Search

Recently, machine translation systems based on neural networks have reached state-of-the-art results for some pairs of languages (e.g., German–English). In this paper, we are investigating the performance of neural machine translation in Chinese–Spanish, which is a challenging language pair. Given that the meaning of a Chinese word can be related to its graphical representation, this work aims to enhance neural machine translation by using as input a combination of: words or characters and their corresponding bitmap fonts. The fact of performing the interpretation of every word or character as a bitmap font generates more informed vectorial representations. Best results are obtained when using words plus their bitmap fonts obtaining an improvement (over a competitive neural MT baseline system) of almost six BLEU, five METEOR points and ranked coherently better in the human evaluation., Peer Reviewed, Postprint (published version)

Recently, translation systems based on neural networks are starting to compete with systems based on phrases. The systems which are based on neural networks use vectorial repre- sentations of words. However, one of the biggest challenges that machine translation still faces, is dealing with large vocabularies and morphologically rich languages. This work aims to adapt a neural machine translation system to translate from Chinese to Spanish, using as input different types of granularity: words, characters, bitmap fonts of Chinese characters or words. The fact of performing the interpretation of every character or word as a bitmap font allows for obtaining more informed vectorial representations. Best results are obtained when using the information of the word bitmap font., Postprint (published version)