Detecting low-complexity unobserved causes

By Dominik Janzing, Eleni Sgouritsa, Oliver Stegle, Jan Peters, Bernhard Schölkopf. 2011

24 citations recorded by the source corpus.

We describe a method that infers whether statistical dependences between two observed variables X and Y are due to a "direct" causal link or only due to a connecting causal path that contains an unobserved variable of low complexity, e.g., a binary variable. This problem is motivated by statistical genetics. Given a genetic marker that is correlated with a phenotype of interest, we want to detect whether this marker is causal or it only correlates with a causal one. Our method is based on the analysis of the location of the conditional distributions P(Y|x) in the simplex of all distributions of Y. We report encouraging results on semi-empirical data.

Subfields: Artificial Intelligence, Genetics, Molecular Biology, Causal Inference & Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bernhard Schölkopf
2. Krikamol Muandet
3. Bharath K. Sriperumbudur
4. Kenji Fukumizu
5. Jan Peters
6. David Lopez-Paz
7. Arthur Gretton
8. Dominik Janzing

In Genetics

1. Jan Peters
2. Kenji Fukumizu
3. Dominik Janzing
4. Bernhard Schölkopf
5. Eleni Sgouritsa

In Molecular Biology

1. Bernhard Schölkopf
2. Dominik Janzing
3. Kenji Fukumizu
4. Jan Peters
5. Eleni Sgouritsa
6. Krikamol Muandet
7. Bharath K. Sriperumbudur
8. Michel Besserve

In Causal Inference & Discovery

1. Bernhard Schölkopf
2. Dominik Janzing
3. Jan Peters
4. Krikamol Muandet
5. David Lopez-Paz
6. Kun Zhang
7. Eleni Sgouritsa
8. B. Schoelkopf

Citations in the corpus

• Kernel Mean Embedding of Distributions: A Review and Beyonds (2016, background): “…al. 2013), Markov decision processes and reinforcement learning (Grünewälder et al. 2012, Nishiyama et al. 2012), latent variable models (Song et al. 2010b; 2011a;b), the kernel Bayes’ rule (Fukumizu et al. 2011), and causal discovery (Janzing et al. 2011, Sgouritsa et al. 2013, Chen et al. 2014).”
• Identifying Finite Mixtures of Nonparametric Product Distributions and Causal Inference of Confounders (2013, background): “Finally, the reconstruction of binary confounders under the assumption of pure conditionals is presented in Janzing et al. (2011).”
• Estimating Causal Effects by Bounding Confounding (2014, background): “[Janzing et al., 2011]. for i = 1, 2.”
• Detecting Confounding in Multivariate Linear Models via Spectral Analysis (2017, background): “[7] assumes a discrete confounder attaining a few values only.”