{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

# lec5 - C260A Lecture 5 Probabilistic Modeling Christopher...

This preview shows pages 1–8. Sign up to view the full content.

C260A Lecture 5: Probabilistic Modeling Christopher Lee October 7, 2009

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Defining Events vs. Variables event : a subset of our total probability space S . p ( e ) = a number. variable : some slicing of S into disjoint subsets, each labeled with a distinct value . Now p ( X ) = f ( X ) 1
Multiple Variables = Multiple Slicings Each variable just represents another way of slicing up S . Different slicings could be very similar, or very different. Each time we draw one item from S , it is “labeled” with a value for each of our different variables. 2

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Independence? Are events A , B statistically independent? 3
Independence is About Variables Independence is a statement about the function p ( X , Y ) = p ( X ) p ( Y ) over the entire space, i.e. true x , y ! Don’t confuse independence with simple set intersection. 4

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Event Independence? Create variables X : { A , ¬ A } , Y : { B , ¬ B } . p ( A , B ) = p ( A ) p ( B ) ⇐⇒ p ( X , Y ) = p ( X ) p ( Y ) 5
Unconditional Sampling “Throw a dart at the dartboard”, get one data point ( x , y , z ) yielding a value for each of the variables defined in our information graph. Each time we sample, we get a new value for every vari- able, including our hidden variables. By contrast, in in- ference we typically draw multiple samples of the observ- ables from one specific inference problem (i.e. one value of θ ).

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}