Bilingual Experiments on Automatic Recovery of Capitalization and Punctuation of Automatic Speech Transcripts

Batista, Fernando; Moniz, Helena; Trancoso, Isabel; Mamede, Nuno J.

doi:10.1109/tasl.2011.2159594

Cited by 52 publications

(54 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Other features involve pitch (f0), energy, jitter and shimmer, including pitch and energy average, median, standard deviation, dynamics, range, and slopes, both within and between IPUs [1]. Pitch related features were calculated based on semitones rather than frequency.…”

Section: Recent Progress In Parameterizations For Paralinguisticsmentioning

confidence: 99%

Acoustic-Prosodic Automatic Personality Trait Assessment for Adults and Children

Solera-Ureña

Moniz

Batista

et al. 2016

Advances in Speech and Language Technologies for Iberian Languages

Self Cite

View full text Add to dashboard Cite

Abstract. This paper investigates the use of heterogeneous speech corpora for automatic assessment of personality traits in terms of the BigFive OCEAN dimensions. The motivation for this work is twofold: the need to develop methods to overcome the lack of children's speech corpora, particularly severe when targeting personality traits, and the interest on cross-age comparisons of acoustic-prosodic features to build robust paralinguistic detectors. For this purpose, we devise an experimental setup with age mismatch utilizing the Interspeech 2012 Personality Subchallenge, containing adult speech, as training data. As test data, we use a corpus of children's European Portuguese speech. We investigate various features sets such as the Sub-challenge baseline features, the recently introduced eGeMAPS features and our own knowledge-based features. The preliminary results bring insights into cross-age and -language detection of personality traits in spontaneous speech, pointing out to a stable set of acoustic-prosodic features for Extraversion and Agreeableness in both adult and child speech.

show abstract

Section: Recent Progress In Parameterizations For Paralinguisticsmentioning

confidence: 99%

Acoustic-Prosodic Automatic Personality Trait Assessment for Adults and Children

Solera-Ureña

Moniz

Batista

et al. 2016

Advances in Speech and Language Technologies for Iberian Languages

Self Cite

View full text Add to dashboard Cite

show abstract

“…Prosodic and textual cues are also combined in [22] and implemented in a decision tree classifier with the goal to detect sentence boundaries. A combination of lexical-, prosodic-, and speaker-based features is also found in [4] for the detection of full stops, commas, and question marks in a bilingual English-Portuguese broadcast news data, while [19] focuses on Czech broadcast news speech to detect commas and sentence boundaries by using a prosodic model in a decision trees and a multi-layer perceptron and N-gram models for language modeling. Similar works deal with the punctuation generation problem by using statistical models of prosodic features [9], the combination of both textual and prosodic features based on adaptive boosting [18], and a cross-linguistic study of prosodic features through two different approaches for feature selection: a forward search wrapper and feature filtering [14].…”

Section: Related Workmentioning

confidence: 99%

Attentional Parallel RNNs for Generating Punctuation in Transcribed Speech

Öktem

Farrús

Wanner

2017

Statistical Language and Speech Processing

View full text Add to dashboard Cite

Abstract. Until very recently, the generation of punctuation marks for automatic speech recognition (ASR) output has been mostly done by looking at the syntactic structure of the recognized utterances. Prosodic cues such as breaks, speech rate, pitch intonation that influence placing of punctuation marks on speech transcripts have been seldom used. We propose a method that uses recurrent neural networks, taking prosodic and lexical information into account in order to predict punctuation marks for raw ASR output. Our experiments show that an attention mechanism over parallel sequences of prosodic cues aligned with transcribed speech improves accuracy of punctuation generation.

show abstract

“…O enriquecimento de transcrições automáticas depende de: i) transcrições manuais, ii) transcrições produzidas pelo reconhecedor automático e iii) análise do sinal acústico (para mais informações sobre este processo, veja-se Batista et al 2012aBatista et al e 2012bMoniz 2013;Moniz et al 2014b;Cabarrão et al 2015). A anotação manual é complexa sobretudo no caso da anotação de corpora de fala espontânea, uma vez que implica pontuar fala, por um lado, e identificar disfluências e marcadores discursivos, por outro -tarefas que, como se sabe, afetam a concordância inter-anotadores em distintos corpora do português europeu (cf.…”

Section: Prosódia E Processamento Automáticounclassified

5. Prosódia, variação e processamento automático

Mata¹,

Moniz²

2016

Manual De Linguística Portuguesa

Self Cite

View full text Add to dashboard Cite

Prosódia, variação e processamento automáticoNeste capítulo apresentamos um olhar panorâmico sobre a variação prosódica e sobre a sua interface com a área do processamento automático de fala. Tendo por base essencialmente a investigação que tem sido desenvolvida no português europeu sobre corpora de fala espontânea e preparada, em contextos de exposição e de interação, nomeadamente na variedade padrão falada em Lisboa, analisamos a variação da entoação em contextos declarativos e interrogativos, e abordamos as funções pragmático-discursivas que podem associar-se também a outros parâmetros prosódicos. Partindo de estudos comparativos inter-estilos (com maior/menor grau de espontaneidade e de planeamento, e natureza mais interativa/expositiva) e inter-falantes (espaço geográfico, género, grupo etário/estatuto), destacamos o papel da variação estilística e sociolinguística da prosódia no português europeu. Mostramos também o papel da variação no processamento automático de proeminência prosódica, pontuação, disfluências e emoções.Palavras-chave: prosódia, entoação, variação estilística, variação social, processamento automático de fala

show abstract

Bilingual Experiments on Automatic Recovery of Capitalization and Punctuation of Automatic Speech Transcripts

Cited by 52 publications

References 26 publications

Acoustic-Prosodic Automatic Personality Trait Assessment for Adults and Children

Acoustic-Prosodic Automatic Personality Trait Assessment for Adults and Children

Attentional Parallel RNNs for Generating Punctuation in Transcribed Speech

5. Prosódia, variação e processamento automático

Contact Info

Product

Resources

About