Bilingual FrameNet Dictionaries for Machine Translation

The goals of GFN are to create lexical entries for German nouns, verbs, and adjectives that correspond to existing FrameNet entries, and to link the parallel lexicon fragments by means of common semantic frames and numerical indexing mechanisms.

Abstract

This paper describes issues surrounding the planning and design of GermanFrameNet (GFN), a counterpart to the English-based FrameNet project. The goals of GFN are (a) to create lexical entries for German nouns, verbs, and adjectives that correspond to existing FrameNet entries, and (b) to link the parallel lexicon fragments by means of common semantic frames and numerical indexing mechanisms. GFN will take a fine-grained approach towards polysemy that seeks to split word senses based on the semantic frames that underlie their analysis. The parallel lexicon fragments represent an important step towards capturing valuable information about the different syntactic realizations of frame semantic concepts across languages, which is relevant for information retrieval, machine translation, and language generation. 1.