{"id":{"repo_id":"nus","oai_identifier":"oai:scholarbank.nus.edu.sg:10635/53706"},"canonical_url":"https://search.dev.ndltd.org/etd/nus/oai:scholarbank.nus.edu.sg:10635/53706","repository":{"repo_id":"nus","name":"National University of Singapore","base_url":"https://scholarbank.nus.edu.sg/oai/request"},"display":{"title":"Noise-Robust Speech Recognition Using Deep Neural Network","abstract":"This thesis addresses the noise robustness of the recently developed Deep Neural Networks (DNNs) based speech recognition systems. Five techniques have been proposed. Firstly, a Mean Variance Normalization technique was developed to integrate noise statistics using Vector Taylor Series. Secondly, a Deep Split Temporal Context system was proposed to separately model sub-contexts of a long temporal span of speech. Thirdly, we revisited the missing feature theory and developed a DNN-based spectral masking system. Fourthly, an Ideal Hidden-activation Mask (IHM) was proposed to remove noise-prone latent detectors. Lastly, a noise code technique was developed to simulate IHM with reduced computational costs. Improved noise robustness has been obtained using the proposed techniques on benchmark tasks, Aurora-2 and Aurora-4. The spectral masking approach successfully ranks first in the literature on both tasks at the time of writing and is one of the most promising noise-robust techniques for DNNs.","abstract_html":"This thesis addresses the noise robustness of the recently developed Deep Neural Networks (DNNs) based speech recognition systems. Five techniques have been proposed. Firstly, a Mean Variance Normalization technique was developed to integrate noise statistics using Vector Taylor Series. Secondly, a Deep Split Temporal Context system was proposed to separately model sub-contexts of a long temporal span of speech. Thirdly, we revisited the missing feature theory and developed a DNN-based spectral masking system. Fourthly, an Ideal Hidden-activation Mask (IHM) was proposed to remove noise-prone latent detectors. Lastly, a noise code technique was developed to simulate IHM with reduced computational costs. Improved noise robustness has been obtained using the proposed techniques on benchmark tasks, Aurora-2 and Aurora-4. The spectral masking approach successfully ranks first in the literature on both tasks at the time of writing and is one of the most promising noise-robust techniques for DNNs.","abstract_has_math":false,"creators":["LI BO"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["SIM KHE CHAI"],"committee_chairs":[],"committee_members":[],"year":2014,"date_issued":"2014-01-06","date_published":"2014-01-06","updated_at":"2026-08-21T16:47:23Z","subjects":["Speech Recognition, Deep Neural Network, Noise Robustness"],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://scholarbank.nus.edu.sg/handle/10635/53706","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"source_record":{"url":"https://scholarbank.nus.edu.sg/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Ascholarbank.nus.edu.sg%3A10635%2F53706","prefix":"dim"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.other","label":"Dc Contributor Other","values":["COMPUTER SCIENCE"]},{"key":"dc:contributor.supervisor","label":"Supervisor","values":["SIM KHE CHAI"]},{"key":"dc:creator","label":"Author","values":["LI BO"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2014-05-31T18:02:52Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2014-05-31T18:02:52Z"]},{"key":"dc:date.issued","label":"Date","values":["2014-01-06"]},{"key":"dc:relation.isreferencedby","label":"Dc Relation Isreferencedby","values":["https://scholarbank.nus.edu.sg/handle/10635/53706"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Speech Recognition, Deep Neural Network, Noise Robustness"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://scholarbank.nus.edu.sg/bitstreams/27d99b87-6955-40e8-b091-902ee8872bc2/download","https://scholarbank.nus.edu.sg/handle/10635/53706"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["This thesis addresses the noise robustness of the recently developed Deep Neural Networks (DNNs) based speech recognition systems. Five techniques have been proposed. Firstly, a Mean Variance Normalization technique was developed to integrate noise statistics using Vector Taylor Series. Secondly, a Deep Split Temporal Context system was proposed to separately model sub-contexts of a long temporal span of speech. Thirdly, we revisited the missing feature theory and developed a DNN-based spectral masking system. Fourthly, an Ideal Hidden-activation Mask (IHM) was proposed to remove noise-prone latent detectors. Lastly, a noise code technique was developed to simulate IHM with reduced computational costs. Improved noise robustness has been obtained using the proposed techniques on benchmark tasks, Aurora-2 and Aurora-4. The spectral masking approach successfully ranks first in the literature on both tasks at the time of writing and is one of the most promising noise-robust techniques for DNNs."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Ph.D"]},{"key":"dc:format.checksum.md5","label":"Dc Format Checksum Md5","values":["ad16ccfcbfc26b1688f6bf7e60ba927b","f399d68533f0e53e6ffbce59cd2947e9"]},{"key":"dc:title","label":"Title","values":["Noise-Robust Speech Recognition Using Deep Neural Network"]}]}],"canonical_facts":{"dc:contributor.other":["COMPUTER SCIENCE"],"dc:contributor.supervisor":["SIM KHE CHAI"],"dc:creator":["LI BO"],"dc:date.accessioned":["2014-05-31T18:02:52Z"],"dc:date.available":["2014-05-31T18:02:52Z"],"dc:date.issued":["2014-01-06"],"dc:description.abstract":["This thesis addresses the noise robustness of the recently developed Deep Neural Networks (DNNs) based speech recognition systems. Five techniques have been proposed. Firstly, a Mean Variance Normalization technique was developed to integrate noise statistics using Vector Taylor Series. Secondly, a Deep Split Temporal Context system was proposed to separately model sub-contexts of a long temporal span of speech. Thirdly, we revisited the missing feature theory and developed a DNN-based spectral masking system. Fourthly, an Ideal Hidden-activation Mask (IHM) was proposed to remove noise-prone latent detectors. Lastly, a noise code technique was developed to simulate IHM with reduced computational costs. Improved noise robustness has been obtained using the proposed techniques on benchmark tasks, Aurora-2 and Aurora-4. The spectral masking approach successfully ranks first in the literature on both tasks at the time of writing and is one of the most promising noise-robust techniques for DNNs."],"dc:description.degree":["Ph.D"],"dc:format.checksum.md5":["ad16ccfcbfc26b1688f6bf7e60ba927b","f399d68533f0e53e6ffbce59cd2947e9"],"dc:identifier.uri":["https://scholarbank.nus.edu.sg/bitstreams/27d99b87-6955-40e8-b091-902ee8872bc2/download","https://scholarbank.nus.edu.sg/handle/10635/53706"],"dc:language.iso":["en"],"dc:relation.isreferencedby":["https://scholarbank.nus.edu.sg/handle/10635/53706"],"dc:subject":["Speech Recognition, Deep Neural Network, Noise Robustness"],"dc:title":["Noise-Robust Speech Recognition Using Deep Neural Network"],"dc:type":["Thesis"]},"updated_at":"2026-08-21T16:47:23Z"}