QA-It: Non-referential It for Question Answering

This project provides the QA-It corpus that can be used for the classification of non-referential it. Our dataset is unique; it is annotated on question answer pairs collected from multiple genres, which is useful for developing advanced question answering systems. Our annotation scheme makes clear distinctions between 4 types of pronominal it, and provides guidelines for many erroneous cases.

• QA-It: Classifying Non-Referential It for Question Answer Pairs, Timothy Lee, Alex Lutz, Jinho D. Choi, In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop, ACL'16, Berlin, Germany, 2016 (poster).
• Corpus: v1.0 (last updated: 10/20/2016).

Format

Our corpus is in the TSV format, where each column contains the following contents:

• 0: the source the data.
• 1: the genre.
• 2: the document size (token count).
• 3: the total count of pronominal it.
• 4: a question answer pair, where the question and the answer are delimited by >>>>> and all pronominal its are surrounded by double square brackets (e.g., [[it]]).
• 5: the classes of the pronominal its in the question answer pair, where each class is delimited by , (e.g., 1,2 implies that the classes of the first and the second pronominal its are 1 and 2, respectively).

The followings describe the meaning of each class:

• 1: Non-referential (pleonastic)
• 2: Referential - nominal
• 3: Referential - others
• 4: Error