MT Technology Tools

Sunda offers specialized language technology tools for customers wishing to build their own natural language processors.

1. Machine Translation Tool

This tool is specifically designed for transforming natural language sentences into another natural language. Sunda's English-Finnish translator has been built using this tool. If your goal is to build an MT system that translates English into another language, the tool has a built-in English analysis component for it.

2. Natural Language Transformer

This tool can be used to produce programs that transform natural language sentences into XML expressions or other formal operations. The tool employs well-defined linguistic rules and comprises several linguistically motivated phases. Possible applications are, for example, speech controlled systems – if speech recognition component is available. The tool also contains built-in linguistic rules for analyzing English sentences.

The Development Environment

The above-mentioned tools are based on Sunda's linguistic development environment. The development environment offers a flexible way for a linguistic developer to write new rules or modify the existing ones and to study the effects of the modifications. It tells the developer which rules have fired and what kind of effect they have had on a given sentence.

The development environment consists of a Windows application that offers a set of tools in an integrated graphical user interface. It contains the following main components:

  • Rule Editor. A tool for browsing and editing structural and lexical rules. » see image
  • Lexicon Editor. A tool for editing basic lexicons, special field lexicons and frozen phrases (idioms and terms). » see image
  • Trace. Traces the translation process by showing which rules and operations have been executed in different translation phases. It also shows the generated parse trees. » see image
  • Rule Compiler. Compiles the text-based rules into binary format compatible with translation engine.
  • Reference Run. The reference run shows possible differences in the translation results of the reference sentences, thus allowing the developer to detect unfavourable consequences of rule modifications.
  • CVS Integration. The development software can be connected to a CVS server allowing a flexible versioning and sharing of the rule files between the development team.

If you are interested in developing new machine translators or natural language transformers in joint ventures, please contact us.