Hiemstra, D. 1998. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In Peter Arno Coppen, Hans van Halteren, and Lisanne Teunissen, editors, Proceedings of the 8th CLIN meeting, pages 41--58.  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Disambiguation Strategies for Cross-language Information.. - Djoerd Hiemstra And (1999) (1 citation) Self-citation (Hiemstra) (Correct)
....available in parallel corpora on sustainable development. Translation pairs can be derived from parallel corpora using statistical co occurrence by so called word alignment algorithms. Within the Twenty One project word alignment algorithms were developed that do the job in a fast and reliable way [6]. Domain speci c translation lexicons were derived from Agenda 21, a UN document on sustainable development that is available in most of the European languages including Dutch and English. For the experiment we merged the automatically derived dictionary with the Van Dale dictionary in the ....
D. Hiemstra. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In P.A. Coppen, H. van Halteren, and L. Teunissen, editors, Proceedings of eightth CLIN meeting, pages 41-58, 1998.
Extracting Bimodal Representations for Language-Based.. - Westerveld, Hiemstra, .. (2000) Self-citation (Hiemstra) (Correct)
.... from a Parallel Corpus If we have a parallel corpus of documents in the modalities text and images (i.e. a set of pictures with reasonably complete descriptions) we could extract a bimodal translation lexicon from this corpus using techniques developed for cross language information retrieval [9]. Such a translation lexicon can be extracted by looking at the number of times language elements and image elements co occur. If this number is relatively high, the elements are probably related. For example, if image descriptions contain the word sun and the corresponding images often contain a ....
....0.33 0.33 brown rectangle 0.75 0.25 1.0 blue rectangle 0.25 0.58 0.33 0.33 0.5 green rectangle 0.25 0.58 0.33 0.33 0. 5 2 The information elements in the captions are in bold face The following bimodal dictionary can be inferred automatically by applying the EM algorithm as described in [9]. The algorithm eliminates certain word imagefeature combinations and infers for example that trees basically consist of green circles and brown rectangles. tree park sun day (null) green circle 1.0 1.0 yellow circle 0.5 0.5 brown rectangle 1.0 1.0 blue rectangle 1.0 0.25 ....
Hiemstra, D., Multilingual domain modeling in Twenty-One: automatic creation of a bidirectional translation lexicon from a parallel corpus, In: Peter-Arno Coppen, Hans van Halteren and Lisanne Teunissen (eds.), Proceedings of the eighth CLIN meeting, pages 41-58, 1998.
Disambiguation Strategies for Cross-language Information.. - Hiemstra, de Jong (1999) (1 citation) Self-citation (Hiemstra) (Correct)
....available in parallel corpora on sustainable development. Translation pairs can be derived from parallel corpora using statistical co occurrence by so called word alignment algorithms. Within the Twenty One project word alignment algorithms were developed that do the job in a fast and reliable way [6]. Domain specific translation lexicons were derived from Agenda 21, a UN document on sustainable development that is available in most of the European languages including Dutch and English. For the experiment we merged the automatically derived dictionary with the Van Dale dictionary in the ....
D. Hiemstra. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In P.A. Coppen, H. van Halteren, and L. Teunissen, editors, Proceedings of eightth CLIN meeting, pages 41--58, 1998.
Cross-language information retrieval in Twenty-One: Using.. - Hiemstra, de Jong Self-citation (Hiemstra) (Correct)
....in parallel corpora on sustainable development. Translation pairs can be derived from parallel corpora using statistical co occurrence by so called word alignment algorithms. Within the Twenty One project word alignment algorithms were developed that do the job in a fast and reliable way [3, 4]. Domain specific translation lexicons were derived from Agenda 21, a UN document on sustainable development that is available in most of the European languages including Dutch and English. For the experiment we merged the automatically derived dictionary with the Van Dale dictionary in the ....
D. Hiemstra. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In Proceedings of eightth CLIN meeting, 1998.
Evaluating the LIHLA lexical aligner on Spanish.. - Caseli, Nunes, Forcada (2005) (Correct)
No context found.
Hiemstra, D. 1998. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In Peter Arno Coppen, Hans van Halteren, and Lisanne Teunissen, editors, Proceedings of the 8th CLIN meeting, pages 41--58.
LIHLA: Shared task system description - Caseli, Nunes, al. (2005) (Correct)
No context found.
Djoerd Hiemstra. 1998. Multilingual domain modeling in Twenty-One: automatic creation of a bi-directional translation lexicon from a parallel corpus. In Peter Arno Coppen, Hans van Halteren, and Lisanne Teunissen, editors, Proceedings of the 8th CLIN meeting, pages 41--58.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC