IJCAI 2016 Tutorial
Coreference Resolution: Successes and Challenges

Vincent Ng
The University of Texas at Dallas

Location and Time: Room Regent Parlor, July 11, 2016, 8:45am-12:45pm



Recent years have seen a surge of interest in the Winograd Schema Challenge in the AI community owing in part to Levesque's (2011) proposal of the task as an appealing alternative to the Turing Test. Being a pronoun resolution task, the Winograd Schema Challenge can be considered part of a task known as coreference resolution in the natural language processing (NLP) community. Coreference resolution involves determining which mentions in a text or dialogue refer to the same entity or event in the real world. It is an enabling technology for many high-level NLP applications, and is generally considered one of the most challenging tasks in NLP. Though extensively investigated by NLP researchers since the 1960s, coreference resolution is far from being solved. This tutorial aims to introduce this long-standing task to AI researchers. In particular, we will discuss the progress made on coreference resolution in the past two decades.


The tutorial will be composed of four parts:

Part 1: Background [45 minutes total]

Part 2: Machine Learning for Coreference Resolution [130 minutes total]

Part 3: Solving Hard Coreference Problems [25 minutes]

Part 4: Future Directions [10 minutes]

Tutorial Slides

The slides that we will use for the tutorial can be accessed from this link.