A novel approach to mapping FrameNet lexical units to WordNet synsets (short paper)

In this paper we present a novel approach to mapping FrameNet lexical units to WordNet synsets in order to automatically enrich the lexical unit set of a given frame. While the mapping approaches proposed in the past mainly rely on the semantic similarity between lexical units in a frame and lemmas in a synset, we exploit the definition of the lexical entries in FrameNet and the WordNet glosses to find the best candidate synset(s) for the mapping. Evaluation results are also reported and discussed. 1 FrameNet and the existing mapping approaches The FrameNet database [1] is a lexical resource of English describing some prototypical situations, the frames, and the frame-evoking words or expressions associated with them, the lexical units (LU). Every frame corresponds to a scenario involving a set of participants, the frame elements, that are typically the syntactic dependents of the lexical units. The FrameNet resource is corpus-based, i.e. every lexical unit should be instantiated by at least one example sentence, even if at the moment the definition and annotation step is still incomplete for several LUs. FrameNet has proved to be useful in a number of NLP tasks, from textual entailment to question answering, but its coverage is still a major problem. In order to expand the resource, it would be a good solution to acquire lexical knowledge encoded in other existing resources and import it into the FrameNet database. WordNet [4], for instance, covers the majority of nouns, verbs, adjectives and adverbs in the English language, organized in synonym sets called synsets. Mapping FrameNet LUs to WordNet synsets would automatically increase the number of LUs per frame by importing all synonyms from the mapped synset, and would allow to exploit the semantic and lexical relations in WordNet to enrich the information encoded in FrameNet.