Graphical models form a successful probabilistic modeling approach: They encode relationships among a set of random variables and provide a representation for the joint probability distribution over these variables. The advantages of the graphical formalism are its origins in probability theory and graph theory, the structural modularity favoring parallel computations, and its visual appeal. In this paper, we discuss a method for constructing a particular instance of graphical models (the Helmholtz machine) by using an evolutionary approach. Particularly, we focus on the explaining away phenomenon difficult to address but potentially improving a graphical model qualitatively. Additionally, we provide initial simulation results for a case study.
CITATION STYLE
Garionis, R. (2002). Synthesizing graphical models employing explaining away. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2439, pp. 749–758). Springer Verlag. https://doi.org/10.1007/3-540-45712-7_72
Mendeley helps you to discover research relevant for your work.