Integrating Collaborative and Linguistic Resources for Word Sense Disambiguation and Semantic Role Labeling
Lexical semantic resources (LSRs) are a basic building block for many natural language processing technologies such as word sense disambiguation or semantic role labeling.
These technologies often do not scale to real-life applications due to insufficient size and quality of the resources.
In this project, we integrate various expert-built (WordNet, FrameNet, VerbNet) and collaboratively created (Wikipedia, Wiktionary, OmegaWiki) LSRs in order to alleviate the coverage problems of the single resources.
Merging them will yield a novel resource of unprecedented coverage and quality – called UBY. It provides semantic operability between the integrated resources and offers uniform access to complementary information from the single resources, e.g., word senses, their definitions, domain labels, and example sentences, which is highly valuable for various natural language processing tasks.
We will explore the benefits of UBY in the tasks of word sense disambiguation and semantic role labeling.
- Prof. Iryna Gurevych, Principal Investigator
- Dr. Kostadin Cholakov, Postdoctoral Researcher
- Silvana Hartmann, Doctoral Researcher
- Tristan Miller, Doctoral Researcher