Source code for AAAI2020 paper:End-to-End Bootstrapping Neural Network for Entity Set Expansion
- pytorch >= 1.2.0
- torchtext >= 0.3.1
- We use the dataset provided by Zupon et al.(2019), and you can also download them as well as all preprocessed data from google driver
- We use the Glove for entity and pattern embedding initialization ( You can use the pre-trained Bert for embedding initialization)
- To accelarate the model training and testing process, we use following preprocess steps:
cd preprocess
python generate_graph.py # generate graph data in numerical representation for quick reading
python generate_dev.py # generate development data
python generate_initialization.py # generate entity and pattern initialized embeddings.
- All prepreocessed files are included in google driver, just unzip it and put them in "data" directory.
- To execute our code, please run:
python run_conoll.py # for conoll dataset
# python run_onto.py # for onto dataset
Please cite the following paper if you find our code is helpful.
@inproceedings{yan_end_2020,
author = "Lingyong, Yan and
Xianpei, Han and
Ben, He and
Le, Sun",
title = "End-to-End Bootstrapping Neural Network for Entity Set Expansion",
booktitle = "Thirty-Fourth AAAI Conference on Artificial Intelligence",
year = "2020"
}