Multilingual Neural Machine Translation With Soft Decoupled Encoding

Abstract

Multilingual training of neural machine translation (NMT) systems has led to impressive accuracy improvements on low-resource languages. However, there are still significant challenges in efficiently learning word representations in the face of paucity of data. In this paper, we propose Soft Decoupled Encoding (SDE), a multilingual lexicon encoding framework specifically designed to share lexical-level information intelligently without requiring heuristic preprocessing such as pre-segmenting the data. SDE represents a word by its spelling through a character encoding, and its semantic meaning through a latent embedding space shared by all languages. Experiments on a standard dataset of four low-resource languages show consistent improvements over strong multilingual NMT baselines, with gains of up to 2 BLEU on one of the tested languages, achieving the new state-of-the-art on all four language pairs.

accepted at ICLR 2019

References in corpus (2)

• Transfer Learning across Low-Resource, Related Languages for Neural Machine Translation
• Google's Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation

Cited by in corpus (8)

• Balancing Training for Multilingual Neural Machine Translation
• MultiFiT: Efficient Multi-lingual Language Model Fine-tuning
• Optimizing Data Usage via Differentiable Rewards
• A Brief Survey of Multilingual Neural Machine Translation
• Adversarial Subword Regularization for Robust Neural Machine Translation
• ABSent: Cross-Lingual Sentence Representation Mapping with Bidirectional GANs
• Char2Subword: Extending the Subword Embedding Space Using Robust Character Compositionality
• Improving Target-side Lexical Transfer in Multilingual Neural Machine Translation