That Sounds Familiar: an Analysis of Phonetic Representations Transfer Across Languages

被引:6
|
作者
Zelasko, Piotr [1 ]
Moro-Velazquez, Laureano [1 ]
Hasegawa-Johnson, Mark [3 ,4 ]
Scharenborg, Odette [5 ]
Dehak, Najim [1 ,2 ]
机构
[1] Johns Hopkins Univ, Ctr Language & Speech Proc, Baltimore, MD 21218 USA
[2] Johns Hopkins Univ, Human Language Technol Ctr Excellence, Baltimore, MD USA
[3] Univ Illinois, ECE Dept, Urbana, IL USA
[4] Univ Illinois, Beckman Inst, Urbana, IL USA
[5] Delft Univ Technol, Multimedia Comp Grp, Delft, Netherlands
来源
关键词
speech recognition; multilingual; crosslingual; transfer learning; zero-shot; phone recognition;
D O I
10.21437/Interspeech.2020-2513
中图分类号
R36 [病理学]; R76 [耳鼻咽喉科学];
学科分类号
100104 ; 100213 ;
摘要
Only a handful of the world's languages are abundant with the resources that enable practical applications of speech processing technologies. One of the methods to overcome this problem is to use the resources existing in other languages to train a multilingual automatic speech recognition (ASR) model, which, intuitively, should learn some universal phonetic representations. In this work, we focus on gaining a deeper understanding of how general these representations might be, and how individual phones are getting improved in a multilingual setting. To that end, we select a phonetically diverse set of languages, and perform a series of monolingual, multilingual and crosslingual (zero-shot) experiments. The ASR is trained to recognize the International Phonetic Alphabet (IPA) token sequences. We observe significant improvements across all languages in the multilingual setting, and stark degradation in the crosslingual setting, where the model, among other errors, considers Javanese as a tone language. Notably, as little as 10 hours of the target language training data tremendously reduces ASR error rates. Our analysis uncovered that even the phones that are unique to a single language can benefit greatly from adding training data from other languages - an encouraging result for the low-resource speech community.
引用
收藏
页码:3705 / 3709
页数:5
相关论文
共 50 条