Difference between revisions of "10-601 GM2"
From Cohen Courses
Jump to navigationJump to search| Line 17: | Line 17: | ||
* what if some variables are latent? | * what if some variables are latent? | ||
* what is the difference between a Markov network and a correlation network? | * what is the difference between a Markov network and a correlation network? | ||
| − | * | + | * learning a Markov network using Graphical Lasso, which assumptions we are making on the model underlying the data? |
| − | * | + | * why neighborhood selection using lasso is correct? |
Revision as of 22:27, 3 November 2013
Slides
Readings
Taking home message
- what is inference and learning in GM
- inference via belief propagation, when such a method is exact? When it is not?
- what is BP, what are the running modes of BP
- why MLE on fully observed GM is easy?
- what if some variables are latent?
- what is the difference between a Markov network and a correlation network?
- learning a Markov network using Graphical Lasso, which assumptions we are making on the model underlying the data?
- why neighborhood selection using lasso is correct?
