SQuAD 2.0

The SQuAD 2.0 dataset is a new challenging task for natural language processing, which requires that machine can read, understand, and answer questions about a text. The dataset contains over 100,000+ questions on 500+ Wikipedia articles.