Self-Training for End-to-End Speech Translation
INTERSPEECH, pp. 1476-1480, 2020.
EI
Abstract:
One of the main challenges for end-to-end speech translation is data scarcity. We leverage pseudo-labels generated from unlabeled audio by a cascade and an end-to-end speech translation model. This provides 8.3 and 5.7 BLEU gains over a strong semi-supervised baseline on the MuST-C English-French and English-German datasets, reaching st...More
Code:
Data:
Get fulltext within 24h
Other Links
Upload PDF
1.Your uploaded documents will be check within 24h, and coins will be credited to your account.
2.As the current system does not support cash withdrawal, you can add staff WeChat (AMxiaomai) to receive it as a red packet.
3.10 coins will be exchanged for 1 yuan.
?
¥
Upload a single paper
for 5 coins
Wechat's Red Packet
?
¥
Upload 50 articles
for 250 coins
Wechat's Red Packet
?
¥
Upload 200 articles
for 1000 coins
Wechat's Red Packet
?
¥
Upload 500 articles
for 2500 coins
Wechat's Red Packet
?
¥
Upload 1000 articles
for 5000 coins
Wechat's Red Packet
Tags
Comments