lenition
use speech as biomarker for medical and linguistic purposes
I have been working as a graduate research assistant since 2021 for this NSF project led by Dr. Ratree Wayland and Dr. Kevin Tang. For this project, we are trying to integrate deep learning models with articulatory and acoustic data to predict degrees of lenition in different types of speech, with an aim to provide a unique form of diagnostics medical and linguistic purposes.
To measure the degree of lenition, we have trained a deep learning Phonet model to generate the posterior probabilities of sonorant and continuant features of Spanish stops in different contexts and compare them with previously reported matrics of lenition (e.g., Broś et al, 2021; Kingston, 2008).
Related Publications
2023
- LAN
- JASAFrom sonority hierarchy to posterior probability as a measure of lenition: The case of Spanish stopsThe Journal of the Acoustical Society of America, 2023
- ICPhSMeasuring gradient effects of alcohol on speech with neural networks’ posterior probability of phonological features (accepted)In Proceedings of the 20th International Congress of the Phonetic Sciences, 2023
- POMANeural networks’ posterior probability as measure of effects of alcohol on speechIn Proceedings of Meetings on Acoustics, 2023
- POMALenition measures: Neural networks’ posterior probability versus acoustic cuesIn Proceedings of Meetings on Acoustics, 2023
- ASANeural networks’ posterior probability as measure of effects of alcohol on speechIn The 184th Meeting of Acoustical Society of America, 2023
2022
- ASALenition measures: Neural networks’ posterior probability versus acoustic cuesIn The 183rd Meeting of Acoustical Society of America, 2022
- ASAMeasuring second language acquisition of spanish lenitionIn The 183rd Meeting of Acoustical Society of America, 2022