UfLexicc – Example Sentence Extraction for Lexicographers
Affiliation: TU Darmstadt
Associated since: December 2016
Short Description
The goal of the project is to assist lexicographers in finding and categorizing sentences that can be used as good examples for the use of particular words. We employ a combination of unsupervised and supervised methods to process data from a large corpus. This approach aims to offer lexicographers a list of appropriate pre-analyzed example candidates. Among the challenges faced in the project are supporting the interactive disambiguation of example sentences according to word senses, as well as the user-driven identification of diversified contexts related to the usage of the target word in one particular sense.
The project involves the use of various state-of-the-art techniques, including clustering and classification methods for word sense induction. User feedback is incorporated into the process of shaping the word senses and selecting the best examples for each sense.
CEDIFOR Project Partners
- Dr. Richard Eckart de Castilho, UKP Lab, TU Darmstadt
- Beto Boullosa, UKP Lab, TU Darmstadt
CEDIFOR-Partners:
- Dr. Alexander Geyken, Berlin-Brandenburgische Akademie der Wissenschaften
- Dr. Lothar Lemnitzer, Berlin-Brandenburgische Akademie der Wissenschaften
Publications
- Boullosa, Beto; Eckart de Castilho, Richard; Geyken, Alexander; Lemnitzer, Lothar; Gurevych, Iryna: A tool for extracting sense-disambiguated example sentences through user feedback. In Proceedings of the Software Demonstrations of the 15th Conference of the European Chapter of the Association for Computational Linguistics, p. 69-72, Valencia, Spain, 2017.