Fine-Tune a BioBERT model on n2c2 dataset

The Dataset

n2c2, a wordplay on i2b2, is a dataset created for a competetion hosted by Harvard University. The task was to create a set of medical documents in which different information related to drugs are annotated.

The task was to be achieved by completeting two subtasks

1. Identify Drug Name, Dosage, Durations and other entities.
2. Create two relations
   • the relation of drug with symptoms and diseases.
   • the relation of drugs with Adverse Dru Event (ADE).

WHO defines Adverse Drug Event (ADE) as “an injury resulting from medical intervention related to a drug”. Eg. Drug administration errors.