NLIP Group Part II Project Proposals
Project Proposals by Tamara Polajnar (and Stephen Clark)
Multiplatform Game for Eliciting Semantic Data
We would like to design a game that will gather data for particular linguistic semantic experiments. The game incentives should be aligned with the main goals of the game: to provide accurate data and to be enjoyable to play in order to lead to wide-spread participation. The game will gather data that will help us assess similarity between words, phrases, and sentences. It should be accessible from both web browsers and mobile devices.
Project Proposal by Stephen Clark
Optimising a Natural Language Parser in Java
I have recently developed a Java version of the well-known C&C natural language (NL) parser (http://svn.ask.it.usyd.edu.au/trac/candc). The Java version is as accurate as the C++ version, but less optimised and not as fast. The goal of the project is to improve the software engineering aspects of the Java NL parser, making it more efficient to train and run, and easier for other NL parser developers to use. As a secondary goal, there may also be opportunities to improve the accuracy of the NL parser.
Feel free to contact Stephen Clark (sc609) if you are interested in this project.