Integrating Rules and Dictionaries from Shallow-Transfer Machine Translation into Phrase-Based Statistical Machine Translation

Abstract

We describe a hybridisation strategy whose objective is to integrate linguistic resources from shallow-transfer rule-based machine translation (RBMT) into phrase-based statistical machine translation (PBSMT). It basically consists of enriching the phrase table of a PBSMT system with bilingual phrase pairs matching transfer rules and dictionary entries from a shallow-transfer RBMT system. This new strategy takes advantage of how the linguistic resources are used by the RBMT system to segment the source-language sentences to be translated, and overcomes the limitations of existing hybrid approaches that treat the RBMT systems as a black box. Experimental results confirm that our approach delivers translations of higher quality than existing ones, and that it is specially useful when the parallel corpus available for training the SMT system is small or when translating out-of-domain texts that are well covered by the RBMT dictionaries. A combination of this approach with a recently proposed unsupervised shallow-transfer rule inference algorithm results in a significantly greater translation quality than that of a baseline PBSMT; in this case, the only hand-crafted resource used are the dictionaries commonly used in RBMT. Moreover, the translation quality achieved by the hybrid system built with automatically inferred rules is similar to that obtained by those built with hand-crafted rules.

The UA-Prompsit hybrid machine translation system for the 2014 Workshop on Statistical Machine Translation

“…Linguistic data from RBMT systems have already been used to enrich SMT systems (Tyers, 2009;Schwenk et al, 2009;Eisele et al, 2008;Sánchez-Cartagena et al, 2011a). We have already proved that using hand-written rules and dictionaries from RBMT yields better results than using only dictionaries (Sánchez-Cartagena et al, 2011a).…”

Conserving Semantic Unit Information and Simplifying Syntactic Constituents to Improve Implicit Discourse Relation Recognition

“…Step 1 in our work aims at extracting the phrases in the sentence. The commonly used methods for this task are usually based on graphs [ 29 , 30 ] or probabilistics [ 31 , 32 ]. Inspiring as they are, the results of these methods are still not satisfying enough for application; accuracy and recall are problems.…”

RuLearn: an Open-source Toolkit for the Automatic Inference of Shallow-transfer Rules for Machine Translation

“…Thus, ruLearn reduces the difficulty of creating Apertiumbased RBMT systems for new language pairs. ruLearn has been successfully used in the development of Apertium-based RBMT systems for Chinese→Spanish (Costa-Jussà and Centelles, 2015) and Serbian↔Croatian (Klubička et al, 2016) The rules obtained with ruLearn can also be integrated into a phrase-based SMT system by means of the hybridisation strategy we developed (Sánchez-Cartagena et al, 2016) and released as an open-source toolkit (Sánchez-Cartagena et al, 2012). When shallow-transfer rules extracted from the same training corpus are integrated into a phrase-based SMT system, the translation knowledge contained in the parallel corpus is generalised to sequences of words that have not been observed in the corpus, thus helping to mitigate data sparseness.…”