Syllable-a promising recognition unit for LVCSR

Ganapathiraju, Aravind; Goel, Vaibhava; Picone, J.; Corrada, A.; Doddington, George R.; Kirchhoff, Katrin; Ordowski, Mark; Wheatley, B.

doi:10.1109/asru.1997.659007

Cited by 16 publications

(6 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Four passes of Baum-Welch reestimation were used to reestimate the triphone model parameters. The number of Gaussians was, however, reduced by tying states [20]. Finally, these models were increased to eight Gaussians per state using a standard divide-by-two clustering algorithm.…”

Section: B Baseline Triphone Systemmentioning

confidence: 99%

Syllable-based large vocabulary continuous speech recognition

Ganapathiraju¹,

Hamaker

Picone

et al. 2001

IEEE Trans. Speech Audio Process.

Self Cite

104

View full text Add to dashboard Cite

Most large vocabulary continuous speech recognition (LVCSR) systems in the past decade have used a context-dependent (CD) phone as the fundamental acoustic unit. In this paper, we present one of the first robust LVCSR systems that uses a syllable-level acoustic unit for LVCSR on telephone-bandwidth speech. This effort is motivated by the inherent limitations in phone-based approaches-namely the lack of an easy and efficient way for modeling long-term temporal dependencies. A syllable unit spans a longer time frame, typically three phones, thereby offering a more parsimonious framework for modeling pronunciation variation in spontaneous speech. We present encouraging results which show that a syllable-based system exceeds the performance of a comparable triphone system both in terms of word error rate (WER) and complexity. The WER of the best syllable system reported here is 49.1% on a standard SWITCHBOARD evaluation, a small improvement over the triphone system. We also report results on a much smaller recognition task, OGI Alphadigits, which was used to validate some of the benefits syllables offer over triphones. The syllable-based system exceeds the performance of the triphone system by nearly 20%, an impressive accomplishment since the alphadigits application consists mostly of phone-level minimal pair distinctions.

show abstract

Section: B Baseline Triphone Systemmentioning

confidence: 99%

Syllable-based large vocabulary continuous speech recognition

Ganapathiraju¹,

Hamaker

Picone

et al. 2001

IEEE Trans. Speech Audio Process.

Self Cite

104

View full text Add to dashboard Cite

show abstract

“…Syllables as a basic recognition unit were again analyzed by a team of researchers in Ganapathiraju et al [1997]. This work created a set of models for the 200 most common monosyllabic words, and a second set of syllable models for the syllables in the remaining monosyllabic words and multiple syllable words.…”

Section: Syllables As a Recognition Unitmentioning

confidence: 99%

Graphical models for integrating syllabic information

Bartels

Bilmes

2010

Computer Speech & Language

View full text Add to dashboard Cite

“…A plausible solution to these constraints was to select the male speaker subset [6] from the WS'97 DevTest, and to reserve 10 utterances from each test speaker for adaptation. This resulted in 1241 utterances consisting o f 2 3 s p e a k e r s , 2 4 c o n v e r s a t i o n s i d e s , a n d approximately 50 minutes of speech.…”

Section: N-best Rescoring Experimentsmentioning

confidence: 99%

“…This resulted in 1241 utterances consisting o f 2 3 s p e a k e r s , 2 4 c o n v e r s a t i o n s i d e s , a n d approximately 50 minutes of speech. We used a baseline context-dependent phone HMM system [6] to generate N-best lists and time alignments for the reference transcription and the 100-best hypotheses. The HDM systems rescored these hypotheses, and the resulting sentence hypotheses were scored using standard NIST scoring software and presented in terms of word error rate (WER).…”

Section: N-best Rescoring Experimentsmentioning

confidence: 99%

Initial evaluation of hidden dynamic models on conversational speech

Pike

Regan

Kamm

et al. 1999

1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258)

Self Cite

View full text Add to dashboard Cite

Conversational speech recognition is a challenging problem primarily because speakers rarely fully articulate sounds. A successful speech recognition approach must infer intended spectral targets from the speech data, or develop a method of dealing with large variances in the data. Hidden Dynamic Models (HDMs) attempt to automatically learn such targets in a hidden feature space using models that integrate linguistic information with constrained temporal trajectory models. HDMs are a radical departure from conventional hidden Markov models (HMMs), which simply account for variation in the observed data. In this paper, we present an initial evaluation of such models on a conversational speech recognition task involving a subset of the SWITCHBOARD corpus. We show that in an N-Best rescoring paradigm, HDMs are capable of delivering performance competitive with Hh4Ms. produce more consistent acoustic scoring for conversational speech, because sounds are rarely fully articulated in such data. Tremendous amounts of variation are observed in the speech data because of the manner in which the realization of a sound was truncated is highly context-dependent. It is the goal of this work to produce acoustic scores that reflect measurements in the hidden (or target) space, rather than directly in the feature space as is currently done in context-dependent phonetic modeling.The work presented here was the culmination of an intense effort at the 1998 NSF Workshop on Language Engineering held at the Center for Language and Speech Processing at Johns Hopkins University. One goal of this work, which is the primary focus of this paper, was to evaluate the HDM approach on a credible conversational speech recognition task involving the SWITCHBOARD (SWB) Corpus [3]. HIDDEN DYNAMIC MODELS 1. INTRODUCTIONHidden dynamic models [ 1,2] (HDMs) attempt to produce acoustic likelihoods of phone-level sound units that reflect intended spectral configurations rather than likelihoods based on the actual realization of the sound in the speech data. This is a radical departure from current statistical modeling approaches that attempt to account for variation in the data by accumulating large numbers of Gaussian mixture components. It is conjectured that this approach will 1.

show abstract

Syllable-a promising recognition unit for LVCSR

Cited by 16 publications

References 3 publications

Syllable-based large vocabulary continuous speech recognition

Syllable-based large vocabulary continuous speech recognition

Graphical models for integrating syllabic information

Initial evaluation of hidden dynamic models on conversational speech

Contact Info

Product

Resources

About