539 research outputs found
Sort by
Jak se Mojžíš s Jozuem učili hindsky (aneb souboj s mizernými daty při překladu z angličtiny do hindštiny)
No full textA Contrastive Lexical description of Basic Verbs. Examples from Swedish and Czech.
Full text linkThis paper aims at a lexical description of frequent, but not enough cognitively salient uses of frequent lexical verbs in Swedish on the background of Czech, with some implications for the lexical description of basic verbs in general. It results in a draft of a production lexicon of Swedish basic verbs for advanced Czech learners of Swedish, with focus on their uses as light verbs
Statistical Machine Translation between Related and Unrelated Languages
No full textIn this paper we describe an attempt to compare how relatedness of languages
can influence the performance of statistical machine translation (SMT). We
apply the Moses toolkit on the Czech-English-Russian corpus UMC 0.1 in order to
train two translation systems: Russian-Czech and English-Czech. The quality
of the translation is evaluated on an independent test set of 1000 sentences
parallel in all three languages using an automatic metric (BLEU score) as well
as manual judgments. We examine whether the quality of Russian-Czech is better
thanks to the relatedness of the languages and similar characteristics of word
order and morphological richness. Additionally, we present and discuss
the most frequent translation errors for both language pairs
Computer Aided Translation Backed by Machine Translation
No full textA number of tools to support translators (computer-aided translation, CAT)
exist, as there are many systems of machine translation (MT). So far, the
integration of the two system types was little or none. The aim of this paper is
to examine a tighter coupling of MT and CAT.
We introduce our web-based CAT tool implemented using the modern AJAX technology
that
communicates with Moses MT system on the server side
to provide the translator with suggested translations of individual phrases of
the source sentence as well as several options of the complete continuation of
the output sentence. The suggested continuation is based on what has been
already translated and what the user has already written as the output.
Hopefully, the proposed user interface and the MT system at the back end will
accelerate and simplify the process of translation
CzEng 0.9, Building a Large Czech-English Automatic Parallel Treebank
No full textWe describe our ongoing efforts in collecting a Czech-English parallel corpus CzEng.
The paper provides full details on the current
version~0.9 and focuses on its new features: (1) data from new sources were added, most importantly a few hundred electronically available books, technical documentation and also some parallel web pages, (2) the full corpus has been automatically annotated up to the tectogrammatical layer (surface and deep syntactic analysis), (3) sentence segmentation has been refined, and (4) several heuristic filters to improve corpus quality were implemented. In total, we provide a sentence-aligned automatic parallel treebank of 8.0 million sentences, 93 English and 82 Czech words. CzEng~0.9 is freely available for non-commercial research purposes
Získávání paralelních textů z webu
No full textWe examine methods for collecting parallel Czech-English corpora from the web. We propose and evaluate automatic methods for finding source web sites, language identification and most importantly the document alignment of obtained pages
Annotation Quality Checking and Its Implications for Design of Treebank (in Building the Prague Czech-English Dependency Treebank)
No full textThe article presents the system for annotation quality checking, proposed and used during the building of the Czech part of the Prague Czech-English Dependency Treebank. At first, the treebank project is introduced, as well as
its basic principles and annotation process. The second part of the article pursues in detail one of the important phases of the annotation process,
namely how the correctness of the annotated data is automatically and continuously checked during the process. The system of annotation quality
checking is demonstrated on several particular checking procedures concerning syntactical phenomena. We try to evaluate the contribution of the system not only to the quality of the data and annotation, but also to the corpus design,
impact on annotation rules and the annotation process as a whole
Tectogrammatical Annotation of the Wall Street Journal
No full textThis paper gives an overview of the current state of the Prague English Dependency Treebank project. It is an updated version of a draft text that was released along with a CD presenting the first 25\% of the PDT-like version of the Penn Treebank -- WSJ section (PEDT 1.0)
Towards English-Czech Parallel Valency Lexicon via Treebank Examples
No full textThe paper describes an ongoing project of building a bilingual valency
lexicon in the framework of Functional Generative Description. The bilingual
lexicon is designed as a result of interlinking frames and frame elements of
two already existing valency lexicons.
First, we give an overall account of the character of the lexicons to be linked,
second, the process of frame linking is explained, and third, a case
study is presented to exemplify what the information contained in frame
links tells us about crosslinguistic differences in general and the linguistic
theory applied