Align-to-Distill: Trainable Attention Alignment for Knowledge Distillation in Neural Machine Translation.
Heegon JinSeonil SonJemin ParkYoungseok KimHyungjong NohYeonsoo LeePublished in: CoRR (2024)
Keyphrases
- machine translation
- word alignment
- cross lingual
- language processing
- language independent
- word level
- natural language
- natural language processing
- word sense disambiguation
- knowledge base
- information extraction
- target language
- statistical machine translation
- natural language generation
- machine translation system
- chinese english
- language resources
- cross language information retrieval
- artificial intelligence
- parallel corpora
- knowledge representation
- multilingual documents
- statistical translation models
- query translation
- precision and recall
- relevant documents
- document collections
- semi supervised
- expert systems