Modelling the Interplay of Multiple Cues in Prosodic Focus Marking

Arnhold, Anja; Kyröläinen, Aki-Juhani

doi:10.5334/labphon.78

Cited by 13 publications

(11 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Al Tamimi and Khattab (2018) discovered the most predictive markers for differentiating between distinct acoustic stops using random forests and linear mixed models. Arnhold and Kyrolainen (2017) evaluated the concentration scoring by random woodlands and the Universalis added hybrid algorithm, emphasizing the variable significance. They discovered that both algorithms performed admirably in their evaluation.…”

Section: Teaching and Learning Phoneticsmentioning

confidence: 99%

The Uses of Machine Learning (ML) in Teaching and Learning English Language: A Methodical Review.

AlHarbi

2022

المجلة التربوية لکلية التربية بسوهاج

View full text Add to dashboard Cite

Regular teaching methods for ESL/EFL may rely on assumptions that do not lead to the desired results in student learning. Consequently, the current study sought to perform a systematic review of the research that revealed numerous applications of machine learning (ML) in English language instruction (Applied Linguistics) and the development of interpretable machine learning or associated practical software. The current research employed a mixed-methods approach (approach) that included qualitative and quantitative analysis to survey the previous studies on machine learning in EFL/ESL instruction. Results showed that machine learning (ML) may teach English as a foreign language or as a second language and that it has a bright future in this sector. Interpretable machine learning (ML) was found to help in teaching many EFL/ESL topics such as dictation, phonetics instruction, language blogging, and all of which contribute to the facilitation of English language learning and teaching processes. Finally, the study concluded with recommendations for researchers in EFL or ESL on how to incorporate machine learning in English classrooms using functional programming languages, tools, and platforms. More studies were also advised to determine the most effective methods and procedures for incorporating machine learning (ML) into English as ESL/EFL in the classroom.

show abstract

Section: Teaching and Learning Phoneticsmentioning

confidence: 99%

The Uses of Machine Learning (ML) in Teaching and Learning English Language: A Methodical Review.

AlHarbi

2022

المجلة التربوية لکلية التربية بسوهاج

View full text Add to dashboard Cite

show abstract

“…The results showed that random forests and k nearest neighbor algorithm outperformed other algorithms. Arnhold and Kyrolainen (2017) investigated the focus marking by both random forests and the generalized additive mixed algorithm with the spotlight on the variable importance. This can help us understand phonetics and develop a speech scoring system that can be applied either in language testing or clinical linguistics, etc.…”

Section: Phoneticsmentioning

confidence: 99%

“…(Fromont et al, 2020;Heikel et al, 2018;Monner et al, 2013;Munsell et al, 2019;Pearl & Enverga, 2014) Phonetics(Al-Tamimi & Khattab, 2018;Arnhold & Kyrolainen, 2017;Bybee & De Souza, 2019;Charalabopoulou et al, 2011;Howell et al, 2017;Litman et al, 2018;Przybyla & Teisseyre, 2014;Xie et al, 2019) 4.1.1 Language Teaching and Language Learning…”

mentioning

confidence: 99%

A Methodological Review of Machine Learning in Applied Linguistics

Liu¹

2020

ELT

View full text Add to dashboard Cite

The traditional linear regression in applied linguistics (AL) suffers from the drawbacks arising from the strict assumptions namely: linearity, and normality, etc. More advanced methods are needed to overcome the shortcomings of the traditional method and grapple with intricate linguistic problems. However, there is no previous review on the applications of machine learning (ML) in AL, the introduction of interpretable ML, and related practical software. This paper addresses these gaps by reviewing the representative algorithms of ML in AL. The result shows that ML is applicable in AL and enjoys a promising future. It goes further to discuss the applications of interpretable ML for reporting the results in AL. Finally, it ends with the recommendations of the practical programming languages, software, and platforms to implement ML for researchers in AL to foster the interdisciplinary studies between AL and ML.

show abstract

“…7 Ladd, 2008, p. 48 Other studies assumed very phonetic definitions of what prominence is by characterising it with single acoustic parameters such as articulatory vocalic targets or jaw openings without providing insight into why those parameters were chosen (see Table 2.1). This view is in conflict with prosodic research which points out a complex relationship of acoustic cues to prominence where multiple cues are shown to contribute to the perception of prominence (see Breen, Fedorenko, Wagner, & Gibson, 2010;Cole, 2015;Arnhold & Kyröläinen, 2017;Kügler & Calhoun, in press;Baumann & Winter, 2018. The cues to prominence depend on multiple factors such as language in question and position in prosodic structure (see Section 2.5.1.2). Accordingly, if particular points in time are needed to be measured for synchronisation tests, this measure should not ignore prosodic structure and how the acoustic measure is used in the language in question.…”

Section: Previous Studiesmentioning

confidence: 98%

“…However, relevant research on prosody has shown that multiple acoustic cues contribute to the perception of prominence. It has also been shown that these cues cannot be expected to be the same for every syllable, regardless of factors (1) position in prosodic structure, (2)how those acoustic cues are used in that language (Breen et al, 2010;Arnhold & Kyröläinen, 2017;Baumann & Winter, 2018;Kügler & Calhoun, in press). Therefore, the prosodic anchors that are tested for synchronisation should be sensitive to these factors.…”

Section: Investigation Of Micro Level Synchronisation In the Present Studymentioning

confidence: 99%

Gesture, Prosody and Information Structure Synchronisation in Turkish

Turk¹

View full text Add to dashboard Cite

<p>This thesis investigates the synchronisation of gesture with prosody and information structure in Turkish. Speech and gesture have a close relationship in human communication, and they are tightly coordinated in production. Research has shown that gestural units are synchronised with prosodic units on a prominence-related micro level (i.e., pitch accents and gesture apexes), however these studies have largely been on a small number of languages of a similar prosodic type, not including Turkish, which has prominence-less prosodic words. It is known that both gesture and speech, through prosody, are hierarchically structured with nested phrasal constituents, but little is known about gesture-prosody synchronisation at this macro level. Even less is known about the timing relationships of gesture with information structure, which is also closely related to prosody. This thesis links gesture to information structure as a part of a three-way synchronisation relationship of gesture, prosody, and information structure. Four participants were filmed in a narrative task, resulting in three hours of Turkish natural speech and gesture data. Selected sections were annotated for prosody using an adapted scheme for Turkish in the Autosegmental-Metrical framework, for information structure and for gesture. In total, there were over 20,000 annotations. The synchronisation of gesture and speech units was systematically investigated at (1) the micro level, and (2) the macro level. At the micro level, this thesis asked which tones apexes are synchronised with, and whether this synchronisation depends on other prosodic and gestural features. It was found that gesture apexes were synchronised with pitch accents if there were pitch accents in the relevant prosodic phrases; if not, they were synchronised with low tones that marked the onsets of prosodic words. This synchronisation pattern was largely consistent across different prosodic and gestural contexts, although it was tighter in the nuclear area. These findings confirm prominence as a constraint on synchronisation with evidence of pitch accent-apex synchronisation. The findings also extend our knowledge of the typology of micro-level synchronisation to cases where prominence is locally absent showing that micro-level synchronisation also obeys the prosodic hierarchy. At the macro level, the aim was to find the prosodic anchor for single gesture phrases while testing for the possible effects of prosodic, gestural and information structural contexts. The findings showed that there was no one-to-one synchronisation of single gesture phrases with single intermediate or intonational phrases. However, it was found that gesture phrases often spanned over multiple consecutive intermediate phrases, and the synchronisation of gesture phrase boundaries was with the boundaries of these intermediate phrase groupings. In addition, these groupings tended to be combinations of pre-nuclear and nuclear intermediate phrases constituting the default focus position in Turkish. This synchronisation behaviour over the focal domain implied that there might be another speech element governing the speech-gesture synchronisation which also informs prosody, i.e. information structure. Based on this finding and a few other associations in the earlier studies, it was hypothesised that gesture is also informed by and synchronised with information structure. In order to test this hypothesis, it was investigated whether gesture phrases were synchronised with information structural units, i.e., topics, foci and background. The findings showed that gesture phrases tended to accompany discursively prominent foci over topics and background. However, gesture phrases did not show perfect synchronisation with any of these information structure units, although there was a systematic overlap in which foci and topics were contained within the duration of complete gesture phrases. Further investigations revealed that gesture phrase parts that bear apex related meaning provided a much better anchor for the synchronisation of information structure units. The preference for accompanying and synchronisation with the parts of gesture bearing gesturally prominent apical meaning also highlighted that prominence is a driving factor of synchronisation at the macro level as well as at the micro level. This thesis has revealed pivotal links between gesture, prosody and information structure through a systematic investigation of synchronisation of these structures. The implications of these links have also been discussed within the thesis, and a model of speech and gesture production integrating synchronisation has been proposed. Overall, the thesis contributes to a deeper understanding of speech and gesture production, explaining how these interact during natural speech.</p>

show abstract

Modelling the Interplay of Multiple Cues in Prosodic Focus Marking

Cited by 13 publications

References 69 publications

The Uses of Machine Learning (ML) in Teaching and Learning English Language: A Methodical Review.

The Uses of Machine Learning (ML) in Teaching and Learning English Language: A Methodical Review.

A Methodological Review of Machine Learning in Applied Linguistics

Gesture, Prosody and Information Structure Synchronisation in Turkish

Contact Info

Product

Resources

About