Worse WER, but Better BLEU? Leveraging Word Embedding as Intermediate in Multitask End-to-End Speech Translation.
Shun-Po ChuangTzu-Wei SungAlexander H. LiuHung-yi LeePublished in: ACL (2020)
Keyphrases
- end to end
- multi task
- word error rate
- language model
- statistical machine translation
- translation model
- speech recognition
- machine translation
- word level
- machine translation system
- automatic speech recognition
- multitask learning
- multi task learning
- word alignment
- language modeling
- learning tasks
- n gram
- target language
- handwriting recognition
- source language
- cross language information retrieval
- multiple tasks
- gaussian processes
- probabilistic model
- learning problems
- cross lingual
- speech signal
- transfer learning
- congestion control
- multi class
- feature selection
- word recognition
- word segmentation
- hidden markov models
- out of vocabulary
- information retrieval
- query translation
- broadcast news
- machine learning
- natural language processing
- information extraction