mctaco

• Description:

MC-TACO is a dataset of 13k question-answer pairs that require temporal commonsense comprehension. The dataset contains five temporal properties:

1. duration (how long an event takes)
2. temporal ordering (typical order of events)
3. typical time (when an event occurs)
4. frequency (how often an event occurs)
5. stationarity (whether a state is maintained for a very long time or indefinitely)

We hope that this dataset can promote the future exploration of this particular class of reasoning problems.

• Homepage: https://github.com/CogComp/MCTACO

• Source code: tfds.question_answering.Mctaco

• Versions:

  • 1.0.0 (default): No release notes.

• Download size: 2.27 MiB

• Dataset size: 3.18 MiB

• Auto-cached (documentation): Yes

• Splits:

Split	Examples
'test'	9,442
'validation'	3,783

• Features:

FeaturesDict({
    'answer': Text(shape=(), dtype=tf.string),
    'category': ClassLabel(shape=(), dtype=tf.int64, num_classes=5),
    'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=2),
    'question': Text(shape=(), dtype=tf.string),
    'sentence': Text(shape=(), dtype=tf.string),
})

• Supervised keys (See as_supervised doc): None

• Citation:

@inproceedings{ZKNR19,
    author = {Ben Zhou, Daniel Khashabi, Qiang Ning and Dan Roth},
    title = {"Going on a vacation" takes longer than "Going for a walk": A Study of Temporal Commonsense Understanding },
    booktitle = {EMNLP},
    year = {2019},
}

• Visualization (tfds.show_examples): Not supported.