Bayesian Networks

Propositional
representation:

nAvoid enumeration of worlds!

nRepresent distribution compactly

Bayesian
networks use locality

Intell_J.Doe

Diffic_CS101

Grade_JDoe_CS101

Intell_FGump

Diffic_Phil101

Grade_FGump_Phil101

Intell_FGump

Diffic_CS101

Grade_FGump_CS101

These “instances” are not independent!

[Pearl]