Query and Extract: Refining Event Extraction as Type Oriented Binary Decoding

Table of Contents (Optional)

• Installation
• Data Preparation
• Trigger Detection
• Argument Detection
• Citation
• License

Installation

To install the dependency packages, run

conda create --name query_extract_EE python=3.8
conda activate query_extract_EE
pip install -r requirements.txt

Data Preparation

1. Follow https://github.com/wilburOne/ACE_ERE_Scripts to process the raw data and save to ./data/ace_en/processed_data and ./data/ere_en/processed_data respectively
2. Save the event data into .txt files, process the .txt file and save as Torch TensorDataset

   ./setup.sh

   Trigger Detection

   To train the trigger detection model, run

   python scripts/run_trigger_detection.py --tr_dataset=${PATH_TO_TRAIN_DATASET} --dev_dataset=${PATH_TO_DEV_DATASET} 

   To evaluate the trigger detection performance, run

   python scripts/run_trigger_detection.py --EPOCH=0 --te_dataset=${PATH_TO_TEST_DATASET} 

Argument Detection

To train the argument detection model, run

python scripts/run_argument_detection.py --train_file_pt=${PATH_TO_TRAIN_FILE} --dev_file_pt=${PATH_TO_DEV_FILE}

To evaluate the argument detection model, run

python scripts/eval.py

Citation

If you find this repo useful, please cite the following paper:

@inproceedings{wang2022query,
    title={Query and Extract: Refining Event Extraction as Type-oriented Binary Decoding},
    author={Wang, Sijia and Yu, Mo and Chang, Shiyu and Sun, Lichao and Huang, Lifu},
    booktitle={Findings of the 2022 Association for Computational Linguistics},  
    year={2022}
}

License

Distributed under the MIT License.