gemire / autoner

Learning Named Entity Tagger from Domain-Specific Dictionary

AutoNER

Check Our New NER Toolkit🚀🚀🚀

• Inference:
  • LightNER: inference w. models pre-trained / trained w. any following tools, efficiently.
• Training:
  • LD-Net: train NER models w. efficient contextualized representations.
  • VanillaNER: train vanilla NER models w. pre-trained embedding.
• Distant Training:
  • AutoNER: train NER models w.o. line-by-line annotations and get competitive performance.

---

No line-by-line annotations, AutoNER trains named entity taggers with distant supervision.

Details about AutoNER can be accessed at: https://arxiv.org/abs/1809.03599

• Model notes
• Benchmarks
• Training
  • Required Inputs
  • Dependencies
  • Command
• Citation

Model Notes

Benchmarks

Method	Precision	Recall	F1
Supervised Benchmark	88.84	85.16	86.96
Dictionary Match	93.93	58.35	71.98
Fuzzy-LSTM-CRF	88.27	76.75	82.11
AutoNER	88.96	81.00	84.80

Training

Required Inputs

• Tokenized Raw Texts
  • Example: data/BC5CDR/raw_text.txt
    • One token per line.
    • An empty line means the end of a sentence.
• Two Dictionaries
  • Core Dictionary w/ Type Info
    • Example: data/BC5CDR/dict_core.txt
      • Two columns (i.e., Type, Tokenized Surface) per line.
      • Tab separated.
    • How to obtain?
      • From domain-specific dictionaries.
  • Full Dictionary w/o Type Info
    • Example: data/BC5CDR/dict_full.txt
      • One tokenized high-quality phrases per line.
    • How to obtain?
      • From domain-specific dictionaries.
      • Applying the high-quality phrase mining tool on domain-specific corpus.
        • AutoPhrase
• Pre-trained word embeddings
  • Train your own or download from the web.
  • The example run uses embedding/bio_embedding.txt, which can be downloaded from our group's server. For example, curl http://dmserv4.cs.illinois.edu/bio_embedding.txt -o embedding/bio_embedding.txt. Since the embedding encoding step consumes quite a lot of memory, we also provide the encoded file in the autoner_train.sh.
• [Optional] Development & Test Sets.
  • Example: data/BC5CDR/truth_dev.ck and data/BC5CDR/truth_test.ck
    • Three columns (i.e., token, Tie or Break label, entity type).
    • I is Break.
    • O is Tie.
    • Two special tokens <s> and <eof> mean the start and end of the sentence.

Dependencies

This project is based on python>=3.6. The dependent package for this project is listed as below:

numpy==1.13.1
tqdm
torch-scope>=0.5.0
pytorch==0.4.1

Command

To train an AutoNER model, please run

./autoner_train.sh

To apply the trained AutoNER model, please run

./autoner_test.sh

You can specify the parameters in the bash files. The variables names are self-explained.

Citation

Please cite the following two papers if you are using our tool. Thanks!

• Jingbo Shang*, Liyuan Liu*, Xiaotao Gu, Xiang Ren, Teng Ren and Jiawei Han, "Learning Named Entity Tagger using Domain-Specific Dictionary", in Proc. of 2018 Conf. on Empirical Methods in Natural Language Processing (EMNLP'18), Brussels, Belgium, Oct. 2018. (* Equal Contribution)
• Jingbo Shang, Jialu Liu, Meng Jiang, Xiang Ren, Clare R Voss, Jiawei Han, "Automated Phrase Mining from Massive Text Corpora", accepted by IEEE Transactions on Knowledge and Data Engineering, Feb. 2018.

@inproceedings{shang2018learning,
  title = {Learning Named Entity Tagger using Domain-Specific Dictionary}, 
  author = {Shang, Jingbo and Liu, Liyuan and Ren, Xiang and Gu, Xiaotao and Ren, Teng and Han, Jiawei}, 
  booktitle = {EMNLP}, 
  year = 2018, 
}

@article{shang2018automated,
  title = {Automated phrase mining from massive text corpora},
  author = {Shang, Jingbo and Liu, Jialu and Jiang, Meng and Ren, Xiang and Voss, Clare R and Han, Jiawei},
  journal = {IEEE Transactions on Knowledge and Data Engineering},
  year = {2018},
  publisher = {IEEE}
}