A toy disease progression model with confounding by severity.
This is a simple simulation of outcomes from a viral infection, designed to produce data from which we can learn causal models where we know what the correct answer should be, or use those models to create analytical simulations.
These are the HTML versions of the notebooks:
Experiments with a constant policy offline and simulated environment are in the directory
- RL_offline/
You will need to install the gym package for python to run them.
This repository was originally hosted under the Virtual Generalist repo, but we split it out to make it easier to find and manage.