tau/mrqa

An effective question answering system should do more than merely interpolate from the training set to answer test examples drawn from the same distribution: it should also be able to extrapolate to out-of-distribution examples — a significantly harder challenge. The dataset is a collection of 18 existing QA dataset (carefully selected subset of them) and converted to the same format (SQuAD format).

Among these 18 datasets, six datasets were made available for training, six datasets were made available for development, and the final six for testing. The dataset is released as part of the MRQA 2019 Shared Task.