Neural machine translation (NMT) takes deterministic sequences for source representations. However, either word-level or subword-level segmentations have multiple choices to split a source sequence with different word segmentors or different subword vocabulary sizes. We hypothesize that the diversity in segmentations may affect the NMT performance. To integrate different segmentations with the state-of-the-art NMT model, Transformer, we propose lattice-based encoders to explore effective word or subword representation in an automatic way during training. We propose two methods: 1) lattice positional encoding and 2) lattice-aware self-attention. These two methods can be used together and show complementary to each other to further improve translation performance. Experiment results show superiorities of lattice-based encoders in word-level and subword-level representations over conventional Transformer encoder.
@article{arxiv.1906.01282,
title = {Lattice-Based Transformer Encoder for Neural Machine Translation},
author = {Fengshun Xiao and Jiangtong Li and Hai Zhao and Rui Wang and Kehai Chen},
journal= {arXiv preprint arXiv:1906.01282},
year = {2019}
}