[특강안내]

Obtaining High-Quality Linguistic Annotation for Natural Language Processing Abstract

• In this talk I will first present some general ideas of Natural Language Processing (NLP), focusing specifically on the area of Intelligent Comptuer-Assisted Language Learning (ICALL).  I will specifically illustrate this with examples from the Korean Learner Language Analysis (KoLLA) project.  The needs of ICALL applications will highlight the need for linguistic abstractions in NLP, including the need for high-quality annotated corpora.  Developing high-quality annotation is a challenge, however, and in the second half of the talk I will present research from the Detecting Anomalous Parse Structures (DAPS) project, outlining some simple (and some more complex) ways to automatically find human-generated and automatically-generated annotation errors in corpora.  The DAPS techniques seem to work well especially for small amounts of annotation, providing promise for bootstrapping larger corpora.

재학생 여러분의 많은 참여 바랍니다.

We look forward for your attendance.