README.md

metadata

language: en
tags:
  - greco
  - grammar
  - grammaticality
  - gec
base_model: microsoft/deberta-v3-large
datasets: w&i+locness
model-index:
  - name: GRECO
    results:
      - task:
          type: grammatical-error-correction
          name: Grammatical Error Correction
        dataset:
          type: conll-2014-shared-task-grammatical-error
          name: CoNLL-2014
          split: test
        metrics:
          - type: f0.5
            value: 71.12
            name: F0.5
        source:
          name: NLP-progress
          url: https://nlpprogress.com/english/grammatical_error_correction.html
license: gpl-3.0

GRECO: Gammaticality-scorer for re-ranking corrections

GRECO is a quality estimation model for grammatical error correction. The model is trained to detect which words are incorrect and whether a word or phrase needs to be inserted after certain words. You can then use the model to get the grammaticality score of a sentence.

Please check the official repository for more implementation details and updates.

The model was published in the following paper:

System Combination via Quality Estimation for Grammatical Error Correction (PDF | ACL Anthology)
Muhammad Reza Qorib and Hwee Tou Ng
The 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP)

Citation

If you find it useful for your work, please cite the paper:

@inproceedings{qorib-ng-2023-system,
    title = "System Combination via Quality Estimation for Grammatical Error Correction",
    author = "Qorib, Muhammad Reza  and
      Ng, Hwee Tou",
    editor = "Bouamor, Houda  and
      Pino, Juan  and
      Bali, Kalika",
    booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.emnlp-main.785",
    doi = "10.18653/v1/2023.emnlp-main.785",
    pages = "12746--12759",
}