An Overview of the 2019 Language and Intelligence Challenge

Quan Wang, Wenquan Wu, Yabing Shi, Hongyu Li, Zhen Guo, Wei He, Hongyu Liu, Ying Chen, Yajuan Lyu, Hua Wu

Abstract: This paper provides an overview of the 2019 Language and Intelligence Challenge (LIC 2019), which assesses the ability of machines to understand language and use language to interact with humans. The challenge comprised three tasks: information extraction (IE), knowledge-driven dialogue, and machine reading comprehension (MRC), all providing large-scale Chinese datasets and open-source baseline systems. There were 2,376 teams that took part in the challenge, with a total of 6,212 system runs submitted. The participating systems performed quite well, offering a 21.65% increase over the baseline in IE, a 37.40% increase in the dialogue task, and a 34.09% increase in MRC.