We Are Not Your Real Parents: Telling Causal from Confounded using MDL

01/21/2019
by   David Kaltenpoth, et al.
14

Given data over variables (X_1,...,X_m, Y) we consider the problem of finding out whether X jointly causes Y or whether they are all confounded by an unobserved latent variable Z. To do so, we take an information-theoretic approach based on Kolmogorov complexity. In a nutshell, we follow the postulate that first encoding the true cause, and then the effects given that cause, results in a shorter description than any other encoding of the observed variables. The ideal score is not computable, and hence we have to approximate it. We propose to do so using the Minimum Description Length (MDL) principle. We compare the MDL scores under the models where X causes Y and where there exists a latent variables Z confounding both X and Y and show our scores are consistent. To find potential confounders we propose using latent factor modeling, in particular, probabilistic PCA (PPCA). Empirical evaluation on both synthetic and real-world data shows that our method, CoCa, performs very well -- even when the true generating process of the data is far from the assumptions made by the models we use. Moreover, it is robust as its accuracy goes hand in hand with its confidence.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/26/2017

Telling Cause from Effect using MDL-based Local and Global Regression

We consider the fundamental problem of inferring the causal direction be...
research
02/21/2017

Causal Inference on Multivariate and Mixed-Type Data

Given data over the joint distribution of two random variables X and Y, ...
research
02/22/2017

Causal Inference by Stochastic Complexity

The algorithmic Markov condition states that the most likely causal dire...
research
02/12/2021

A Critical Look At The Identifiability of Causal Effects with Deep Latent Variable Models

Using deep latent variable models in causal inference has attracted cons...
research
02/12/2021

Do-calculus enables causal reasoning with latent variable models

Latent variable models (LVMs) are probabilistic models where some of the...
research
11/18/2020

Detecting Hierarchical Changes in Latent Variable Models

This paper addresses the issue of detecting hierarchical changes in late...
research
02/14/2012

Noisy-OR Models with Latent Confounding

Given a set of experiments in which varying subsets of observed variable...

Please sign up or login with your details

Forgot password? Click here to reset