This repository contains the implementation of the Causal Variational Approach (CVA) method, developed for causal inference in conditioned stochastic dynamics.
👉 Associated publication:
Inference in conditioned dynamics through causality restoration — Scientific Reports (2023)
https://www.nature.com/articles/s41598-023-33770-3
📁 RandomWalk
Contains a Jupyter notebook demonstrating the CVA method on a conditioned random walk problem. The notebook includes both the implementation and the results.
📁 src
Core source code implementing the CVA framework, used for more complex inference problems such as epidemic dynamics.
📁 Epidemics
Introductory examples of CVA applied to epidemic inference:
Example.ipynb: shows how to simulate and infer simple epidemic instances using the code in src.
Results/: contains data produced by simulation and inference.