metadata

license: cc-by-4.0
task_categories:
  - translation
language:
  - hi
  - en
tags:
  - code-mixing
  - Data-annotation
  - Hinglish
size_categories:
  - 10K<n<100K

Abstract

Code-mixing is the phenomenon of using more than one language in a sentence. In the multilingual communities, it is a very frequently observed pattern of communication on social media platforms. Flexibility to use multiple languages in one text message might help to communicate efficiently with the target audience. But, the noisy user-generated code-mixed text adds to the challenge of processing and understanding natural language to a much larger extent. Machine translation from monolingual source to the target language is a well-studied research problem. Here, we demonstrate that widely popular and sophisticated translation systems such as Google Translate fail at times to translate code-mixed text effectively. To address this challenge, we present a parallel corpus of the 13,738 code-mixed Hindi-English sentences and their corresponding human translation in English. In addition, we also propose a translation pipeline build on top of Google Translate. The evaluation of the proposed pipeline on PHINC demonstrates an increase in the performance of the underlying system. With minimal effort, we can extend the dataset and the proposed approach to other code-mixing language pairs.

Dataset Overview

PHINC (Parallel Hinglish Social Media Code-Mixed Corpus for Machine Translation): The dataset tackles challenges in translating noisy, informal, code-mixed social media text, offering 13,738 Hinglish-English sentence pairs manually annotated by 54 annotators for low-resource machine translation task.

The dataset contains the following fields:

Hinglish Code-Mixed Sentence: The original sentence in Romanized Hindi-English (Hinglish).
Human Translated English Sentence: The corresponding English translation provided by human annotators.

Dataset Attribution & Licensing

Curated by: Lingo Research Group at IIT Gandhinagar
Language(s) : Bilingual (Hindi [hi], English [en])
Licensed by: cc-by-4.0

Dataset Usage Disclaimer ‼️

This dataset may contain biased or offensive content and includes limited safety measures. Use it responsibly.

Citation

If you use this dataset, please cite the following work:

@inproceedings{srivastava-singh-2020-phinc,
    title = "{PHINC}: A Parallel {H}inglish Social Media Code-Mixed Corpus for Machine Translation",
    author = "Srivastava, Vivek  and
      Singh, Mayank",
    editor = "Xu, Wei  and
      Ritter, Alan  and
      Baldwin, Tim  and
      Rahimi, Afshin",
    booktitle = "Proceedings of the Sixth Workshop on Noisy User-generated Text (W-NUT 2020)",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2020.wnut-1.7/",
    doi = "10.18653/v1/2020.wnut-1.7",
    pages = "41--49"
}