DSpace Repository

Tackling Perception Bias in Unsupervised Phoneme Discovery Using DPGMM-RNN Hybrid Model and Functional Load

Show simple item record

dc.contributor.author Wu, Bin en
dc.contributor.author Sakti, Sakriani en
dc.contributor.author Zhang, Jinsong en
dc.contributor.author Nakamura, Satoshi en
dc.date.accessioned 2021-01-07T06:30:57Z en
dc.date.available 2021-01-07T06:30:57Z en
dc.date.issued 2020-12-02 en
dc.identifier.uri http://hdl.handle.net/10061/14206 en
dc.description.abstract The human perception of phonemes is biased against speech sounds. The lack of correspondence between perceputal phonemes and acoustic signals forms a big challenge in designing unsupervised algorithms to distinguish phonemes from sound. We propose the DPGMM-RNN hybrid model that improves phoneme categorization by relieving the fragmentation problem. We also merge segments with low functional load, which is the work done by segment contrasts to differentiate between utterances, just like humans who convert unambiguous segments into phonemes as units for immediate perception. Our results show that the DPGMM-RNN hybrid model relieves the fragmentation problem and improves phoneme discriminability. The minimal functional load merge compresses a segment system, preserves information and keeps phoneme discriminability. en
dc.language.iso en en
dc.publisher IEEE en
dc.relation.isreplacedby https://ieeexplore.ieee.org/document/9276474 en
dc.rights This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ en
dc.subject Acoustics en
dc.subject Clustering algorithms en
dc.subject Auditory system en
dc.subject Ear en
dc.subject Visualization en
dc.subject Load modeling en
dc.subject Context modeling en
dc.title Tackling Perception Bias in Unsupervised Phoneme Discovery Using DPGMM-RNN Hybrid Model and Functional Load en
dc.type.nii Journal Article en
dc.contributor.transcription ナカムラ, サトシ ja
dc.contributor.alternative 中村, 哲 ja
dc.textversion none en
dc.identifier.eissn 2329-9304 en
dc.identifier.jtitle IEEE/ACM Transactions on Audio, Speech, and Language Processing en
dc.identifier.volume 29 en
dc.identifier.spage 348 en
dc.identifier.epage 362 en
dc.relation.doi 10.1109/TASLP.2020.3042016 en
dc.identifier.NAIST-ID 85629731 en
dc.identifier.NAIST-ID 73297715 en
dc.identifier.NAIST-ID 73296626 en


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account