• Corpus ID: 333683

Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks

  title={Distinguishing Cause from Effect Using Observational Data: Methods and Benchmarks},
  author={Joris M. Mooij and J. Peters and Dominik Janzing and Jakob Zscheischler and Bernhard Sch{\"o}lkopf},
The discovery of causal relationships from purely observational data is a fundamental problem in science. The most elementary form of such a causal discovery problem is to decide whether X causes Y or, alternatively, Y causes X, given joint observations of two variables X,Y. An example is to decide whether altitude causes temperature, or vice versa, given only joint measurements of both variables. Even under the simplifying assumptions of no confounding, no feedback loops, and no selection bias… 

Comparing Causal Discovery Methods using Synthetic and Real Data

The goal of this work is to evaluate and to compare a number of state-of-the-art techniques in a joint benchmark and employ synthetic data, where it can give detailed reasoning about the resulting performance of the individual methods given their underlying assumptions.

Bivariate Causal Discovery via Conditional Divergence

This work proposes Conditional Divergence based Causal Inference (CDCI), a novel algorithm for detecting causal direction in purely observational data that can relax multiple strict assumptions commonly adopted in the causal discovery literature, including functional form and noise model.

Leveraging directed causal discovery to detect latent common causes

A general heuristic is devised which takes a causal discovery algorithm that can only distinguish purely directed causal relations and modifies it to also detect latent common causes, and demonstrates that both the modified algorithms preserve the performance of the original in distinguishing directed causal Relations.

The Effect of Noise Level on Causal Identification with Additive Noise Models

  • B. Kap
  • Computer Science
  • 2021
This work considered bivariate cases, which is the most elementary form of a causal discovery problem where one needs to decide whetherX causes Y or Y causesX, given joint distributions of two variablesX, Y, and found that these methods can fail to capture the true causal direction for some levels of noise.

Joint Causal Inference from Multiple Contexts

This work introduces Joint Causal Inference, a novel approach to causal discovery from multiple data sets from different contexts that elegantly unifies both approaches and concludes that JCI implementations can considerably outperform state-of-the-art causal discovery algorithms.

Joint Causal Inference from Multiple Datasets

JCI deals with several different types of interventions in a unified fashion, it can learn intervention targets, it pools data across different datasets which improves the statistical power of independence tests, and by exploiting differences in distribution between contexts it improves on the accuracy and identifiability of the predicted causal relations.

Causal Discovery Beyond Conditional Independences

This thesis addresses the problem of causal discovery, that is, recovering the underlying causal structure based on the joint probability distribution of the observed random variables, and estimates a set of Markov equivalent graphs.

Effectiveness of classification approach in recovering pairwise causal relations from data.

This thesis extends the bivariate case to the possibility of having at least one confounder between X and Y and improves the Bayesian network scores for the network structures of some medium to large-sized networks using the standard ordering based search algorithms.

The Randomized Causation Coefficient

This short paper proposes to learn how to perform causal inference directly from data, without the need of feature engineering, and poses causality as a kernel mean embedding classification problem, where inputs are samples from arbitrary probability distributions on pairs of random variables, and labels are types of causal relationships.

Results of the Cause-Effect Pair Challenge

A challenge in causal discovery from observational data with the aim of devising a “causation coefficient” to score pairs of variables to discover whether the data supports the hypothesis that Y = f(X, noise), which for the purpose of this challenge was the definition of causality.



Inference of Cause and Effect with Unsupervised Inverse Regression

This work addresses the problem of causal discovery in the two-variable case, given a sample from their joint distribution and proposes an implicit notion of independence, namely that pY|X cannot be estimated based on pX (lower case denotes density), however, it may be possible to estimate pY |X based on the density of the effect, pY.

Distinguishing Cause and Effect via Second Order Exponential Models

This work proposes a method to infer causal structures containing both discrete and continuous variables and describes that sometimes a causal hypothesis must be rejected because P(effect|cause) and P(cause) share algorithmic information (which is untypical if they are chosen independently).

Causal Inference on Discrete Data Using Additive Noise Models

It is proved that it almost never occurs that additive noise models can be fit in both directions, and an efficient algorithm is proposed that is able to perform this way of causal inference on finite samples of discrete variables.

Regression by dependence minimization and its application to causal inference in additive noise models

This work proposes a novel method for regression that minimizes the statistical dependence between regressors and residuals, and proposes an algorithm for efficiently inferring causal models from observational data for more than two variables.

A Linear Non-Gaussian Acyclic Model for Causal Discovery

This work shows how to discover the complete causal structure of continuous-valued data, under the assumptions that (a) the data generating process is linear, (b) there are no unobserved confounders, and (c) disturbance variables have non-Gaussian distributions of non-zero variances.

Causal discovery with continuous additive noise models

If the observational distribution follows a structural equation model with an additive noise structure, the directed acyclic graph becomes identifiable from the distribution under mild conditions, which constitutes an interesting alternative to traditional methods that assume faithfulness and identify only the Markov equivalence class of the graph, thus leaving some edges undirected.

Probabilistic latent variable models for distinguishing between cause and effect

A novel method for inferring whether X causes Y or vice versa from joint observations of X and Y is proposed, which considers the hypothetical effect variable to be a function of the hypothetical cause variable and an independent noise term.

Causal Inference Using Nonnormality

Path analysis, often applied to observational data to study causal structures, describes causal relationship between observed variables. The path analysis is of confirmatory nature and can make