Create an ADR (Adverse drug reaction) extraction model for unstructured text:
An adverse drug reaction (ADR) can be defined as an appreciably harmful or unpleasant reaction resulting from an intervention related to the use of a medicinal product. The goal of this project is to extracts the “ADR” term from the unstructured text . Like social media text or any EHR document.
For example : “Rash in hands caused by omeprazole.” here we have ADR: “Rash” which is caused by omeprazole.
Model(s)
The bio_bert model are good to start for the ADR extraction task.
https://huggingface.co/dmis-lab/biobert-base-cased-v1.2
https://huggingface.co/emilyalsentzer/Bio_ClinicalBERT
https://huggingface.co/gsarti/biobert-nli
Datasets
ade_corpus_v2 is usually a good corpus to test the ADR extraction model and you can find other open source ADR corups s well.
Challenges
you are given a social media post or raw text from EHR document and you need to extract “ADR” mention as accurately as possible.
Desired project outcomes
Create a Streamlit or Gradio app on huggingface, Spaces that can Extract “ADR” mention from the raw text.
Don’t forget to push all your models and datasets to the Hub so others can build on them!
Additional resources
https://huggingface.co/datasets/ade_corpus_v2#additional-information
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6947008/
https://towardsdatascience.com/automated-adverse-drug-event-ade-detection-from-text-in-spark-nlp-with-biobert-837c700f5d8c
Discord channel
To chat and organise with other people interested in this project, head over to our Discord 1 and:
Follow the instructions on the #join-course channel
Join the #adr-extraction channel
Just make sure you comment here to indicate that you’ll be contributing to this project