Modeling confounding by half-sibling regression

By Bernhard Schölkopf, David W. Hogg, Dun Wang, Daniel Foreman-Mackey, Dominik Janzing, Carl-Johann Simon-Gabriel, Jan Peters. 2016

76 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability, 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. Yoshua Bengio
3. Anirudh Goyal
4. Stefan Bauer
5. Francesco Locatello
6. Nan Rosemary Ke
7. Kyunghyun Cho
8. Sergey Levine

In Statistics and Probability

1. Bernhard Schölkopf
2. Francesco Locatello
3. Stefan Bauer
4. Anirudh Goyal
5. Yoshua Bengio
6. Nan Rosemary Ke
7. Nal Kalchbrenner
8. José Miguel Hernández-Lobato

In Causal Inference & Discovery

1. Yoshua Bengio
2. Anirudh Goyal
3. Stefan Bauer
4. Francesco Locatello
5. Bernhard Schölkopf
6. Nan Rosemary Ke
7. Nal Kalchbrenner
8. Paul K. Rubenstein

Citations in the corpus

• Towards Causal Representation Learning (2021, background): “patterns in measurement noise across stars light-years apart can be removed in order to reduce the instrument’s influence on the measurement [219], which is critical especially in the context of partial technical failures as experienced in the Kepler exoplanet search mission.”
• Causality for Machine Learning (2019, methodology): “The method recovers the random variable representing the desired signal almost surely (up to a constant offset), for an additive noise model, and subject to the assumption that the instrument’s effect on the star is in principle predictable from the other stars (Schölkopf et al., 2016a).”
• The Incomplete Rosetta Stone problem: Identifiability results for Multi-view Nonlinear ICA (2019, methodology): “Half-sibling regression [35] is a method to reconstruct a source from noisy observations by exploiting other sources that are affected by the same noise process but otherwise independent from it.”
• Deconfounding Reinforcement Learning in Observational Settings (2018, background): “A common practice is then to leverage observed proxy variables that contain information about the unobserved confounders (Angrist and Pischke, 2008; Maddala and Lahiri, 1992; Montgomery et al., 2000; Schölkopf et al., 2016).”
• A Causal Inference Method for Reducing Gender Bias in Word Embedding Relations (2019, unknown): “Sch¨olkopf et al. (2016) show that HSR approximates the cleaned variable not less than the mean-subtraction method does.”