Article Structure
Abstract
We present a corpus analysis of how Italian connectives are translated into LIS, the Italian Sign Language.
Introduction
Automatic translation between a spoken language and a signed language gives rise to some of the same difficulties as translation between spoken languages, but adds unique challenges of its own.
Corpus Analysis
The corpus consists of 40 weather forecasts in Italian and LIS.
The effect of the Italian connectives on the LIS translation
From this basic frequency analysis we can already notice that a large number of connectives only appear in Italian, or have far more occurrences in Italian than in LIS.
Rule extraction
We trained a classifier to help a LIS generator determine how an Italian connective should be translated.
Conclusions and Future Work
The small size of our corpus, with around 375 bilingual sentences, posed a large challenge to the use of statistical methods; on the other hand, having no access to a LIS speaker prevented us from simply relying on a rule-based approach.
Topics
subtrees
- In effect, since we have hypothesized that the presence of a connective can affect the translation of the two subtrees that it connects, we would like to be able to align each of those subtrees to its translation.Page 6, “The effect of the Italian connectives on the LIS translation”
- We make the observation that, if two words belong to two different subtrees linked by a connective, so that the path between the two words goes through the connective, then the frontier between the LIS counterparts of those two subtrees should also lie along the path between the signs aligned with those two words.Page 6, “The effect of the Italian connectives on the LIS translation”
- Then, each pair of words belonging to different subtrees is linked by a path that goes through the connective in the original tree.Page 6, “The effect of the Italian connectives on the LIS translation”
- Of these words, we select the ones that have aligned signs, and then we compute the path between each pair of signs aligned to words belonging to different subtrees .Page 6, “The effect of the Italian connectives on the LIS translation”
- o This node connects two subtrees : a child subtree containing “qualche, breve, scroscio, di, pioggia”, and a parent subtree containing the rest of the sentence.Page 6, “The effect of the Italian connectives on the LIS translation”
- words are connected by a connective, and they map to signs which have a very short path between them (up to 3 nodes, including the two signs), the connective may be reflected simply in this close connection between the translated subtrees in the LIS syntax tree;Page 7, “The effect of the Italian connectives on the LIS translation”
See all papers in Proc. ACL 2013 that mention subtrees.
See all papers in Proc. ACL that mention subtrees.
Back to top.
rule-based
- Still, it is a very small corpus, hence the main project shied away from statistical NLP techniques, relying instead on rule-based approaches developed with the help of a native ItaliarflLIS bilingual speaker; a similar approach is taken e.g.Page 2, “Introduction”
- Because the translation pipeline we plan to integrate with is rule-based , we chose a Decision Tree as our classifier: this allows rules to be easily extracted from the classification model.Page 8, “Rule extraction”
- The small size of our corpus, with around 375 bilingual sentences, posed a large challenge to the use of statistical methods; on the other hand, having no access to a LIS speaker prevented us from simply relying on a rule-based approach.Page 9, “Conclusions and Future Work”
See all papers in Proc. ACL 2013 that mention rule-based.
See all papers in Proc. ACL that mention rule-based.
Back to top.
translation system
- The resulting lack of a shared written form does nothing to improve the availability of sign language corpora; bilingual corpora, which are of particular importance to a translation system , are especially rare.Page 1, “Introduction”
- Tree alignment in a variety of forms has been extensively used in machine translation systems (Gildea, 2003; Eisner, 2003; May and Knight, 2007).Page 5, “The effect of the Italian connectives on the LIS translation”
- We have produced initial results in terms of rule extraction, and we will be integrating these rules into the full Italian-LIS translation system to produce improved translation of connec-t1ves.Page 9, “Conclusions and Future Work”
See all papers in Proc. ACL 2013 that mention translation system.
See all papers in Proc. ACL that mention translation system.
Back to top.
Word Aligner
- Word Alignment .Page 6, “The effect of the Italian connectives on the LIS translation”
- For this purpose we used the Berkeley Word Aligner (BWA) 1 (Denero, 2007), a general tool for aligning sentences in bilingual corpora.Page 6, “The effect of the Italian connectives on the LIS translation”
- Bold dashed lines show word alignment .Page 6, “The effect of the Italian connectives on the LIS translation”
See all papers in Proc. ACL 2013 that mention Word Aligner.
See all papers in Proc. ACL that mention Word Aligner.
Back to top.
