable to those achieved with discrete hidden Markov models using mixture distributions. >

Fast algorithms for phone classification and recognition using segment-based models

Methods for reducing the computation requirements of joint segmentation and recognition of phones using the stochastic segment model are presented. The approach uses a fast segment classification method that reduces computation by a factor of two to four, depending on the confidence of choosing the most probable model. A split-and-merge segmentation algorithm is proposed as an alternative to the typical dynamic programming solution of the segmentation and recognition problem, with computation savings increasing proportionally with model complexity. Although the current recognizer uses context-independent phone models, the results reported for the TIMIT database for speaker-independent joint segmentation and recognition are comparable to those of systems that use context information. >

A stochastic segment model for phoneme-based continuous speech recognition

The authors introduce a novel approach to modeling variable-duration phonemes, called the stochastic segment model. A phoneme X is observed as a variable-length sequence of frames, where each frame is represented by a parameter vector and the length of the sequence is random. The stochastic segment model consists of (1) a time warping of the variable-length segment X into a fixed-length segment Y called a resampled segment and (2) a joint density function of the parameters of X which in this study is a Gaussian density. The segment model represents spectra/temporal structure over the entire phoneme. The model also allows the incorporation in Y of acoustic-phonetic features derived from X, in addition to the usual spectral features that have been used in hidden Markov modeling and dynamic time warping approaches to speech recognition. The authors describe the stochastic segment model, the recognition algorithm, and an iterative training algorithm for estimating segment models from continuous speech. They present several results using segment models in two speaker-dependent recognition tasks and compare the performance of the stochastic segment model to the performance of the hidden Markov models. >

On the use of some robust modeling techniques for speech recognition

Recently, hidden Markov models (HMM) have been applied successfully to both isolated and connected word recognition. However, when the same formulation is adopted for recognition of more confusable vocabularies, like English alphabets, the recognition performance is often less satisfactory. One main reason is that robustness issues, such as model validity, recognition parameter selection, model parameter initialization, model parameter estimation, training sample size, and durational information incorporation, can no longer be ignored. In this paper, a stochastic segment model (SSM) is proposed, which is a simplified HMM, for speech recognition. Three specific robustness issues are then discussed, namely the choice of observation densities, the initialization of model parameters, and the incorporation of duration information. In a step-by-step attempt to address those issues, it was found that the same SSM formulation can still be adopted if acoustic and phonetic knowledge about the vocabulary is taken into account in the model parameter estimation and recognition phases. Testing on the 39-word English alpha-digit vocabulary indicates that the recognition performance, based on conventional HMM techniques, can be signficantly improved if model parameters are adequately initialized and durational information is properly incorporated.