Cause Effect Pairs in Machine Learning

1. Auflage 2019
ISBN: 978-3-030-21809-6
Verlag: Springer International Publishing

Buch, Englisch, 372 Seiten, Format (B × H): 160 mm x 241 mm, Gewicht: 746 g

Reihe: The Springer Series on Challenges in Machine Learning

This book presents ground-breaking advances in the domain of causal structure learning. The problem of distinguishing cause from effect (“Does altitude cause a change in atmospheric pressure, or vice versa?”) is here cast as a binary classification problem, to be tackled by machine learning algorithms.  Based on the results of the ChaLearn Cause-Effect Pairs Challenge, this book reveals that the joint distribution of two variables can be scrutinized by machine learning algorithms to reveal the possible existence of a “causal mechanism”, in the sense that the values of one variable may have been generated from the values of the other.  
This book provides both tutorial material on the state-of-the-art on cause-effect pairs and exposes the reader to more advanced material, with a collection of selected papers. Supplemental material includes videos, slides, and code which can be found on the workshop website.

Discovering causal relationships from observational data will become increasingly important in data science with the increasing amount of available data, as a means of detecting potential triggers in epidemiology, social sciences, economy, biology, medicine, and other sciences.

Zielgruppe

Graduate

Autoren/Hrsg.

Guyon, Isabelle

Batu, Berna Bakir

Statnikov, Alexander

Fachgebiete

• Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Maschinelles Lernen

Inhaltsverzeichnis

1. The cause-effect problem: motivation, ideas, and popular misconceptions.- 2. Evaluation methods of cause-effect pairs.- 3. Learning Bivariate Functional Causal Models.- 4. Discriminant Learning Machines.- 5. Cause-Effect Pairs in Time Series with a Focus on Econometrics.- 6. Beyond cause-effect pairs.- 7. Results of the Cause-Effect Pair Challenge.- 8. Non-linear Causal Inference using Gaussianity Measures.- 9. From Dependency to Causality: A Machine Learning Approach.- 10. Pattern-based Causal Feature Extraction.- 11. Training Gradient Boosting Machines using Curve-fitting and Information-theoretic Features for Causal Direction Detection.- 12. Conditional distribution variability measures for causality detection.- 13. Feature importance in causal inference for numerical and categorical variables.- 14. Markov Blanket Ranking using Kernel-based Conditional Dependence Measures.