Causal modeling & machine learning
Causality is a fundamental notion in science, and plays an important role in explanation, prediction, decision making, and control. It has attracted much attention in philosophy, economics, computer science, and statistics.
In the last decade, interesting advances were made in machine learning for tackling some long-standing problems in causality, such as how to distinguish cause from effect for two random variables, and how to infer the effect of interventions using observational data. For instance, modern machine learning methodologies provided efficient methods for causal structure learning and powerful tools for conditional independence test, which is a key component in traditional constraint-based causal discovery.
On the other hand, causal models provide compact descriptions of the properties of data distributions, and it has recently been demonstrated that causal information can facilitate various machine learning tasks, including semi-supervised learning and domain adaptation (or transfer learning). For instance, causal analysis inspired efficient methods to characterize the information transfer across regimes, environments, and sampling schemes, which is frequently encountered in data analysis.
This workshop aims to foster the research at the intersection of causal modeling and machine learning, and will take place in Beijing, China, on June 25, 2014. In particular, we are interested in how machine learning could help develop better methods for causal discovery and inference, and moreover, how causal knowledge could help understand learning problems and inspire more natural ways to solve them.
Topics of interest
The workshop is concerned with all topics at the intersection of causality and machine learning, including but not limited to
- causal structure learning and inference
- characterization of causal information in observational data
- understanding machine learning tasks in light of causality
- machine learning methods exploiting causal knowledge
- efficient causal discovery in large-scale data
- big data and causality
- real-world problems for causal analysis.
We encourage both full paper submissions (up to 9 papges including references) and short submissions (extended abstracts, up to 4 pages in length). Submissions should be in the ICML 2014 format and need not be anonymous. All accepted contributions will be available online at the workshop website, and will be selected for an oral or poster presentation. We will also include presentations on recently published high-impact work; to give such presentations, please submit a one-page abstract together with published papers for review.
Please click here to submit your contribution by 11 April, 2014.
Other causality-related events
- The cause-effect pairs challenge on the Codalab platform has been relaunched (deadline: June 15, 2014)
- JMLR Special Topic on Causality: Large-scale Experiment Design and Inference of Causal Mechanisms (submission deadline: September 15, 2014)