Can We Translate Letters?

By David Vilar, Jan-Thorsten Peter, Hermann Ney. 2007

96 citations recorded by the source corpus.

Current statistical machine translation systems handle the translation process as the transformation of a string of symbols into another string of symbols. Normally the symbols dealt with are the words in different languages, sometimes with some additional information included, like morphological data. In this work we try to push the approach to the limit, working not on the level of words, but treating both the source and target sentences as a string of letters. We try to find out if a nearly unmodified state-of-the-art translation system is able to cope with the problem and whether it is capable to further generalize translation rules, for example at the level of word suffixes and translation of unseen words. Experiments are carried out for the translation of Catalan to Spanish.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

Neural Machine Translation of Rare Words with Subword Units (2015, methodology): “ transliteration is required (Durrani et al., 2014). Character-based translation has also been investigated with phrase-based models, which proved especially successful for closely related languages (Vilar et al., 2007; Tiedemann, 2009; Neubig et al., 2012). The segmentation of morphologically compl”
Who Needs Words? Lexicon-Free Speech Recognition (2019, background): “Character-based models permeated text classification [1], language modeling [2, 3, 4], machine translation [5, 6, 7, 8, 9, 10], and automatic speech recognition (ASR) [11, 12, 13].”
Why Catalan-Spanish Neural Machine Translation? Analysis, comparison and combination with standard Rule and Phrase-based technologies (2017, background): “Best results for the out-of-domain corpus vary depending on the translation direction and measure: for Catalan-to-Spanish, best results are obtained with the phrase-based system; and for Spanish-to-Catalan, best re-sults are obtained with the rule-based system in terms of BLEU, but with the phrase-b”
Orthographic Syllable as basic unit for SMT between Related Languages (2016, methodology): “A different paradigm is to drop the notion of word boundary and consider the character n-gram as the basic unit of translation (Vilar et al., 2007; Tiedemann, 2009a).”
Character-based Neural Machine Translation (2015, methodology): “However, unlike previous approaches that attempt to discard the notion of words completely (Vilar et al., 2007; Neubig et al., 2013), we propose an hierarhical architecture, which replaces the word lookup tables (steps 1 and 3) and the word softmax (step 6) with character-based alternatives, which…”
Character-Based PSMT for Closely Related Languages (2009, methodology): “However, we still loose a lot of training data and, therefore, we also apply the two-step procedure as proposed by (Vilar et al., 2007) to compare our results with that approach.”
Combining Word-Level and Character-Level Models for Machine Translation Between Closely-Related Languages (2012, methodology): “The language similarities and the regularities in morphological variation and spelling motivate the use of character-level translation models, which were applied to translation (Vilar et al., 2007; Tiede-mann, 2009a) and transliteration (Matthews, 2007).”
Character-Based Pivot Translation for Under-Resourced Languages and Domains (2012, unknown): “Thereafter, MERT can run as usual.”
Similar Southeast Asian Languages: Corpus-Based Case Study on Thai-Laotian and Malay-Indonesian (2016, background): “In engineering practice, however, statistical metrics, i.e., word alignment precision, or word order differences, are more addressed in NLP applications.”
Machine Translation without Words through Substring Alignment (2012, methodology): “However, while the approach is attractive conceptually, previous research has only been shown effective for closely related language pairs (Vilar et al., 2007; Tiedemann, 2009; Sornlertlamvanich et al., 2008).”