Introduction

Prague Czech-English Dependency Treebank 1.0 (PCEDT) was produced by Linguistic Data Consortium (LDC) catalog number LDC2004T25 and ISBN 1-58563-321-6. This corpus was developed at the Center for Computational Linguistics in cooperation with the Institute of Formal and Applied Linguistics.

PCEDT 1.0 is a corpus of Czech-English parallel resources suitable for experiments in machine translation, with a special emphasis on dependency-based (structural) translation (with evaluation data provided for Czech-to-English systems).

Data

The core part of PCEDT 1.0 is a Czech translation of 21,600 English sentences from the Wall Street Journal, which are part of the Penn Treebank corpus. Sentences of the Czech translation were automatically morphologically annotated and parsed into two levels (analytical and tectogrammatical) of dependency structures introduced in the theory of Functional Generative Description and closely related to the Prague Dependency Treebank project. The original English sentences were transformed from the Penn Treebank phrase-structure trees into dependency representations. A heldout (development and evaluation) set of 515 sentence pairs was selected and manually annotated on tectogrammatical level in both Czech and English; for the purposes of quantitative evaluation, this set has been retranslated from Czech into English by four different translation companies.

PCEDT 1.0 also contains a parallel Czech-English corpus of plain text from Reader's Digest 1993-1996 consisting of 53,000 parallel sentences, and a large monolingual corpus of Czech (2.4 M sentences). The included Czech-English translation dictionary consists of 46,150 translation pairs in its lemmatized version and 496,673 pairs of word forms, where for each entry-translation pair all corresponding word form pairs have been generated. Also included is an English-Czech dictionary provided by Milan Svoboda under GNU/FDL license; this dictionary contains multi-word translations in 115,929 translation pairs.

The next version of PCEDT intends to translate the whole Wall Street Journal part of the Penn Treebank, and to include reference retranslations for Czech. As the manual for tectogrammatical annotation of English gets created, the proportion of manually annotated data will increase.

Sponsorship

• Ministry of Education of the Czech Republic project No. LN00A063 (Center for Computational Linguistics)
• National Science Foundation grant No. IIS-0121285

Updates

Content Copyright

Portions © 2004 Trustees of the University of Pennsylvania, © 1988-1989 Wall Street Journal, © 1993-1996 Reader's Digest, © 1991-1995 Lidové noviny, © 2004 Milan Svoboda, © 2002-2004 Center for Computational Linguistics, Charles University in Prague