Seminar in applied mathematics and statistics

SPEAKER: Mohamad Elmasri (McGill University)

TITLE: Sub-clustering in decomposable graphs and size-varying junction trees

ABSTRACT: Decomposable graphs are known for their tedious and complicated Markov update steps. Instead of modeling them directly, this work introduces a novel representation of decomposable graphs based on a class of tree-dependent bipartite graphs that span the projective space of the former. The novel representation has few main benefits. First it allows for a new node-driven Markov chain Monte Carlo sampler of decomposable graphs that can easily parallelize and scale. The proposed sampler also benefits from the computational efficiency of junction-tree-based samplers. Second, it enables a form of sub-clustering within maximal cliques of the graph, adding informational richness to the general use of decomposable graphs that could be harnessed in applications with behavioral type of data. Moreover, a likelihood factorization theorem in terms of the novel representation is possible.

Tea and chocolate will be served in room 04.4.19 after the seminar.

------

Upcoming seminars:

January 26, 13:15: Gherardo Varando (Universidad Politécnica de Madrid)

February 2, 13:15: Axel Munk (Göttingen)

February 9, 13:15: Peter Harremoës (Copenhagen Business College)

April 6, 13:15: Kayvan Sadeghi (Cambridge)