Learning Bayesian Networks (original) (raw)
Published on 2007-08-1267593 Views
Bayesian networks are graphical structures for representing the probabilistic relationships among a large number of variables and doing probabilistic inference with those variables. The 1990's saw the
Related categories
Presentation
Statistical Causality 00:03
slide 201:23
A common way to learn (perhaps define) causation is via manipulation experiments (non-passive data)02:46
slide 404:30
slide 505:03
slide 605:33
slide 706:20
slide 807:22
Causal Graphs08:05
The Causal Markov Assumption10:36
Examples12:41
slide 1218:35
slide 1320:14
slide 14 24:18
Experimental evidence for the Causal Markov Assumption29:24
Exceptions to the Causal Markov Assumption29:42
1. Hidden common causes29:45
2. Causal feedback31:21
3. Selection bias31:44
4. The entities in the population are units of time34:22
slide 2135:24
slide 2235:41
Causal Faithfulness Assumption39:31
Exceptions to the Causal Faithfulness Assumption40:43
Learning Causal Influences Under the Causal Faithfulness Assumption41:13
slide 2641:33
Example42:32
slide 2843:18
Example 143:43
Example 245:18
Example 346:18
Example 451:48
Theorem53:29
Example 554:14
How much data do we need?57:39
Empirical Results59:17
Conflicting Empirical Results01:00:35