Project Description
This project aims to provide a graphical platform for meaning analysis experiments, on the web and as a stand-alone tool, for both undergraduate and graduate instruction. The purpose of the platform is to give students hands-on experience with machine learning techniques, which are central to many tasks in computational linguistics, and to illustrate a core area of computational linguistics: lexical semantics. Meaning analyses like this, providing word sense and predicate-argument structure (identifying events and their participants in text), support natural language processing applications like advanced search, or information extraction from large document collections. The project will extend an existing meaning analysis system to a teaching-friendly platform with an easy to use, intuitive interface, with sample classifiers for both English and German. The platform will enable a variety of uses, from a safe experimentation platform for undergraduate homeworks up to challenging course project for graduate students.
You can, of course, still get the command-line version of Shalmaneser from the SALSA Project.
This project is supported by a grant from UT Austin's Liberal Arts Instructional Technology Services.
| PI | Katrin Erk | ||
|---|---|---|---|
| RAs | Trevor Fountain, | Andrew Young, | Dan Velleman. |
Web interface to Shalmaneser
Would you like to try out the Shalmaneser shallow semantic analysis? Use the Web interface to Shalmaneser..
The Shalmaneser tutorial
Topics of the Shalmaneser tutorial
- Word sense disambiguation and semantic role labeling
- Frame semantics
- Automatic semantic analysis with the Shalmaneser GUI
- Feature engineering with the Feature Description Language