Vocabulary Selection Strategies for Neural Machine Translation

Abstract

Classical translation models constrain the space of possible outputs by selecting a subset of translation rules based on the input sentence. Recent work on improving the efficiency of neural translation models adopted a similar strategy by restricting the output vocabulary to a subset of likely candidates given the source. In this paper we experiment with context and embedding-based selection methods and extend previous work by examining speed and accuracy trade-offs in more detail. We show that decoding time on CPUs can be reduced by up to 90% and training time by 25% on the WMT15 English-German and WMT16 English-Romanian tasks at the same or only negligible change in accuracy. This brings the time to decode with a state of the art neural translation system to just over 140 msec per sentence on a single CPU core for English-German.

References in corpus (3)

• Effective Approaches to Attention-based Neural Machine Translation
• BilBOWA: Fast Bilingual Distributed Representations without Word Alignments
• Efficient softmax approximation for GPUs

Cited by in corpus (14)

• Convolutional Sequence to Sequence Learning
• Learning to Compute Word Embeddings On the Fly
• A Convolutional Encoder Model for Neural Machine Translation
• Cross-Lingual Transfer Learning for Multilingual Task Oriented Dialog
• Neural Machine Translation with Word Predictions
• Attention-based Vocabulary Selection for NMT Decoding
• Neural Machine Translation via Binary Code Prediction
• Neural Machine Translation: A Review and Survey
• Cue-word Driven Neural Response Generation with a Shrinking Vocabulary
• HSCJN: A Holistic Semantic Constraint Joint Network for Diverse Response Generation
• Grammatical Sequence Prediction for Real-Time Neural Semantic Parsing
• Open Vocabulary Learning for Neural Chinese Pinyin IME
• An Empirical Study of Adequate Vision Span for Attention-Based Neural Machine Translation
• Single-Queue Decoding for Neural Machine Translation