Publikace
Detail publikace
Citace
p. 236-245, 2010. : Speaker adaptation of language and prosodic models for automatic dialog act segmentation of speech . Speech Communication, vol. 52,
Další informace
Full paper at Sciencedirect.com
Abstrakt
Speaker-dependent modeling has a long history in speech recognition, but has received less attention in speech understanding. This study explores speaker-specific modeling for the task of automatic segmentation of speech into dialog acts (DAs), using a linear combination of speaker-dependent and speaker-independent language and prosodic models. Data come from 20 frequent speakers in the ICSI meeting corpus; adaptation data per speaker ranges from 5k to 115k words. We compare performance for both reference transcripts and automatic speech recognition output. We find that: (1) speaker adaptation in this domain results both in a significant overall improvement and in improvements for many individual speakers, (2) the magnitude of improvement for individual speakers does not depend on the amount of adaptation data, and (3) language and prosodic models differ both in degree of improvement, and in relative benefit for specific DA classes. These results suggest important future directions for speaker-specific modeling in spoken language understanding tasks.
Abstrakt v češtině
Speaker-dependent modeling has a long history in speech recognition, but has received less attention in speech understanding. This study explores speaker-specific modeling for the task of automatic segmentation of speech into dialog acts (DAs), using a linear combination of speaker-dependent and speaker-independent language and prosodic models. Data come from 20 frequent speakers in the ICSI meeting corpus; adaptation data per speaker ranges from 5k to 115k words. We compare performance for both reference transcripts and automatic speech recognition output. We find that: (1) speaker adaptation in this domain results both in a significant overall improvement and in improvements for many individual speakers, (2) the magnitude of improvement for individual speakers does not depend on the amount of adaptation data, and (3) language and prosodic models differ both in degree of improvement, and in relative benefit for specific DA classes. These results suggest important future directions for speaker-specific modeling in spoken language understanding tasks.
Detail publikace
Název: | Speaker adaptation of language and prosodic models for automatic dialog act segmentation of speech |
---|---|
Autor: | Jáchym Kolář ; Yang Liu ; Elizabeth Shriberg |
Název - česky: | Speaker adaptation of language and prosodic models for automatic dialog act segmentation of speech |
Jazyk publikace: | česky |
Datum vydání: | 1.3.2010 |
Rok vydání: | 2010 |
Typ publikace: | Článek z časopisu |
Název časopisu / knihy: | Speech Communication |
Číslo vydání: | 52 |
Strana: | 236 - 245 |
Klíčová slova
Spoken language understanding, Dialog act segmentation, Speaker adaptation, Prosody modeling, Language modeling
Klíčová slova v češtině
Spoken language understanding, Dialog act segmentation, Speaker adaptation, Prosody modeling, Language modeling
BibTeX
@ARTICLE{JachymKolar_2010_Speakeradaptationof, author = {J\'{a}chym Kol\'{a}\v{r} and Yang Liu and Elizabeth Shriberg}, title = {Speaker adaptation of language and prosodic models for automatic dialog act segmentation of speech}, year = {2010}, journal = {Speech Communication}, volume = {52}, pages = {236-245}, note = {Available Online}, url = {http://www.kky.zcu.cz/en/publications/JachymKolar_2010_Speakeradaptationof}, }