Shared task at SemEval 2018: Machine Comprehension using Commonsense Knowledge

Manfred Pinkal's group is organizing a shared task at next year's SemEval, officially titled "SemEval 2018 Shared Task 11: Machine Comprehension using Commonsense Knowledge". As in other machine comprehension tasks, reading texts will be provided together with multiple choice questions. Participating systems have to pick the correct answer to each question, given the reading text. In contrast to previous machine comprehension tasks, this task will require a model to make use of commonsense knowledge: some of the questions explicitly require inference over such knowledge and cannot be answered purely from the given text.