mrpeerat / mReFinED

mReFinED: An Efficient End-to-End Multilingual Entity Linking System

Overview

We propose mReFinED, the first end-to-end MEL model. mReFinED supports 9 languages: AR, EN, ES, DE, FA, JA, TA, and TR. Our experimental results in the research paper demonstrated that mReFinED outperformed the best existing work in the end-to-end MEL task while being 44 times faster compared to existing state-of-the-art (mGENRE).

mReFinED's Paper

The mReFinED model architecture is described in the paper below (https://aclanthology.org/2023.findings-emnlp.1007):

@inproceedings{limkonchotiwat-etal-2023-mrefined,
    title = "m{R}e{F}in{ED}: An Efficient End-to-End Multilingual Entity Linking System",
    author = "Limkonchotiwat, Peerat  and
      Cheng, Weiwei  and
      Christodoulopoulos, Christos  and
      Saffari, Amir  and
      Lehmann, Jens",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.findings-emnlp.1007",
    doi = "10.18653/v1/2023.findings-emnlp.1007",
    pages = "15080--15089",
}

mReFinED

• This is the replica of mReFinED from Amazon's mReFinED.
• We improve the training and inference codes to make them easier to reproduce.
• We also provide the mReFinED model and training data :)

Hardware Requirements

• mReFinED has a low hardware requirement. For fast inference speed, a GPU should be used, but this is not a strict requirement.
• We create training data for 15 days (CPU only). However, the process can be sped up using GPUs (~2 days).
• We use 8 V100 in the training step for ~10 days.
• For the inference setting, we use only a single V100.

Model, Data, and Codes

Materials

• Model: XXXXXXX
• Training data: XXXXXXXX

Example Script

• mReFinED: Creating training data

cd mReFinED/src/
export PYTHONPATH=$PYTHONPATH:src
python refined/offline_data_generation/preprocess_all_multilingual_combine.py

• mReFinED: Training

cd mReFinED/src/
export PYTHONPATH=$PYTHONPATH:src
bash refined/training/train/multilingual_train.sh

• Mention Detection For Unlabeled Entity in Wikipedia. However, we can skip this step using WikiNN instead.

cd mReFinED/src/refined/training/train
python multilingual_md_train_xtreme.py
python md_on_wiki.py
python multilingual_md_train_xtreme_wikipedia.py

• mReFinED: Inference

print('hi')

• mReFinED on Mewsli-9

print('hi')

Security

See CONTRIBUTING for more information.

License

This library is licensed under the CC-BY-NC 4.0 License.

Contact us

If you have questions please open Github issues instead of sending us emails, as some of the listed email addresses are no longer active.