---

---

Dataset Card for "scan"

Dataset Summary

SCAN tasks with various splits.

SCAN is a set of simple language-driven navigation tasks for studying compositional learning and zero-shot generalization.

See https://github.com/brendenlake/SCAN for a description of the splits.

Example usage: data = datasets.load_dataset('scan/length')

Supported Tasks

More Information Needed

Languages

More Information Needed

Dataset Structure

We show detailed information for up to 5 configurations of the dataset.

Data Instances

addprim_jump

• Size of downloaded dataset files: 17.82 MB
• Size of the generated dataset: 3.86 MB
• Total amount of disk used: 21.68 MB

An example of 'train' looks as follows.

addprim_turn_left

• Size of downloaded dataset files: 17.82 MB
• Size of the generated dataset: 3.90 MB
• Total amount of disk used: 21.71 MB

An example of 'train' looks as follows.

filler_num0

• Size of downloaded dataset files: 17.82 MB
• Size of the generated dataset: 2.72 MB
• Total amount of disk used: 20.53 MB

An example of 'train' looks as follows.

filler_num1

• Size of downloaded dataset files: 17.82 MB
• Size of the generated dataset: 2.99 MB
• Total amount of disk used: 20.81 MB

An example of 'train' looks as follows.

filler_num2

• Size of downloaded dataset files: 17.82 MB
• Size of the generated dataset: 3.28 MB
• Total amount of disk used: 21.10 MB

An example of 'train' looks as follows.

Data Fields

The data fields are the same among all splits.

addprim_jump

• commands: a string feature.
• actions: a string feature.

addprim_turn_left

• commands: a string feature.
• actions: a string feature.

filler_num0

• commands: a string feature.
• actions: a string feature.

filler_num1

• commands: a string feature.
• actions: a string feature.

filler_num2

• commands: a string feature.
• actions: a string feature.

Data Splits Sample Size

name	train	test
addprim_jump	14670	7706
addprim_turn_left	21890	1208
filler_num0	15225	1173
filler_num1	16290	1173
filler_num2	17391	1173

Dataset Creation

Curation Rationale

More Information Needed

Source Data

More Information Needed

Annotations

More Information Needed

Personal and Sensitive Information

More Information Needed

Considerations for Using the Data

Social Impact of Dataset

More Information Needed

Discussion of Biases

More Information Needed

Other Known Limitations

More Information Needed

Additional Information

Dataset Curators

More Information Needed

Licensing Information

More Information Needed

Citation Information

@inproceedings{Lake2018GeneralizationWS,
  title={Generalization without Systematicity: On the Compositional Skills of
         Sequence-to-Sequence Recurrent Networks},
  author={Brenden M. Lake and Marco Baroni},
  booktitle={ICML},
  year={2018},
  url={https://arxiv.org/pdf/1711.00350.pdf},
}

Contributions

Thanks to @lewtun, @patrickvonplaten, @mariamabarham, @thomwolf for adding this dataset.