Decision trees for phonological rules in continuous speech

Bahl, Lalit R.; deSouza, P.V.; Gopalakrishnan, P.S.; Nahamoo, D.; Picheny, Michael

doi:10.1109/icassp.1991.150308

Cited by 95 publications

(47 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In automatic speech recognition (ASR), rare triphones are tied on the model [1] or the state [2] level, and such context modeling based on either data-driven or decision tree clustering significantly improves the recognition performance. It was already shown that the state tying system consistently out-performs the model clustered system.…”

Section: Introductionmentioning

confidence: 99%

Rule-Based Triphone Mapping for Acoustic Modeling in Automatic Speech Recognition

Darjaa

Cerňak

Beňuš

et al. 2011

Text, Speech and Dialogue

View full text Add to dashboard Cite

Abstract. This paper presents rule-based triphone mapping for acoustic models training in automatic speech recognition. We test if the incorporation of expanded knowledge at the level of parameter tying in acoustic modeling improves the performance of automatic speech recognition in Slovak. We propose a novel technique of knowledge-based triphone tying, which allows the synthesis of unseen triphones. The proposed technique is compared with decision tree-based state tying, and it is shown that for bigger acoustic models, at a size of 3000 states and more, a triphone mapped HMM system achieves better performance than a tree-based state tying system on a large vocabulary continuous speech transription task. Experiments, performed using 350 hours of a Slovak audio database of mixed read and spontaneous speech, are presented. Relative decrease of word error rate was 4.23% for models with 7500 states, and 4.13% at 11500 states.

show abstract

Section: Introductionmentioning

confidence: 99%

Rule-Based Triphone Mapping for Acoustic Modeling in Automatic Speech Recognition

Darjaa

Cerňak

Beňuš

et al. 2011

Text, Speech and Dialogue

View full text Add to dashboard Cite

show abstract

“…Starting from the classical scheme [1], some attempts have been made in order to improve the accuracy and the discriminative power of the models. An alternative methodology, the fast and efficient Growing and Pruning algorithm [4], has also been applied to build the decision trees.…”

Section: Introductionmentioning

confidence: 99%

“…Decision Trees (DT) are one of the most common approaches to the problem of selecting a suitable set of context dependent sublexical units (DT-CD-SLUs) for speech recognition [1][2] [3].…”

Section: Introductionmentioning

confidence: 99%

Decision Tree-Based Context Dependent Sublexical Units for Continuous Speech Recognition of Basque

Ipiña¹,

Grana²,

Ezeiza

et al. 2003

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper presents a new methodology, based on the classical decision trees, to get a suitable set of context dependent sublexical units for Basque Continuous Speech Recognition (CSR). The original method proposed by Bahl [1] was applied as the benchmark. Then two new features were added: a data massaging to emphasise the data and a fast and efficient Growing and Pruning algorithm for DT construction. In addition, the use of the new context dependent units to build word models was addressed. The benchmark Bahl approach gave recognition rates clearly outperforming those of context independent phone-like units. Finally the new methodology improves over the benchmark DT approach.

show abstract

“…The probability of cluster membership P(Kjtr u ) can be ex- (5) This last approximation rests upon the assumption that left and right contextual inuences are independent of each other which is of course not exactly right. However, this simplication is needed to achieve generalization capabilities and is somewhat analog to the use of questions concerning the left (right) only contexts for constructing decision trees [4]. The probabilities in (5) ; (6) where #Tr o is the total number of observed triphone states and #K the number of triphone states in cluster K.…”

Section: Decoding With Unseen Triphonesmentioning

confidence: 99%

“…So far, this generalization problem has been mainly solved using decision-trees [4] [1] [5] that proceed \top-down" by going from less-to more specic contexts through successive (binary) splitting steps. In contrast, bottom-up techniques start from the most detailed level and aim at building robust entities by means of a merging process typically based on agglomerative clustering.…”

Section: Introductionmentioning

confidence: 99%

A bottom-up approach for handling unseen triphones in large vocabulary continuous speech recognition

Aubert

Beyerlein²,

Ullrich³

Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96

View full text Add to dashboard Cite

This paper presents an extension of bottom-up state-tying towards improved handling of unseen triphones. As opposed to the usual backing-o to diphones and monophones, the current method aims at nding a triphone model that has proven to exhibit some similarity with the unseen triphone. It is based on a probabilistic mapping of unseen contexts to clusters of triphone-states observed in the training data.This algorithm has been applied to dictation tasks for three languages with vocabulary sizes ranging from 20k to 64k. The results compare favorably with those obtained using standard back-o rules. This technique also oers an alternative to top-down decision-tree procedures which are frequently used especially for their generalization capabilities.

show abstract

Decision trees for phonological rules in continuous speech

Cited by 95 publications

References 13 publications

Rule-Based Triphone Mapping for Acoustic Modeling in Automatic Speech Recognition

Rule-Based Triphone Mapping for Acoustic Modeling in Automatic Speech Recognition

Decision Tree-Based Context Dependent Sublexical Units for Continuous Speech Recognition of Basque

A bottom-up approach for handling unseen triphones in large vocabulary continuous speech recognition

Contact Info

Product

Resources

About