CoRTEx

Publish the model for the paper "CoRTEx: Contrastive Learning for Representing Terms via Explanations with Applications on Constructing Biomedical Knowledge Graphs"

Because of license of UMLS, we cannot share our training data. So we only provide our model and how to use our model. If you need the training details please contact [email protected]

Quick start

---

Installation

pip install InstructorEmbedding

Compute your customized embeddings

Then you can use the model like this to calculate domain-specific and task-aware embeddings:

from InstructorEmbedding import INSTRUCTOR
model = INSTRUCTOR('yinghy2018/CoRTEx')
term = "fever"
term_definition = "xxxx"
instruction = "Represent the biomedical term for identifying synonymous terms. Input: "
instruction2 = "Represent the meaning of the biomedical term for retrieval. Input: "
embeddings = model.encode([[instruction,term]])
embeddings = model.encode([[instruction2,term_definition]])
print(embeddings)

Cite

@article{ying2024cortex,
  title={CoRTEx: contrastive learning for representing terms via explanations with applications on constructing biomedical knowledge graphs},
  author={Ying, Huaiyuan and Zhao, Zhengyun and Zhao, Yang and Zeng, Sihang and Yu, Sheng},
  journal={Journal of the American Medical Informatics Association},
  pages={ocae115},
  year={2024},
  publisher={Oxford University Press}
}