This repository provides the code for ARM (Attention Retrieval Model), a novel self-attention framework for flexible entity relation extraction from biological literature.
Authors and Contributors: Prashant Srivastava, Saptarshi Bej, Kristian Schultz, Kristina Yordanova, Olaf Wolkenhauer
Steps
-
Create a new folder in root dir named 'Data'.
-
Download Datasets for AIR, TRRUST, GAD, ChemProt, BioGRID, and Elangovan et al. and save them in /Data.
-
Register at Entrez Programming Utilities (E-utilities) and obtain Email and api-key.
-
Create a file named 'config.py' containing entrez email and api-key as:
entrez_api_key = 'Your-Entrez-email'
entrez_email = 'Your-Entrez-apikey'
- Navigate to /Preprocessing dir and run the jupyter-notebooks for all datasets.
Running all the notebooks in this directory with produce entity normalized datasets that are required in next step. - Run Typed Interactions.ipynb for typed relations Case Study.
- Run Untyped Interactions.ipynb for untyped relations Case Study.