Resolving Coreferences to Answer Questions
Dataset Description
Quoref (Q&A for Coreference Resolution)
Resolving Coreferences to Answer Questions
Source
About this dataset
This dataset, called Quoref, is a unique dataset meant to test the coreferential reasoning capability of reading comprehension systems. The dataset contains 24,000 questions over 4,700 paragraphs from Wikipedia pages. A system must resolve complex coreferences before selecting the appropriate span(s) in the paragraphs for answering questions. The data fields in this dataset are question, context, title, url, answers. This allows for systems to not only answer the questions but also provide evidence from the context to back up their answers
Research Ideas
This dataset could be used to test the coreferential reasoning capability of reading comprehension systems.
A system must resolve hard coreferences before selecting the appropriate span(s) in the paragraphs for answering questions
Acknowledgements
Original Author: AllenAI
License
> License: CC0 1.0 Universal (CC0 1.0) - Public Domain Dedication
> No Copyright - You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission. See Other Information.
Columns
File: validation.csv
| Column name | Description |
|---|---|
| question | The question text. (String) |
| context | The context paragraph(s) for the question. (String) |
| title | The title of the Wikipedia page from which the context was extracted. (String) |
| url | The URL of the Wikipedia page from which the context was extracted. (String) |
| answers | The answer span(s) for the question. (List of strings) |
File: train.csv
| Column name | Description |
|---|---|
| question | The question text. (String) |
| context | The context paragraph(s) for the question. (String) |
| title | The title of the Wikipedia page from which the context was extracted. (String) |
| url | The URL of the Wikipedia page from which the context was extracted. (String) |
| answers | The answer span(s) for the question. (List of strings) |
Related Datasets
-
WikiQA (Open-Domain Q&A)
@kaggle
-
Eucalyptus Growth And Environmental Data
@euremarkable
-
Dummy Monster
@owid
