Papers and slides
Reading group
A few of the papers below are being discussed at our
reading group. This is a opportunity to ask questions to the experts.
Reviews
A few cause-effect pair algorithms
- Shimizu S, Hoyer PO, Hyvarinen A, Kerminen A: A Linear Non-Gaussian Acyclic Model for Causal Discovery. The Journal of Machine Learning Research 2006, 7:2003-2030. [See below LINGAM software]
- Zhang K, Hyvärinen A: Distinguishing causes from effects using nonlinear acyclic causal models. Journal of Machine Learning Research, Workshop and Conference Proceedings (NIPS 2008 causality workshop) 2008, 6:157-164.
- Hoyer PO, Janzing D, Mooij J, Peters J, Scholkopf B: Nonlinear causal discovery with additive noise models. Advances in Neural Information Processing Systems 2009, 21:689-696. [See below ANM software]
- Peters J, Janzing D, Scholkopf B: Causal Inference on Discrete Data using Additive Noise Models IEEE TPAMI vol. 33 no. 12 (2011) 2436-2450. [See below ANMD software]
- Daniusis P, Janzing D, Mooij J, Zscheischler J, Steudel B, Zhang K, Schölkopf B: Inferring deterministic causal relations. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence (UAI-2010) 2010, 143-150.
- Mooij JM, Stegle O, Janzing D, Zhang K, Schölkopf B: Probabilistic latent variable models for distinguishing between cause and effect. Advances in Neural Information Processing Systems 2010, (23):1. [See below GPI software]
- Janzing D, Mooij J, Zhang K, Lemeire J, Zscheischler J, Daniusis P, Steudel B, Scholkopf B: Information-geometric approach to inferring causal directions. Artificial Intelligence 2012, in press. [See below IGCI software]
- Zhang K, Hyvarinen A: On the Identifiability of the Post-Nonlinear Causal Model UAI 2009. [See below PNL software]
See also the
cause-effect pair webpage at the Max Planck Institute.
Software
We provide sample code that interfaces to
number of algorithms. The source code is in the directory: Code/mfunc/causa. The algorithms include:
- LINGAM (Linear Non-Gaussian Acyclic Model) by Patrik O. Hoyer, Shohei Shimizu and Antti Kerminen
- ANM (Additive Noise Model) by Joris Mooij
- ANMD (Additive Noise Model for Discrete data) by Jonas Peters, Dominik Janzing, Bernhard Schoelkopf
- GPI (Gaussian Process Inference) by Oliver Stegle, Joris Mooij
- IGCI (Information Geometric Causal Inference) by Povilas Daniusis, Joris Mooij
- PNL (Post Non-Linear causal modeling) by Kun Zhang
Further help
Help for may be obtained by writing to the
Kaggle website forum or emailing the organizers to
causality@chalearn.org. See also the
Frequently Asked Questions.