Causal Discovery and Machine Learning – Advanced Deep Learning and Beyond
Welcome to Chapter 14!
We're inevitably moving towards the end of our book, but we still have something to learn!
In the previous chapter, we introduced four families of causal discovery models: constraint-based, score-based, functional, and gradient-based. Each of the families and methods that we discussed came with unique strengths and unique limitations.
In this chapter, we’ll introduce methods and ideas that aim to solve some of these limitations. We’ll discuss an advanced deep learning causal discovery framework, Deep End-to-end Causal Inference (DECI), and implement it using the Microsoft open source library Causica and PyTorch.
We’ll see how to approach data with hidden confounding using the fast causal inference (FCI) algorithm and introduce other algorithms that can be used in similar scenarios.
After that, we’ll introduce two frameworks that allow...
