Mazen Melibari

Learn More
Sum-Product Networks (SPNs), which are probabilistic inference machines, have attracted a lot of interests in recent years. They have a wide range of applications, including but not limited to activity modeling, language modeling and speech modeling. Despite their practical applications and popularity, little research has been done in understanding what is(More)
Sum-Product Networks (SPN) have recently emerged as a new class of tractable probabilistic graphical models. Unlike Bayesian networks and Markov networks where inference may be exponential in the size of the network, inference in SPNs is in time linear in the size of the network. Since SPNs represent distributions over a fixed set of variables only, we(More)
Sum-Product Networks (SPN) have recently emerged as a new class of tractable probabilistic models. Unlike Bayesian networks and Markov networks where inference may be exponential in the size of the network, inference in SPNs is in time linear in the size of the network. Since SPNs represent distributions over a fixed set of variables only, we propose(More)
Inference in dynamic graphical models is known to be hard, except for models with low treewidth structure. This restricts severely the expressive power of these kinds of models. In this document we are proposing a new type of dynamic graphical model that allows one to model complex stochastic processes with unbounded treewidth while guaranteeing tractable(More)
We use 1 : N to abbreviate the notation {1, 2, . . . , N}. We use a capital letter X to denote a random variable and a bolded capital letter X1:N to denote a set of random variables X1:N = {X1, . . . , XN}. Similarly, a lowercase letter x is used to denote a value taken by X and a bolded lowercase letter x1:N denotes a joint value taken by the corresponding(More)
Investigations into probabilistic graphical models for decision making have predominantly centered on influence diagrams (IDs) and decision circuits (DCs) for representation and computation of decision rules that maximize expected utility. Since IDs are typically handcrafted and DCs are compiled from IDs, in this paper we propose an approach to learn the(More)
  • 1