Multi-Layer Softmaxing during Training Neural Machine Translation for Flexible Decoding with Fewer Layers.
Raj DabreAtsushi FujitaPublished in: CoRR (2019)
Keyphrases
- multi layer
- machine translation
- feed forward neural networks
- error back propagation
- neural network
- finite state transducers
- cross lingual
- language independent
- single layer
- natural language processing
- information extraction
- word alignment
- neural nets
- multiple layers
- cross language information retrieval
- training set
- statistical machine translation
- brazilian portuguese
- target language
- machine translation system
- natural language
- query translation
- language resources
- supervised learning
- word sense disambiguation
- word level
- source language
- chinese english
- data analysis
- training data
- multilingual documents
- data mining