Show simple item record

dc.contributor.authorDing, K
dc.contributor.authorLuo, N
dc.contributor.authorXu, Y
dc.contributor.authorKe, D
dc.contributor.authorSu, K
dc.date.accessioned2021-02-04T06:53:30Z
dc.date.available2021-02-04T06:53:30Z
dc.date.issued2018
dc.identifier.isbn9781538637883en_US
dc.identifier.issn1051-4651en_US
dc.identifier.doi10.1109/ICPR.2018.8546090en_US
dc.identifier.urihttp://hdl.handle.net/10072/401689
dc.description.abstractIn the context of Automatic Speech Recognition (ASR), improving the noise robustness remains an intractable task. Speech enhancement, combined with Generative Adversarial Networks (GAN), such as SEGAN, has effective performance in denoising raw waveform speech signals. Instead of waveforms, using Mel filterbank spectra in GAN is proposed, which has better performance in the task of ASR. However, these techniques will still miss useful information when GAN is used in them. In this paper, we investigate to protect the useful information in GAN, and propose a novel model, called Discriminator Generator Classifier-GAN (DGC-GAN). While normal GAN combining just two networks will lead the model to denoising rather than recognition, DGC-GAN has another network called classifier, which is an ASR system that will tune GAN to be recognized easier. By adding a classifier into previous GAN to get DGC-GAN, we achieve 29.1% Phone Error Rate (PER) relative improvement in a tiny dataset and 47.4% PER relative improvement in a large dataset.en_US
dc.description.peerreviewedYesen_US
dc.languageEnglishen_US
dc.publisherIEEEen_US
dc.relation.ispartofconferencename24th International Conference on Pattern Recognition (ICPR 2018)en_US
dc.relation.ispartofconferencetitleProceedings - International Conference on Pattern Recognitionen_US
dc.relation.ispartofdatefrom2018-08-20
dc.relation.ispartofdateto2018-08-24
dc.relation.ispartoflocationBeijing, Chinaen_US
dc.relation.ispartofpagefrom2699en_US
dc.relation.ispartofpageto2704en_US
dc.subject.fieldofresearchPattern Recognition and Data Miningen_US
dc.subject.fieldofresearchcode080109en_US
dc.subject.keywordsScience & Technologyen_US
dc.subject.keywordsComputer Science, Artificial Intelligenceen_US
dc.subject.keywordsautomatic speech recognitionen_US
dc.titleMutual-optimization Towards Generative Adversarial Networks for Robust Speech Recognitionen_US
dc.typeConference outputen_US
dc.type.descriptionE1 - Conferencesen_US
dcterms.bibliographicCitationDing, K; Luo, N; Xu, Y; Ke, D; Su, K, Mutual-optimization Towards Generative Adversarial Networks for Robust Speech Recognition, Proceedings - International Conference on Pattern Recognition, 2018, pp. 2699-2704en_US
dc.date.updated2021-02-04T06:51:12Z
gro.hasfulltextNo Full Text
gro.griffith.authorSu, Kaile


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

  • Conference outputs
    Contains papers delivered by Griffith authors at national and international conferences.

Show simple item record