Abstrakt
English to Czech machine translation as it is implemented in the TectoMT system consists of three phases: analysis, transfer and synthesis. The system uses tectogrammatical (deep-syntactic dependency) trees as the transfer medium.
Each phase is divided into so-called blocks, which are processing units that solve linguistically interpretable tasks (e.g., statistical part-of-speech tagging or rule-based placement of clitics). This paper shortly introduces linguistic layers of language description which are used for the translation and describes basic concepts of the TectoMT framework.
The translation results are evaluated using both automatic metric BLEU and human judgments from the WMT 2010 evaluation.