Revisiting the Weaknesses of Reinforcement Learning for Neural Machine Translation.
Samuel KiegelandJulia KreutzerPublished in: CoRR (2021)
Keyphrases
- machine translation
- reinforcement learning
- natural language processing
- language processing
- target language
- language independent
- cross lingual
- information extraction
- natural language generation
- statistical machine translation
- word sense disambiguation
- cross language information retrieval
- natural language
- machine translation system
- markov decision processes
- state space
- optimal policy
- source language
- machine learning
- language resources
- parallel corpora
- multilingual documents
- associative memory
- word level
- transfer learning
- expert systems
- learning algorithm