NICT at MixMT 2022: Synthetic Code-Mixed Pre-training and Multi-way Fine-tuning for Hinglish-English Translation.
Raj DabrePublished in: WMT (2022)
Keyphrases
- fine tuning
- machine translation
- fine tuned
- statistical machine translation
- query translation
- minimum error rate
- fine tune
- cross language
- cross language information retrieval
- viable alternative
- source language
- machine translation system
- language learning
- target language
- training program
- training process
- chinese english
- training phase
- english language
- natural language processing
- cross language retrieval
- language resources
- cross language ir
- training set
- parallel corpora
- training corpus
- real world
- bi level
- broadcast news
- parallel corpus
- translation model
- cross lingual
- source code
- domain specific
- supervised learning
- proper names
- hidden markov models
- pronominal anaphora