Datasets:
annotations_creators:
- crowdsourced
language_creators:
- found
- crowdsourced
language:
- en
license:
- mit
multilinguality:
- monolingual
size_categories:
- 10K<n<100K
source_datasets:
- original
task_categories:
- text2text-generation
task_ids: []
paperswithcode_id: commongen
pretty_name: CommonGen
tags:
- concepts-to-text
dataset_info:
features:
- name: concept_set_idx
dtype: int32
- name: concepts
sequence: string
- name: target
dtype: string
splits:
- name: train
num_bytes: 6724166
num_examples: 67389
- name: validation
num_bytes: 408740
num_examples: 4018
- name: test
num_bytes: 77518
num_examples: 1497
download_size: 3434865
dataset_size: 7210424
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
Dataset Card for "common_gen"
Table of Contents
- Dataset Description
- Dataset Structure
- Dataset Creation
- Considerations for Using the Data
- Additional Information
Dataset Description
- Homepage: https://inklab.usc.edu/CommonGen/index.html
- Repository: https://github.com/INK-USC/CommonGen
- Paper: CommonGen: A Constrained Text Generation Challenge for Generative Commonsense Reasoning
- Point of Contact: More Information Needed
- Size of downloaded dataset files: 1.85 MB
- Size of the generated dataset: 7.21 MB
- Total amount of disk used: 9.06 MB
Dataset Summary
CommonGen is a constrained text generation task, associated with a benchmark dataset, to explicitly test machines for the ability of generative commonsense reasoning. Given a set of common concepts; the task is to generate a coherent sentence describing an everyday scenario using these concepts.
CommonGen is challenging because it inherently requires 1) relational reasoning using background commonsense knowledge, and 2) compositional generalization ability to work on unseen concept combinations. Our dataset, constructed through a combination of crowd-sourcing from AMT and existing caption corpora, consists of 30k concept-sets and 50k sentences in total.
Supported Tasks and Leaderboards
Languages
Dataset Structure
Data Instances
default
- Size of downloaded dataset files: 1.85 MB
- Size of the generated dataset: 7.21 MB
- Total amount of disk used: 9.06 MB
An example of 'train' looks as follows.
{
"concept_set_idx": 0,
"concepts": ["ski", "mountain", "skier"],
"target": "Three skiers are skiing on a snowy mountain."
}
Data Fields
The data fields are the same among all splits.
default
concept_set_idx
: aint32
feature.concepts
: alist
ofstring
features.target
: astring
feature.
Data Splits
name | train | validation | test |
---|---|---|---|
default | 67389 | 4018 | 1497 |
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 dataset is licensed under MIT License.
Citation Information
@inproceedings{lin-etal-2020-commongen,
title = "{C}ommon{G}en: A Constrained Text Generation Challenge for Generative Commonsense Reasoning",
author = "Lin, Bill Yuchen and
Zhou, Wangchunshu and
Shen, Ming and
Zhou, Pei and
Bhagavatula, Chandra and
Choi, Yejin and
Ren, Xiang",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2020.findings-emnlp.165",
doi = "10.18653/v1/2020.findings-emnlp.165",
pages = "1823--1840"
}
Contributions
Thanks to @JetRunner, @yuchenlin, @thomwolf, @lhoestq for adding this dataset.