{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,3]],"date-time":"2026-06-03T13:05:19Z","timestamp":1780491919819,"version":"3.54.1"},"reference-count":83,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2021,1,1]],"date-time":"2021-01-01T00:00:00Z","timestamp":1609459200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,1,1]],"date-time":"2021-01-01T00:00:00Z","timestamp":1609459200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2021,1,1]],"date-time":"2021-01-01T00:00:00Z","timestamp":1609459200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Hong Kong Research","award":["14200218"],"award-info":[{"award-number":["14200218"]}]},{"name":"Hong Kong Research","award":["14200220"],"award-info":[{"award-number":["14200220"]}]},{"DOI":"10.13039\/501100010428","name":"Innovation and Technology Fund","doi-asserted-by":"publisher","award":["ITS\/254\/19"],"award-info":[{"award-number":["ITS\/254\/19"]}],"id":[{"id":"10.13039\/501100010428","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100010428","name":"Innovation and Technology Fund","doi-asserted-by":"publisher","award":["PiH\/350\/20"],"award-info":[{"award-number":["PiH\/350\/20"]}],"id":[{"id":"10.13039\/501100010428","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100010428","name":"Innovation and Technology Fund","doi-asserted-by":"publisher","award":["Shun Hing Institute of Advanced Engineering"],"award-info":[{"award-number":["Shun Hing Institute of Advanced Engineering"]}],"id":[{"id":"10.13039\/501100010428","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100010428","name":"Innovation and Technology Fund","doi-asserted-by":"publisher","award":["MMT-p1-19"],"award-info":[{"award-number":["MMT-p1-19"]}],"id":[{"id":"10.13039\/501100010428","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE\/ACM Trans. Audio Speech Lang. Process."],"published-print":{"date-parts":[[2021]]},"DOI":"10.1109\/taslp.2021.3069080","type":"journal-article","created":{"date-parts":[[2021,3,26]],"date-time":"2021-03-26T19:54:35Z","timestamp":1616788475000},"page":"1514-1529","source":"Crossref","is-referenced-by-count":15,"title":["Bayesian Learning of LF-MMI Trained Time Delay Neural Networks for Speech Recognition"],"prefix":"10.1109","volume":"29","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3345-6923","authenticated-orcid":false,"given":"Shoukang","family":"Hu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xurong","family":"Xie","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6202-5615","authenticated-orcid":false,"given":"Shansong","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2449-1436","authenticated-orcid":false,"given":"Jianwei","family":"Yu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zi","family":"Ye","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mengzhe","family":"Geng","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6725-1160","authenticated-orcid":false,"given":"Xunying","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Helen","family":"Meng","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref73","first-page":"1245","article-title":"Fast speaker adaptive training for speech recognition","author":"povey","year":"0","journal-title":"Proc 9th Annu Conf Int Speech Commun Assoc"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.1997.596119"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/ICSLP.1996.607807"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1006\/csla.1998.0043"},{"key":"ref76","first-page":"3145","article-title":"The cambridge university 2014 bolt conversational telephone mandarin chinese LVCSR system for speech translation","author":"liu","year":"0","journal-title":"Proc 16th Annu Conf Int Speech Commun Assoc"},{"key":"ref77","article-title":"Parameterised sigmoid and relu hidden activation functions for DNN acoustic modelling","author":"zhang","year":"0","journal-title":"Proc 16th Annu Conf Int Speech Commun Assoc"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1109\/TASL.2010.2064307"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2019.8683066"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-2162"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/72.317730"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2014.6854049"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1001\/archneur.1994.00540180063015"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/ASRU.2013.6707747"},{"key":"ref32","article-title":"Scalable minimum bayes risk training of deep neural network acoustic models using distributed hessian-free optimization","author":"kingsbury","year":"2012","journal-title":"Annu Conf Int Speech Commun Assoc"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2017.7953152"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-2680"},{"key":"ref37","article-title":"Parallel training of deep neural networks with natural gradient and parameter averaging","author":"povey","year":"0","journal-title":"3rd Int Conf Learn Representations Workshop Track Proc"},{"key":"ref36","first-page":"849","article-title":"Topmoumoute online natural gradient algorithm","author":"roux","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1162\/089976698300017007"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2016.2626289"},{"key":"ref60","article-title":"Steps toward deep kernel methods from infinite neural networks","author":"hazan","year":"2015"},{"key":"ref62","first-page":"207","article-title":"Deep Gaussian processes","author":"damianou","year":"0","journal-title":"Proc Artif Intell Statist"},{"key":"ref61","article-title":"Deep neural networks as Gaussian processes","author":"lee","year":"0"},{"key":"ref63","article-title":"Auto-encoding variational bayes","author":"kingma","year":"0","journal-title":"Proc 2nd Int Conf Learn Representations"},{"key":"ref28","first-page":"-105i","article-title":"Minimum phone error and I-smoothing for improved discriminative training","author":"povey","year":"0","journal-title":"Proc IEEE Int Conf Acoust Speech Signal Process"},{"key":"ref64","first-page":"2980","article-title":"A recurrent latent variable model for sequential data","author":"chung","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref27","article-title":"Automatic speech recognition: A deep learning approach","author":"benesty","year":"2008"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2017-832"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-1927"},{"key":"ref29","first-page":"3586","article-title":"Audio Augmentation for Speech Recognition","author":"ko","year":"2015","journal-title":"InterSpeech"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/29.21701"},{"key":"ref68","article-title":"Convolutional networks for images, speech, and time series","volume":"3361","author":"lecun","year":"1995","journal-title":"Handbook Brain Theory Neural Netw"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1006\/csla.1995.0010"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2008.4518545"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.1986.1169179"},{"key":"ref20","first-page":"1","article-title":"A time delay neural network architecture for efficient modeling of long temporal contexts","author":"peddinti","year":"0","journal-title":"Proc INTERSPEECH"},{"key":"ref22","article-title":"Rwth asr systems for librispeech: Hybrid vs attention-w\/o data augmentation","author":"l\u00fcscher","year":"2019"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2018-1417"},{"key":"ref24","first-page":"1","article-title":"On the comparison of popular end-to-end models for large scale speech recognition","author":"li","year":"2020","journal-title":"InterSpeech"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9053573"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1006\/csla.2001.0182"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1016\/S0167-6393(97)00029-0"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2015.2499302"},{"key":"ref51","article-title":"Single headed attention based sequence-to-sequence model for state-of-the-art results on switchboard-300","author":"t\u00fcske","year":"0"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4612-0745-0_2"},{"key":"ref58","volume":"1","author":"rasmussen","year":"2006","journal-title":"Gaussian Processes for Machine Learning"},{"key":"ref57","first-page":"1613","article-title":"Weight uncertainty in neural network","volume":"37","author":"blundell","year":"2015","journal-title":"Ser Proceedings Machine Learning Research"},{"key":"ref56","first-page":"2348","article-title":"Practical variational inference for neural networks","author":"graves","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref55","author":"bishop","year":"2006","journal-title":"Pattern Recognition and Machine Learning"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1992.4.3.448"},{"key":"ref53","volume":"118","author":"neal","year":"2012","journal-title":"Bayesian learning for neural networks"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/ASRU.2017.8268947"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2016-595"},{"key":"ref11","first-page":"1209","article-title":"Sequence discriminative distributed training of long short-term memory recurrent neural networks","author":"sak","year":"0","journal-title":"Proc INTERSPEECH"},{"key":"ref40","first-page":"1929","article-title":"Dropout: A simple way to prevent neural networks from overfitting","author":"srivastava","year":"0","journal-title":"Proc J Mach Learn Res"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9053499"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2016.7472621"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/1143844.1143891"},{"key":"ref15","article-title":"Sequence transduction with recurrent neural networks","author":"graves","year":"2012"},{"key":"ref82","first-page":"1076","article-title":"Maximum a posteriori adaptation of network parameters in deep models","author":"huang","year":"0","journal-title":"Proc INTERSPEECH"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9054345"},{"key":"ref81","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2015.7178964"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2018-1423"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-2254"},{"key":"ref83","first-page":"2020","article-title":"Neural architecture search with Bayesian optimisation and optimal transport","author":"kandasamy","year":"0","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.1989.266376"},{"key":"ref80","article-title":"Development of the CUHK elderly speech recognition system for neurocognitive disorder detection using the dementiabank corpus","author":"ye","year":"2021","journal-title":"Proc IEEE Int Conf Acoust Speech Signal Process"},{"key":"ref4","first-page":"887","article-title":"A novel loss function for the overall risk criterion based discriminative training of HMM models","author":"kaiser","year":"0","journal-title":"Proc Int Conf Spoken Lang Process"},{"key":"ref3","article-title":"Discriminative training for large vocabulary speech recognition","author":"povey","year":"2005"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2007.366914"},{"key":"ref5","article-title":"Hypothesis spaces for minimum bayes risk training in large vocabulary speech recognition","author":"gibson","year":"0","journal-title":"Proc Int Conf Spoken Lang Process"},{"key":"ref8","first-page":"2345","article-title":"Sequence-discriminative training of deep neural networks","volume":"2013","author":"vesel?","year":"0","journal-title":"InterSpeech"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2009.4960445"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1109\/SLT.2014.7078569"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2013.6638951"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.21437\/Odyssey.2020-51"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2019.8682667"},{"key":"ref48","article-title":"Implementation of the standard i-vector system for the kaldi speech recognition toolkit","author":"madikeri","year":"2016","journal-title":"IDIAP tech report"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/TASL.2010.2064307"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2018-1823"},{"key":"ref41","first-page":"1050","article-title":"Dropout as a bayesian approximation: Representing model uncertainty in deep learning","author":"gal","year":"0","journal-title":"Proc 33nd Int Conf Mach Learn"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2019-2379"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2019.8682487"}],"container-title":["IEEE\/ACM Transactions on Audio, Speech, and Language Processing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/6570655\/9289074\/09387600.pdf?arnumber=9387600","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,5,10]],"date-time":"2022-05-10T14:53:55Z","timestamp":1652194435000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9387600\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021]]},"references-count":83,"URL":"https:\/\/doi.org\/10.1109\/taslp.2021.3069080","relation":{},"ISSN":["2329-9290","2329-9304"],"issn-type":[{"value":"2329-9290","type":"print"},{"value":"2329-9304","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021]]}}}