PART-OF-SPEECH MODELS COMPRESSION METHODS FOR ON-DEVICE GRAPHEME-TO-PHONEME CONVERSION

被引：0
|
作者：
Kubis, Marek [1 ]
Meloux, Maxime [2 ]
Skorzewski, Pawel [1 ]
Lewandowski, Marcin [2 ]
Jho, Gunu [3 ]
Park, Hyoungmin [3 ]
机构：
[1] Adam Mickiewicz Univ, Poznan, Poland
[2] Samsung R&D Inst Poland, Warsaw, Poland
[3] Samsung Elect, Mobile Commun Business, Suwon, South Korea
来源：
2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) | 2022年
关键词：
part-of-speech tagging; model compression; grapheme-to-phoneme conversion;
D O I：
10.1109/ICASSP43922.2022.9746710
中图分类号：
O42 [声学];
学科分类号：
070206 ; 082403 ;
摘要：
The paper investigates methods of compressing part-of-speech models that are developed for an on-device graphemeto-phoneme conversion module. The performance of partof-speech models is analyzed under different compression regimes. The evaluation is done with respect to French, German and Italian datasets that consist of TTS input prompts. The study shows that a proper selection of a compression method reduces the model size significantly without deteriorating the grapheme-to-phoneme conversion performance.
引用
收藏
页码：7117 / 7121
页数：5
相关论文
共 50 条
  • [1] Joint-sequence models for grapheme-to-phoneme conversion
    Bisani, Maximilian
    Ney, Hermann
    [J]. SPEECH COMMUNICATION, 2008, 50 (05) : 434 - 451
  • [2] NEURAL GRAPHEME-TO-PHONEME CONVERSION WITH PRE-TRAINED GRAPHEME MODELS
    Dong, Lu
    Guo, Zhi-Qiang
    Tan, Chao-Hong
    Hu, Ya-Jun
    Jiang, Yuan
    Ling, Zhen-Hua
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 6202 - 6206
  • [3] GRAPHEME-TO-PHONEME CONVERSION METHODS FOR MINORITY LANGUAGE CONDITIONS
    Cao, Mengxue
    Renals, Steve
    Bell, Peter
    Li, Aijun
    Fang, Qiang
    [J]. 2012 INTERNATIONAL CONFERENCE ON SPEECH DATABASE AND ASSESSMENTS, 2012, : 151 - 156
  • [4] Compression of exception lexicons for small footprint grapheme-to-phoneme conversion
    Meron, J
    Veprek, P
    [J]. 2005 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1-5: SPEECH PROCESSING, 2005, : 293 - 296
  • [5] Fast Bilingual Grapheme-To-Phoneme Conversion
    Kim, Hwa-Yeon
    Kim, Jong-Hwan
    Kim, Jae-Min
    [J]. 2022 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, NAACL-HLT 2022, 2022, : 289 - 296
  • [6] Transformer based Grapheme-to-Phoneme Conversion
    Yolchuyeva, Sevinj
    Nemeth, Geza
    Gyires-Toth, Balint
    [J]. INTERSPEECH 2019, 2019, : 2095 - 2099
  • [7] Multitask Learning for Grapheme-to-Phoneme Conversion of Anglicisms in German Speech Recognition
    Pritzen, Julia
    Gref, Michael
    Zuehlke, Dietlind
    Schmidt, Christoph
    [J]. LREC 2022: THIRTEEN INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION, 2022, : 3242 - 3249
  • [8] Multitask Sequence-to-Sequence Models for Grapheme-to-Phoneme Conversion
    Milde, Benjamin
    Schmidt, Christoph
    Koehler, Joachim
    [J]. 18TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2017), VOLS 1-6: SITUATED INTERACTION, 2017, : 2536 - 2540
  • [9] Grapheme-to-Phoneme Conversion for Thai using Neural Regression Models
    Yamasaki, Tomohiro
    [J]. NAACL 2022: THE 2022 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES, 2022, : 4251 - 4255
  • [10] BAYESIAN JOINT-SEQUENCE MODELS FOR GRAPHEME-TO-PHONEME CONVERSION
    Hannemann, Mirko
    Trmal, Jan
    Ondel, Lucas
    Kesiraju, Santosh
    Burget, Lukas
    [J]. 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2017, : 2836 - 2840
12345