Most existing Neural Machine Translation models use groups of characters or whole words as their unit of input and output. We propose a model with a hierarchical char2word encoder, that takes individual characters both as input and output. We first argue that this hierarchical representation of the character encoder reduces computational complexity, and show that it improves translation performance. Secondly, by qualitatively studying attention plots from the decoder we find that the model learns to compress common words into a single embedding whereas rare words, such as names and places, are represented character by character.
@article{arxiv.1610.06550,
title = {Neural Machine Translation with Characters and Hierarchical Encoding},
author = {Alexander Rosenberg Johansen and Jonas Meinertz Hansen and Elias Khazen Obeid and Casper Kaae Sønderby and Ole Winther},
journal= {arXiv preprint arXiv:1610.06550},
year = {2016}
}