MedCPT-Article-Encoder

1. Query-to-article search with both encoders.
2. Query representation for clustering or query-to-query search with the query encoder.
3. Article representation for clustering or article-to-article search with the article encoder.

For more details, please check out our paper (Bioinformatics, 2023). Please note that the released version is slightly different from the version reported in the paper.

Case 1. Using the MedCPT Article Encoder

import torch
from transformers import AutoTokenizer, AutoModel

model = AutoModel.from_pretrained("ncbi/MedCPT-Article-Encoder")
tokenizer = AutoTokenizer.from_pretrained("ncbi/MedCPT-Article-Encoder")

# each article contains a list of two texts (usually a title and an abstract)
articles = [
	[
		"Diagnosis and Management of Central Diabetes Insipidus in Adults",
		"Central diabetes insipidus (CDI) is a clinical syndrome which results from loss or impaired function of vasopressinergic neurons in the hypothalamus/posterior pituitary, resulting in impaired synthesis and/or secretion of arginine vasopressin (AVP). [...]",
	],
	[
		"Adipsic diabetes insipidus",
		"Adipsic diabetes insipidus (ADI) is a rare but devastating disorder of water balance with significant associated morbidity and mortality. Most patients develop the disease as a result of hypothalamic destruction from a variety of underlying etiologies. [...]",
	],
	[
		"Nephrogenic diabetes insipidus: a comprehensive overview",
		"Nephrogenic diabetes insipidus (NDI) is characterized by the inability to concentrate urine that results in polyuria and polydipsia, despite having normal or elevated plasma concentrations of arginine vasopressin (AVP). [...]",
	],
]

with torch.no_grad():
	# tokenize the articles
	encoded = tokenizer(
		articles,
		truncation=True,
		padding=True,
		return_tensors='pt',
		max_length=512,
	)

	# encode the queries (use the [CLS] last hidden states as the representations)
	embeds = model(**encoded).last_hidden_state[:, 0, :]

	print(embeds)
	print(embeds.size())

The output will be:

tensor([[-0.0189,  0.0115,  0.0988,  ..., -0.0655,  0.3155, -0.0357],
        [-0.3402, -0.3064, -0.0749,  ..., -0.0799,  0.3332,  0.1263],
        [-0.2764, -0.0506, -0.0608,  ...,  0.0389,  0.2532,  0.1580]])
torch.Size([3, 768])

These embeddings are also in the same space as those generated by the MedCPT query encoder.

Case 2. Use the Pre-computed Embeddings

We have provided the embeddings of all PubMed articles generated by the MedCPT article encoder at https://ftp.ncbi.nlm.nih.gov/pub/lu/MedCPT/pubmed_embeddings/.

Acknowledgments

This work was supported by the Intramural Research Programs of the National Institutes of Health, National Library of Medicine.

Disclaimer

This tool shows the results of research conducted in the Computational Biology Branch, NCBI/NLM. The information produced on this website is not intended for direct diagnostic use or medical decision-making without review and oversight by a clinical professional. Individuals should not change their health behavior solely on the basis of information produced on this website. NIH does not independently verify the validity or utility of the information produced by this tool. If you have questions about the information produced on this website, please see a health care professional. More information about NCBI's disclaimer policy is available.

Citation

If you find this repo helpful, please cite MedCPT by:

@article{jin2023medcpt,
  title={MedCPT: Contrastive Pre-trained Transformers with large-scale PubMed search logs for zero-shot biomedical information retrieval},
  author={Jin, Qiao and Kim, Won and Chen, Qingyu and Comeau, Donald C and Yeganova, Lana and Wilbur, W John and Lu, Zhiyong},
  journal={Bioinformatics},
  volume={39},
  number={11},
  pages={btad651},
  year={2023},
  publisher={Oxford University Press}
}