FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2892025684> ?p ?o ?g. }

• W2892025684 abstract "This paper proposes a framework that combines teacher-student training and permutation invariant training (PIT) for single-channel multi-talker speech recognition. In contrast to most of conventional teacher-student training methods that aim at compressing the model, the proposed method distills knowledge from the single-talker model to improve the multi-talker model in the PIT framework. The inputs to the teacher and student networks are the single-talker clean speech and the multi-talker mixed speech, respectively. The knowledge is transferred to the student through the soft labels generated by the teacher. Furthermore, the ensemble of multiple teachers is exploited with a progressive training scheme to further improve the system. In this framework it is easy to take advantage of data augmentation and perform domain adaptation for multi-talker speech recognition using only untranscribed data. The proposed techniques were evaluated on artificially mixed two-talker AMI speech data. The experimental results show that the teacher-student training can cut the word error rate (WER) by relative 20% against the baseline PIT model. We also evaluated our unsupervised domain adaptation method on an artificially mixed WSJO corpus and achieved relative 30% WER reduction against the AMI PIT model." @default.
• W2892025684 created "2018-09-27" @default.
• W2892025684 creator A5015824704 @default.
• W2892025684 creator A5034476404 @default.
• W2892025684 creator A5039523706 @default.
• W2892025684 date "2018-04-01" @default.
• W2892025684 modified "2023-09-23" @default.
• W2892025684 title "Knowledge Transfer in Permutation Invariant Training for Single-Channel Multi-Talker Speech Recognition" @default.
• W2892025684 cites W1524333225 @default.
• W2892025684 cites W1989549063 @default.
• W2892025684 cites W1995562189 @default.
• W2892025684 cites W2015633636 @default.
• W2892025684 cites W2033310064 @default.
• W2892025684 cites W2036351241 @default.
• W2892025684 cites W2060822897 @default.
• W2892025684 cites W2062164080 @default.
• W2892025684 cites W2079623482 @default.
• W2892025684 cites W2085628288 @default.
• W2892025684 cites W2087006792 @default.
• W2892025684 cites W2091425152 @default.
• W2892025684 cites W2094147890 @default.
• W2892025684 cites W2121415728 @default.
• W2892025684 cites W2134797427 @default.
• W2892025684 cites W2136933783 @default.
• W2892025684 cites W2144763279 @default.
• W2892025684 cites W2147768505 @default.
• W2892025684 cites W2181607856 @default.
• W2892025684 cites W2221409856 @default.
• W2892025684 cites W2402040300 @default.
• W2892025684 cites W2402146185 @default.
• W2892025684 cites W2407022425 @default.
• W2892025684 cites W2460742184 @default.
• W2892025684 cites W2508418541 @default.
• W2892025684 cites W2515753980 @default.
• W2892025684 cites W2515863432 @default.
• W2892025684 cites W2519224033 @default.
• W2892025684 cites W2558649592 @default.
• W2892025684 cites W2610674366 @default.
• W2892025684 cites W2711861986 @default.
• W2892025684 cites W2734774145 @default.
• W2892025684 cites W2962715207 @default.
• W2892025684 cites W2962894366 @default.
• W2892025684 cites W2963773971 @default.
• W2892025684 cites W2963843276 @default.
• W2892025684 cites W2963864497 @default.
• W2892025684 cites W3099832538 @default.
• W2892025684 cites W330298975 @default.
• W2892025684 cites W1922452246 @default.
• W2892025684 cites W2346991294 @default.
• W2892025684 cites W2416983298 @default.
• W2892025684 cites W2469481911 @default.
• W2892025684 cites W2481515426 @default.
• W2892025684 cites W2613889282 @default.
• W2892025684 cites W2709779212 @default.
• W2892025684 cites W2766410052 @default.
• W2892025684 cites W2769236969 @default.
• W2892025684 cites W2777648877 @default.
• W2892025684 cites W2781406697 @default.
• W2892025684 cites W2782011388 @default.
• W2892025684 cites W2812722839 @default.
• W2892025684 cites W2846467470 @default.
• W2892025684 cites W2890505457 @default.
• W2892025684 cites W2928921291 @default.
• W2892025684 cites W2930281027 @default.
• W2892025684 cites W2930358753 @default.
• W2892025684 cites W2931802168 @default.
• W2892025684 cites W2932556634 @default.
• W2892025684 cites W2932911757 @default.
• W2892025684 cites W2957200381 @default.
• W2892025684 cites W2957422282 @default.
• W2892025684 cites W2958473437 @default.
• W2892025684 cites W2960136726 @default.
• W2892025684 cites W3140637029 @default.
• W2892025684 cites W335191797 @default.
• W2892025684 doi "https://doi.org/10.1109/icassp.2018.8461883" @default.
• W2892025684 hasPublicationYear "2018" @default.
• W2892025684 type Work @default.
• W2892025684 sameAs 2892025684 @default.
• W2892025684 citedByCount "14" @default.
• W2892025684 countsByYear W28920256842018 @default.
• W2892025684 countsByYear W28920256842019 @default.
• W2892025684 countsByYear W28920256842020 @default.
• W2892025684 countsByYear W28920256842021 @default.
• W2892025684 crossrefType "proceedings-article" @default.
• W2892025684 hasAuthorship W2892025684A5015824704 @default.
• W2892025684 hasAuthorship W2892025684A5034476404 @default.
• W2892025684 hasAuthorship W2892025684A5039523706 @default.
• W2892025684 hasConcept C121332964 @default.
• W2892025684 hasConcept C150899416 @default.
• W2892025684 hasConcept C154945302 @default.
• W2892025684 hasConcept C155635449 @default.
• W2892025684 hasConcept C190470478 @default.
• W2892025684 hasConcept C21308566 @default.
• W2892025684 hasConcept C24890656 @default.
• W2892025684 hasConcept C28490314 @default.
• W2892025684 hasConcept C33923547 @default.
• W2892025684 hasConcept C37914503 @default.
• W2892025684 hasConcept C40969351 @default.
• W2892025684 hasConcept C41008148 @default.
• W2892025684 hasConcept C51632099 @default.

• next