Paper: MCTest: A Challenge Dataset for the Open-Domain Machine Comprehension of Text

ACL ID D13-1020
Title MCTest: A Challenge Dataset for the Open-Domain Machine Comprehension of Text
Venue Conference on Empirical Methods in Natural Language Processing
Session Main Conference
Year 2013
Authors

We present MCTest, a freely available set of stories and associated questions intended for research on the machine comprehension of text. Previous work on machine comprehen- sion (e.g., semantic modeling) has made great strides, but primarily focuses either on lim- ited-domain datasets, or on solving a more re- stricted goal (e.g., open-domain relation extraction). In contrast, MCTest requires ma- chines to answer multiple-choice reading comprehension questions about fictional sto- ries, directly tackling the high-level goal of open-domain machine comprehension. Read- ing comprehension can test advanced abilities such as causal reasoning and understanding the world, yet, by being multiple-choice, still provide a clear metric. By being fictional, the answer typically can be found...