The Negative Effect of Machine Translation on Cross–Lingual Question Answering

Abstract

Abstract. This paper presents a study of the negative effect of Machine Translation (MT) on the precision of Cross-Lingual Question Answering (CL-QA). For this research, a English-Spanish Question Answering (QA) system is used. Also, the sets of 200 official questions from CLEF 2004 and 2006 are used. The CL experimental evaluation using MT reveals that the precision of the system drops around 30% with regard to the monolingual Spanish task. Our main contribution consists on a taxonomy of the identified errors caused by using MT and how the errors can be overcome by using our proposals. An experimental evaluation proves that our approach performs better than MT tools, at the same time contributing to this CL-QA system being ranked first at English-Spanish QA CLEF 2006.

Exploiting Wikipedia and EuroWordNet to solve Cross-Lingual Question Answering

“…The impact of the errors caused by the use of MT services has been studied, evaluated, controlled and surpassed. These problems are analyzed in our paper [16]. This recent research presents a study detailing the common errors produced by MT based systems and in this research work we propose an alternative approach (without any MT developing, unlike the current CL-QA system) to overcome such errors.…”

“…The problems caused by the use of MT services are analyzed in our paper [16]. This recent research presents a study detailing the common errors produced by MT based systems and proposes an alternative approach to overcome such errors.…”

“…15 The exact answers are assessed as: (1) Right: if correct; (2) Wrong: if incorrect; (3) Inexact: if contained less or more information than that required by the query; or (4) Unsupported: the supporting snippet did not contain the exact answer. 16 http://www.google.com/intl/en/press/zeitgeist200X.html. 17 Some English-Spanish examples: ''South Africa" () ''Sudáfrica"; ''Channel Tunnel" () ''Eurotúnel"; ''The Little Prince" () ''El Principito".…”

Applying Wikipedia’s Multilingual Knowledge to Cross–Lingual Question Answering

“…The last edition of CLEF (2006) [15] has confirmed that most of the implementations of current CL-QA systems [5,13,18,20,21]are based on the use of on-line translation services. However, a recent research [8] presents a study detailing the common errors produced by Machine Translation (MT) based systems and proposes an alternative approach to overcome such errors.…”

“…For instance, MT systems generate errors [8] such as translations of names that should be left untranslated, translations of polysemous words where the sense translated is not the correct one, syntactic errors in the translation, wrong translations of interrogative particles, incorrect lexical-syntactic category of the translated words and unknown words by the MT and thus left untranslated. The impact of this kind of mistakes should be controlled and evaluated.…”