TablERT: Named Entity Recognition and Relation Extraction using Enhanced Table Filling by Contextualized Representations

This is the PyTorch code for the preprint 'Named Entity Recognition and Relation Extraction using Enhanced Table Filling by Contextualized Representations' (TablERT). An extended version of this paper is accepted as a journal paper at JNLP. The general framework of this software adopts that of SpERT [1].

Setup

Requirements

Requirments are listed in requirements.txt, same as SpERT [1].

Required

• Python 3.5+
• PyTorch 1.1.0+ (tested with version 1.3.1)
• transformers 2.2.0+ (tested with version 2.2.0)
• scikit-learn (tested with version 0.21.3)
• tqdm (tested with version 4.19.5)
• numpy (tested with version 1.17.4)

Optional

• jinja2 (tested with version 2.10.3) - if installed, used to export relation extraction examples
• tensorboardX (tested with version 1.6) - if installed, used to save training process to tensorboard

Examples

We assume the processed CoNLL04 [2] data are already located in the folder data/datasets/, following the data split of [1] and [3]. (not provided along with this software)

To obtain the data, it might be convenient to use our pre-processing script, i.e, data_processing.py. To envoke the script, json files provided by SpERT, and files recording the BILOU-format annotations for the dataset provided by another project GlobalNormalization should be used.

Training

To train a model on CoNLL04 train set and evaluate on CoNLL04 development set, run

python ./table_filling.py train --config configs_example/train_conll04.conf

Evaluation

To evalute a model on CoNLL04 test set, fill in the field model_path in configs_example/eval_conll04.conf with the directory of the model and run

python ./table_filling.py eval --config configs_example/eval_conll04.conf

Citation

If you use the provided code in your work, please cite the following paper:

@article{Youmi Ma2022,
  title={Named Entity Recognition and Relation Extraction Using Enhanced Table Filling by Contextualized Representations},
  author={Youmi Ma and Tatsuya Hiraoka and Naoaki Okazaki},
  journal={Journal of Natural Language Processing},
  volume={29},
  number={1},
  pages={187-223},
  year={2022},
  doi={10.5715/jnlp.29.187}
}

References

[1]Markus Eberts and Adrian Ulges, 2020, 'Span-based joint entity and relation extraction with transformerpre-training' In 24th European Conference on Artifi-cial Intelligence (ECAI).
[2]Dan Roth and Wen-tau Yih, 2004, ‘A Linear Programming Formulation forGlobal Inference in Natural Language Tasks’, in Proc. of CoNLL 2004 at HLT-NAACL 2004, pp. 1–8.
[3]Pankaj Gupta, Hinrich Schütze, and Bernt Andrassy, 2016, ‘Table Filling Multi-Task Recurrent  Neural  Network  for  Joint  Entity  and  Relation Extraction’, in Proc. of COLING 2016, pp. 2537–2547.
[4]Heike Adel and Hinrich Schütze, 2017, 'Global Normalization of Convolutional Neural Networks for Joint Entity and Relation Classification', EMNLP 2017, pp. 1723--1729.