How Good is the Model in Model-in-the-loop Event Coreference Resolution Annotations

Accompanying code for the paper, How Good is the Model in Model-in-the-loop Event Coreference Resolution Annotations published in the 17th Linguistics Annotation Workshop, ACL 2023.

Annotation interface

Prereqs

pip install bs4 lxml spacy
python -m spacy download en_core_web_md

Download and install Prodigy: https://prodi.gy/docs/install

Preparing ECB+ Dataset

Event Coref Bank Plus Corpus: https://github.com/cltl/ecbPlus

git clone https://github.com/cltl/ecbPlus.git

python parse_ecb.py en_core_web_md test

Running the Prodigy Recipe

prodigy evt-coref ecb_test_coref en_core_web_md ./ecb_test_set.json -F coref_recipe.py

Evaluation Methodology - Simulation

Please see recall_comparisons.ipynb

Lambda Analysis on Dev sets

python simulation.py

Citation

If you find this code to be useful, please use the following citation:

@inproceedings{ahmed-etal-2023-good,
    title = "How Good Is the Model in Model-in-the-loop Event Coreference Resolution Annotation?",
    author = "Ahmed, Shafiuddin Rehan  and
      Nath, Abhijnan  and
      Regan, Michael  and
      Pollins, Adam  and
      Krishnaswamy, Nikhil  and
      Martin, James H.",
    editor = "Prange, Jakob  and
      Friedrich, Annemarie",
    booktitle = "Proceedings of the 17th Linguistic Annotation Workshop (LAW-XVII)",
    month = jul,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.law-1.14",
    doi = "10.18653/v1/2023.law-1.14",
    pages = "136--145",
}