Georgetown takes silver and gold in several categories at DISRPT 2019

Thanks to the hard work of students in LING-765, Computational Discourse Modeling, Georgetown’s entry at the DISRPT 2019 shared task was able to score first place in one category and second place in three categories, advancing the state of the art in multilingual discourse processing:

The task challenges researchers to identify the location of discourse units and phrases indicating the presence of a discourse relation, such as “on account of” (signaling a reason) or “in an apparent attempt…” (purpose), across a collection of 15 datasets in 10 languages as diverse as English, Chinese, Basque and Turkish. The research was conducted as part of a course project and supported by GPU computing resources made available by the UIS High Performance Computing team, thanks to an internal infrastructure grant from Georgetown.