chr_en / README.md
albertvillanova's picture
Convert dataset to Parquet (#2)
1b111ec verified
metadata
annotations_creators:
  - expert-generated
  - found
  - no-annotation
language_creators:
  - found
language:
  - chr
  - en
license:
  - other
multilinguality:
  - monolingual
  - multilingual
  - translation
size_categories:
  - 100K<n<1M
  - 10K<n<100K
  - 1K<n<10K
source_datasets:
  - original
task_categories:
  - fill-mask
  - text-generation
  - translation
task_ids:
  - language-modeling
  - masked-language-modeling
paperswithcode_id: chren
config_names:
  - monolingual
  - monolingual_raw
  - parallel
  - parallel_raw
dataset_info:
  - config_name: monolingual
    features:
      - name: sentence
        dtype: string
    splits:
      - name: chr
        num_bytes: 882824
        num_examples: 5210
      - name: en5000
        num_bytes: 615275
        num_examples: 5000
      - name: en10000
        num_bytes: 1211605
        num_examples: 10000
      - name: en20000
        num_bytes: 2432298
        num_examples: 20000
      - name: en50000
        num_bytes: 6065580
        num_examples: 49999
      - name: en100000
        num_bytes: 12130164
        num_examples: 100000
    download_size: 16967664
    dataset_size: 23337746
  - config_name: monolingual_raw
    features:
      - name: text_sentence
        dtype: string
      - name: text_title
        dtype: string
      - name: speaker
        dtype: string
      - name: date
        dtype: int32
      - name: type
        dtype: string
      - name: dialect
        dtype: string
    splits:
      - name: full
        num_bytes: 1210056
        num_examples: 5210
    download_size: 410646
    dataset_size: 1210056
  - config_name: parallel
    features:
      - name: sentence_pair
        dtype:
          translation:
            languages:
              - en
              - chr
    splits:
      - name: train
        num_bytes: 3089562
        num_examples: 11639
      - name: dev
        num_bytes: 260401
        num_examples: 1000
      - name: out_dev
        num_bytes: 78126
        num_examples: 256
      - name: test
        num_bytes: 264595
        num_examples: 1000
      - name: out_test
        num_bytes: 80959
        num_examples: 256
    download_size: 2143266
    dataset_size: 3773643
  - config_name: parallel_raw
    features:
      - name: line_number
        dtype: string
      - name: sentence_pair
        dtype:
          translation:
            languages:
              - en
              - chr
      - name: text_title
        dtype: string
      - name: speaker
        dtype: string
      - name: date
        dtype: int32
      - name: type
        dtype: string
      - name: dialect
        dtype: string
    splits:
      - name: full
        num_bytes: 5010734
        num_examples: 14151
    download_size: 2018726
    dataset_size: 5010734
configs:
  - config_name: monolingual
    data_files:
      - split: chr
        path: monolingual/chr-*
      - split: en5000
        path: monolingual/en5000-*
      - split: en10000
        path: monolingual/en10000-*
      - split: en20000
        path: monolingual/en20000-*
      - split: en50000
        path: monolingual/en50000-*
      - split: en100000
        path: monolingual/en100000-*
  - config_name: monolingual_raw
    data_files:
      - split: full
        path: monolingual_raw/full-*
  - config_name: parallel
    data_files:
      - split: train
        path: parallel/train-*
      - split: dev
        path: parallel/dev-*
      - split: out_dev
        path: parallel/out_dev-*
      - split: test
        path: parallel/test-*
      - split: out_test
        path: parallel/out_test-*
    default: true
  - config_name: parallel_raw
    data_files:
      - split: full
        path: parallel_raw/full-*

Dataset Card for ChrEn

Table of Contents

Dataset Description

Dataset Summary

ChrEn is a Cherokee-English parallel dataset to facilitate machine translation research between Cherokee and English. ChrEn is extremely low-resource contains 14k sentence pairs in total, split in ways that facilitate both in-domain and out-of-domain evaluation. ChrEn also contains 5k Cherokee monolingual data to enable semi-supervised learning.

Supported Tasks and Leaderboards

The dataset is intended to use for machine-translation between Enlish (en) and Cherokee (chr).

Languages

The dataset contains Enlish (en) and Cherokee (chr) text. The data encompasses both existing dialects of Cherokee: the Overhill dialect, mostly spoken in Oklahoma (OK), and the Middle dialect, mostly used in North Carolina (NC).

Dataset Structure

Data Instances

[More Information Needed]

Data Fields

[More Information Needed]

Data Splits

[More Information Needed]

Dataset Creation

Curation Rationale

[More Information Needed]

Source Data

Initial Data Collection and Normalization

Many of the source texts were translations of English materials, which means that the Cherokee structures may not be 100% natural in terms of what a speaker might spontaneously produce. Each text was translated by people who speak Cherokee as the first language, which means there is a high probability of grammaticality. These data were originally available in PDF version. We apply the Optical Character Recognition (OCR) via Tesseract OCR engine to extract the Cherokee and English text.

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

The sentences were manually aligned by Dr. Benjamin Frey a proficient second-language speaker of Cherokee, who also fixed the errors introduced by OCR. This process is time-consuming and took several months.

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

The dataset was gathered and annotated by Shiyue Zhang, Benjamin Frey, and Mohit Bansal at UNC Chapel Hill.

Licensing Information

The copyright of the data belongs to original book/article authors or translators (hence, used for research purpose; and please contact Dr. Benjamin Frey for other copyright questions).

Citation Information

@inproceedings{zhang2020chren,
  title={ChrEn: Cherokee-English Machine Translation for Endangered Language Revitalization},
  author={Zhang, Shiyue and Frey, Benjamin and Bansal, Mohit},
  booktitle={EMNLP2020},
  year={2020}
}

Contributions

Thanks to @yjernite, @lhoestq for adding this dataset.