Perturbed Top-K Optimization

Here we present the Perturbed Top-K Optimization objective with extensions to the synthetic data experiments in the IMLH 2023 paper Learning where to intervene with a differentiable top-K operator: Towards data-driven strategies to prevent fatal opioid overdoses.

Requirements:

Numpy, Tensorflow 2.0+, perturbations.py from this repository

Code overview

• Perturbed Top-K Demonstration.ipynb A notebook demonstrating the creation of synthetic data and model training
• create_data.py Generates and saves the synthetic data
• create_model.py Creates a tensorflow model that can utilize the Perturbed-Top-K module, or without
• data_utils.py Miscellaneous utilities for generating and visualizing the synthetic data