Datasets:
annotations_creators:
- crowdsourced
language_creators:
- found
language:
- en
license:
- cc-by-nc-sa-3.0
multilinguality:
- monolingual
size_categories:
- 10M<n<100M
source_datasets:
- extended|natural_questions
task_categories:
- multiple-choice
task_ids:
- multiple-choice-qa
paperswithcode_id: asnq
pretty_name: Answer Sentence Natural Questions (ASNQ)
dataset_info:
features:
- name: question
dtype: string
- name: sentence
dtype: string
- name: label
dtype:
class_label:
names:
'0': neg
'1': pos
- name: sentence_in_long_answer
dtype: bool
- name: short_answer_in_sentence
dtype: bool
splits:
- name: train
num_bytes: 3656881376
num_examples: 20377568
- name: validation
num_bytes: 168005155
num_examples: 930062
download_size: 3563857920
dataset_size: 3824886531
Dataset Card for "asnq"
Table of Contents
- Dataset Description
- Dataset Structure
- Dataset Creation
- Considerations for Using the Data
- Additional Information
Dataset Description
- Homepage: https://github.com/alexa/wqa_tanda#answer-sentence-natural-questions-asnq
- Repository: More Information Needed
- Paper: TANDA: Transfer and Adapt Pre-Trained Transformer Models for Answer Sentence Selection
- Point of Contact: More Information Needed
- Size of downloaded dataset files: 3398.76 MB
- Size of the generated dataset: 3647.70 MB
- Total amount of disk used: 7046.46 MB
Dataset Summary
ASNQ is a dataset for answer sentence selection derived from Google's Natural Questions (NQ) dataset (Kwiatkowski et al. 2019).
Each example contains a question, candidate sentence, label indicating whether or not the sentence answers the question, and two additional features -- sentence_in_long_answer and short_answer_in_sentence indicating whether ot not the candidate sentence is contained in the long_answer and if the short_answer is in the candidate sentence.
For more details please see https://arxiv.org/abs/1911.04118
and
https://research.google/pubs/pub47761/
Supported Tasks and Leaderboards
Languages
Dataset Structure
Data Instances
default
- Size of downloaded dataset files: 3398.76 MB
- Size of the generated dataset: 3647.70 MB
- Total amount of disk used: 7046.46 MB
An example of 'validation' looks as follows.
{
"label": 0,
"question": "when did somewhere over the rainbow come out",
"sentence": "In films and TV shows ( edit ) In the film Third Finger , Left Hand ( 1940 ) with Myrna Loy , Melvyn Douglas , and Raymond Walburn , the tune played throughout the film in short sequences .",
"sentence_in_long_answer": false,
"short_answer_in_sentence": false
}
Data Fields
The data fields are the same among all splits.
default
question: astringfeature.sentence: astringfeature.label: a classification label, with possible values includingneg(0),pos(1).sentence_in_long_answer: aboolfeature.short_answer_in_sentence: aboolfeature.
Data Splits
| name | train | validation |
|---|---|---|
| default | 20377568 | 930062 |
Dataset Creation
Curation Rationale
Source Data
Initial Data Collection and Normalization
Who are the source language producers?
Annotations
Annotation process
Who are the annotators?
Personal and Sensitive Information
Considerations for Using the Data
Social Impact of Dataset
Discussion of Biases
Other Known Limitations
Additional Information
Dataset Curators
Licensing Information
The data is made available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License: https://github.com/alexa/wqa_tanda/blob/master/LICENSE
Citation Information
@article{Garg_2020,
title={TANDA: Transfer and Adapt Pre-Trained Transformer Models for Answer Sentence Selection},
volume={34},
ISSN={2159-5399},
url={http://dx.doi.org/10.1609/AAAI.V34I05.6282},
DOI={10.1609/aaai.v34i05.6282},
number={05},
journal={Proceedings of the AAAI Conference on Artificial Intelligence},
publisher={Association for the Advancement of Artificial Intelligence (AAAI)},
author={Garg, Siddhant and Vu, Thuy and Moschitti, Alessandro},
year={2020},
month={Apr},
pages={7780–7788}
}
Contributions
Thanks to @mkserge for adding this dataset.