A phonetic search approach to the 2006 NIST spoken term detection evaluation

Wallace, Roy; Vogt, Robbie; Sridharan, Subramanian

doi:10.21437/interspeech.2007-180

Cited by 44 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, we propose to use a multilingual vector representation to characterize the spoken utterances from English, Malay and Tamil. Representing the speech signal with phonetic features from other languages have shown to be useful in many tasks, including speech recognition [26], spoken term detection [27], speech summarization [28], and spoken language identification [29]. However, such multilingual representations have not been applied to speech evaluation.…”

Section: Multilingual Phonemic Embeddingmentioning

confidence: 99%

Multilingual Speech Evaluation: Case Studies on English, Malay and Tamil

Zhang¹,

Shi

Chen

2021

Preprint

View full text Add to dashboard Cite

Speech evaluation is an essential component in computerassisted language learning (CALL). While speech evaluation on English has been popular, automatic speech scoring on low resource languages remains challenging. Work in this area has focused on monolingual specific designs and handcrafted features stemming from resource-rich languages like English. Such approaches are often difficult to generalize to other languages, especially if we also want to consider suprasegmental qualities such as rhythm. In this work, we examine three different languages that possess distinct rhythm patterns: English (stresstimed), Malay (syllable-timed), and Tamil (mora-timed). We exploit robust feature representations inspired by music processing and vector representation learning. Empirical validations show consistent gains for all three languages when predicting pronunciation, rhythm and intonation performance.

show abstract

Section: Multilingual Phonemic Embeddingmentioning

confidence: 99%

Multilingual Speech Evaluation: Case Studies on English, Malay and Tamil

Zhang¹,

Shi

Chen

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Hence, alternative representations that can also capture out-of-vocabulary words are required -for example, each word in the recognized lattice can be expressed in terms of its constituent phones. During retrieval, out-of-vocabulary words can then be found by matching their sub-word (for example, phonetic) representation against such phone-based lattices (Wallace et al, 2007). Other sub-word decompositions have also been employed successfully (Szöke, 2010), but in most cases detection of out-of-vocabulary words remains substantially inferior to that of in-vocabulary words.…”

Section: Introductionmentioning

confidence: 99%

“…Other sub-word decompositions have also been employed successfully (Szöke, 2010), but in most cases detection of out-of-vocabulary words remains substantially inferior to that of in-vocabulary words. Although the query by example methods in Hazen et al (2009), Zhang and Glass (2009), Zhang et al (2012) achieve promising detection rates, retrieval is significantly more demanding computationally than with index-based approaches (Wallace et al, 2007). In Hazen et al (2009), it is therefore recommended that such methods be used as a rescoring mechanism for terms retrieved by a cruder (index-based) approach.…”

Section: Introductionmentioning

confidence: 99%

Language independent search in MediaEval's Spoken Web Search task

Metze

Anguera

Barnard

et al. 2014

Computer Speech & Language

View full text Add to dashboard Cite

“…Typically, a phoneme-based system is used to handle OOV terms, e.g., [3], [4], [9], [10]. In this approach, search terms are converted to pronunciations by letter-to-sound (LTS) models, and the pronunciations are searched for in a phoneme lattice generated by a speech recogniser.…”

Section: Introductionmentioning

confidence: 99%