CI-ADDO-EN: FrameNet 3: Upgrading FrameNet for the NLP Community

点击翻译按钮获取中文摘要

The FrameNet project has been building a lexical database containinghighly detailed information about the syntax and semantics of thedifferent senses of English words, with applications such as questionanswering and information extraction. The database is based onCharles Fillmore's theory of Frame Semantics, relating word meaningsto conceptualizations of situations and their participants, rangingfrom simple actions like Placing to complex events such as a Criminalprocess, with subevents like Arraignment and Trial. Thegeneralizations are derived from annotated example sentences fromlarge text corpora.In this planning grant, input will be gathered from researchers in theNLP field regarding how the FrameNet data is currently being used, howit can be made more useful to a wider range of users, and how it canbe enhanced so as to enable new directions of research. As part ofthis, a series of conferences will be held, both face-to-face and byvideoconference, to plan future FrameNet development and cooperationand alignment of data with other lexical resources and annotationprojects, such as WordNet, PropBank, OntoNotes, etc., with theultimate goal of improving results in sense disambiguation, semanticrole recognition, and natural language understanding in general.