25th Annual Conference on Neural Information Processing Systems (NIPS), Granada 2011

From kernels to causal inference

author: Bernhard Schölkopf, Max Planck Institute for Biological Cybernetics, Max Planck Institute
published: Jan. 25, 2012, recorded: December 2011, views: 640

Slides

Slides
0:00	From Kernels to Causal Inference
0:55	Dependence vs. Causation
2:05	Statistical Implications of Causality
3:17	Functional Causal Model
4:38	Functional Model, ctd.
5:59	Causal Inference from Observational Data
7:54	Restricting the Functional Model
9:47	Causal Inference with Additive Noise, 2-Variable Case
11:08	Identifiability Result
12:37	Causal Inference Method
13:08	Experiments - 1
13:30	Experiments - 2
13:47	Experiments - 3
14:42	Independence-based Regression
15:36	Kernel Independence Testing
17:54	The Kernel Trick
18:04	Detection of Confounders
19:30	Inferring deterministic causal relations
21:28	Causal independence implies anticausal dependence - 1
22:46	Causal independence implies anticausal dependence - 2
22:48	Causal independence implies anticausal dependence - 1
22:51	Causal independence implies anticausal dependence - 2
24:17	80 Cause-Effect Pairs
24:35	80 Cause-Effect Pairs − Examples - 1
24:56	80 Cause-Effect Pairs − Examples - 2
26:54	Causal Learning and Anticausal Learning
29:20	Covariate Shift and Semi-Supervised Learning
33:32	Lens error correction - 1
34:08	Lens error correction - 2
34:47	Lens error correction - 3
34:50	Lens error correction - 2
34:54	Lens error correction - 3
35:09	Inverting a nontrivial convolution model - 1
35:15	Inverting a nontrivial convolution model - 2
35:23	Blurred image
35:26	Our approach
35:28	Blurred image
35:28	Our approach
35:30	Raw Sequence
35:49	Shift-Invariant Kernel Mean Maps
37:01	Causal Inference for Individual Objects
38:00	Causal Markov Conditions - 1
38:27	Causal Markov Conditions - 2
38:49	Kolmogorov complexity
39:07	Conditional Kolmogorov complexity
39:20	Algorithmic mutual information
39:40	Conditional algorithmic mutual information
39:54	Algorithmic mutual information: example
39:54	Postulate: Local Algorithmic Markov Condition
40:10	Equivalence of Algorithmic Markov Conditions
40:41	Algorithmic model of causality
41:41	Applications
41:48	Acknowledgements

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.

Lecture popularity: You need to login to cast your vote.

Description

Kernel methods in machine learning have expanded from tricks to construct nonlinear algorithms to general tools to assay higher order statistics and properties of distributions. They find applications also in causal inference, an intriguing field that examines causal structures by testing their probabilistic footprints. However, the links between causal inference and modern machine learning go beyond this and the talk will outline some initial thoughts how problems like covariate shift adaptation and semi-supervised learning can benefit from the causal methodology.