ember

Bridging the Gap between Reality and Ideality of Entity Matching:
A Revisiting and Benchmark Re-Construction

Arxiv Conference

Description

Code and data for the paper:

Bridging the Gap between Reality and Ideality of Entity Matching: A Revisiting and Benchmark Re-Construction

Data

Details of the released data can be found in the REAME of the data.

How to run

First, install dependencies

# clone project
git clone https://github.com/tshu-w/ember
cd ember

# [SUGGESTED] use conda environment
conda env create -n ember -f environment.yaml
conda activate ember

# [ALTERNATIVE] install requirements directly
pip install -r requirements.txt

Next, to obtain the main results of the paper:

bash scripts/download_images.sh

python scripts/run_ali.py --gpus 0 1 2 3
python scripts/test_ali.py --gpus 0 1 2 3
python scripts/run_dm_ali.py --gpus 0 1 2 3
python scripts/test_dm_ali.py --gpus 0 1 2 3

python scripts/print_results results/test -k test/f1 test/prc test/rec

You can also run experiments with the run script.

# fit with the TextMatcher config
./run fit --config configs/ali_tm.yaml
# or specific command line arguments
./run fit --model TextMatcher --data AliDataModule --data.batch_size 32 --trainer.gpus 0,

# evaluate with the checkpoint
./run test --config configs/ali_tm.yaml --ckpt_path ckpt_path

# get the script help
./run --help
./run fit --help

Citation

@inproceedings{ijcai2022p552,
  title     = {Bridging the Gap between Reality and Ideality of Entity Matching: A Revisting and Benchmark Re-Constrcution},
  author    = {Wang, Tianshu and Lin, Hongyu and Fu, Cheng and Han, Xianpei and Sun, Le and Xiong, Feiyu and Chen, Hui and Lu, Minlong and Zhu, Xiuwen},
  booktitle = {Proceedings of the Thirty-First International Joint Conference on
               Artificial Intelligence, {IJCAI-22}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Lud De Raedt},
  pages     = {3978--3984},
  year      = {2022},
  month     = {7},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2022/552},
  url       = {https://doi.org/10.24963/ijcai.2022/552},
}