In addition to creating FrameNet-style lexicons for other languages, several projects have been building frame semantic databases in specialized domains. Here are some of the important examples:
Soccer FrameNet: kicktionary.de Redesigned and Publicly Accessible!
FrameNet visitor Thomas Schmidt from Germany created and launched the Kicktionary, a domain-specific trilingual (English, German, and French) lexical resource of the language of soccer. Kictionary is based on Frame Semantics and uses WordNet style semantic relations as an additional layer of structure. The lexicon currently contains around 2,000 lexical units organized in 104 frames and 16 scenarios. Each LU is illustrated by a number of examples from a multilingual corpus of soccer match reports.
Check it out for the latest additions and updates (password or permission no longer required!).
For further information contact email@example.com.
Suggested Upper-Merged Ontology (SUMO)
The Suggested Upper-Merged Ontology (SUMO) is being created by the IEEE Standard Upper Ontology Working Group with the goal of developing an upper ontology that will promote data interoperability, information search and retrieval, automated inferencing, and natural language processing. The links from FN to SUMO are the work of Jan Scheffczyk, FrameNet visitor from 2005-2006.
Deductive reasoning with natural language requires combining lexical resources with the world knowledge provided by ontologies. Jan's work aligns FrameNet Semantic Types (ST) with SUMO classes, which are expressed in SUO-KIF, the language of SUMO. Based on this general-domain alignment, Jan has developed a semi-automatic, domain-specific approach for linking FrameNet Frame Elements (FE) to SUMO classes, which is based on typical fillers of FEs in a particular domain. This provides restricted, ontology-based types on the fillers of FEs. This basic work can improve semantic parsing and ontology lexicalization.
BioOntoFN at Linköpings Universitet
This project aims to build a corpus with frame semantic annotations for the domain of biomedicine. We have proposed a novel method of building this kind of corpus using domain knowledge provided by ontologies. We have successfully built a corpus for transport events strictly following the domain knowledge provided by the Gene Ontology (GO) biological process ontology. Ontological domain knowledge eases all the tasks in the corpus construction and leads to well-defined semantics exposed on the corpus, which will be very valuable in text mining applications. More details of the project and corpus data are available at http://www.ida.liu.se/~hetan/bio-onto-frame-corpus.
For futher information, please contact the project leader, He Tan (firstname.lastname@example.org)