Please use this identifier to cite or link to this item:
http://hdl.handle.net/10397/76009
DC Field | Value | Language |
---|---|---|
dc.contributor | Department of Electronic and Information Engineering | en_US |
dc.creator | Lin, WW | en_US |
dc.creator | Mak, MW | en_US |
dc.creator | Chien, JT | en_US |
dc.date.accessioned | 2018-05-10T02:55:08Z | - |
dc.date.available | 2018-05-10T02:55:08Z | - |
dc.identifier.issn | 0885-2308 | en_US |
dc.identifier.uri | http://hdl.handle.net/10397/76009 | - |
dc.language.iso | en | en_US |
dc.publisher | Academic Press | en_US |
dc.rights | © 2017 Elsevier Ltd. All rights reserved. | en_US |
dc.rights | © 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/. | en_US |
dc.rights | The following publication Lin, W. W., Mak, M. W., & Chien, J. T. (2017). Fast scoring for PLDA with uncertainty propagation via i-vector grouping. Computer Speech & Language, 45, 503-515 is available at https://doi.org/10.1016/j.csl.2017.02.009. | en_US |
dc.subject | Speaker verification | en_US |
dc.subject | I-Vector/PLDA | en_US |
dc.subject | Uncertainty Propagation | en_US |
dc.subject | Duration mismatch | en_US |
dc.title | Fast scoring for PLDA with uncertainty propagation via i-vector grouping | en_US |
dc.type | Journal/Magazine Article | en_US |
dc.identifier.spage | 503 | en_US |
dc.identifier.epage | 515 | en_US |
dc.identifier.volume | 45 | en_US |
dc.identifier.doi | 10.1016/j.csl.2017.02.009 | en_US |
dcterms.abstract | The i-vector/PLDA framework has gained huge popularity in text-independent speaker verification. This approach, however, lacks the ability to represent the reliability of i-vectors. As a result, the framework performs poorly when presented with utterances of arbitrary duration. To address this problem, a method called uncertainty propagation (UP) was proposed to explicitly model the reliability of an i-vector by an utterance-dependent loading matrix. However, the utterance-dependent matrix greatly complicates the evaluation of likelihood scores. As a result, PLDA with UP, or PLDA-UP in short, is far more computational intensive than the conventional PLDA. In this paper, we propose to group i-vectors with similar reliability, and for each group the utterance-dependent loading matrices are replaced by a representative one. This arrangement allows us to pre-compute a set of representative matrices that cover all possible i-vectors, thereby greatly reducing the computational cost of PLDA-UP while preserving its ability in discriminating the reliability of i-vectors. Experiments on NIST 2012 SRE show that the proposed method can perform as good as the PLDA with UP while the scoring time is only 3.18% of it. | en_US |
dcterms.accessRights | open access | en_US |
dcterms.bibliographicCitation | Computer speech and language, Sept. 2017, v. 45, p. 503-515 | en_US |
dcterms.isPartOf | Computer speech and language | en_US |
dcterms.issued | 2017-09 | - |
dc.identifier.isi | WOS:000403510500026 | - |
dc.identifier.eissn | 1095-8363 | en_US |
dc.identifier.rosgroupid | 2017004686 | - |
dc.description.ros | 2017-2018 > Academic research: refereed > Publication in refereed journal | en_US |
dc.description.validate | 201805 bcrc | en_US |
dc.description.oa | Accepted Manuscript | en_US |
dc.identifier.FolderNumber | EIE-0663 | - |
dc.description.fundingSource | RGC | en_US |
dc.description.pubStatus | Published | en_US |
dc.identifier.OPUS | 6732232 | - |
Appears in Collections: | Journal/Magazine Article |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Mak_Fast_Scoring_Plda.pdf | Pre-Published version | 966.47 kB | Adobe PDF | View/Open |
Page views
90
Last Week
0
0
Last month
Citations as of Apr 21, 2024
Downloads
28
Citations as of Apr 21, 2024
SCOPUSTM
Citations
8
Last Week
0
0
Last month
Citations as of Apr 19, 2024
WEB OF SCIENCETM
Citations
8
Last Week
0
0
Last month
Citations as of Apr 25, 2024
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.