CODER

CODER: Knowledge infused cross-lingual medical term embedding for term normalization. Paper

CODER++: Automatic Biomedical Term Clustering by Learning Fine-grained Term Representations. Paper

Use the model by transformers

Models have been uploaded to huggingface/transformers repo.

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("GanjinZero/UMLSBert_ENG")
model = AutoModel.from_pretrained("GanjinZero/UMLSBert_ENG")

English checkpoint: GanjinZero/coder_eng or GanjinZero/UMLSBert_ENG (old name)

English checkpoint CODER++: GanjinZero/coder_eng_pp (with hard negative sampling)

Multilingual checkpoint: GanjinZero/coder_all or GanjinZero/UMLSBert_ALL (discarded old name)

Train your model

cd pretrain
python train.py --umls_dir your_umls_dir --model_name_or_path monologg/biobert_v1.1_pubmed

your_umls_dir should contain MRCONSO.RRF, MRREL.RRF and MRSTY.RRF. UMLS Download path:UMLS.

A small tool for load UMLS RRF

from pretrain.load_umls import UMLS
umls = UMLS(your_umls_dir)

Test CODER or other embeddings

CADEC

cd test
python cadec/cadec_eval.py bert_model_name_or_path
python cadec/cadec_eval.py word_embedding_path

MANTRA GSC

Download the Mantra GSC and unzip the xml files to /test/mantra/dataset, run

cd test/mantra
python test.py

MCSM

cd test/embeddings_reimplement
python mcsm.py

DDBRC

Only sampled data is provided.

cd test/diseasedb
python train.py your_embedding embedding_type freeze_or_not gpu_id

• embedding_type should be in [bert, word, cui]
• freeze_or_not should be in [T, F], T means freeze the embedding, and F means fine-tune the embedding

Citation

@article{YUAN2022103983,
title = {CODER: Knowledge-infused cross-lingual medical term embedding for term normalization},
journal = {Journal of Biomedical Informatics},
pages = {103983},
year = {2022},
issn = {1532-0464},
doi = {https://doi.org/10.1016/j.jbi.2021.103983},
url = {https://www.sciencedirect.com/science/article/pii/S1532046421003129},
author = {Zheng Yuan and Zhengyun Zhao and Haixia Sun and Jiao Li and Fei Wang and Sheng Yu},
keywords = {medical term normalization, cross-lingual, medical term representation, knowledge graph embedding, contrastive learning}
}

@misc{https://doi.org/10.48550/arxiv.2204.00391,
  doi = {10.48550/ARXIV.2204.00391},
  url = {https://arxiv.org/abs/2204.00391},
  author = {Zeng, Sihang and Yuan, Zheng and Yu, Sheng},
  title = {Automatic Biomedical Term Clustering by Learning Fine-grained Term Representations},
  publisher = {arXiv},
  year = {2022}
}