Query complex graph patterns

Abstract

In his world-widely renowned book [6], Nobel laureate Herbert Simon pointed out that it is more the complexity of the environment, than the complexity of the individual persons, that determines the complex behavior of humans. The emergence of online social network sites and web 2.0 applications provides a new connected environment/context, where people generate, share and search massive human knowledge; and interact and collaborate with each other to collectively perform some complex tasks.In this talk, we focus on how to make sense of the collaboration data in the context of graphs/networks. To be specific, we will introduce a suite of tools for querying complex patterns from such graphs. Exemplar questions we aim to answer include (a) what makes a team more successful than others, and how to find the best replacement if one of its team members becomes unavailable? [2,4]; (b) how to find a group of authors from databases, data mining and bioinformatics and they collaborate with each other in a starshape? [8,5]; (c) given a set of querying authors of interest, how to find somebody who initiates the research field these querying authors belong to, and how to summarize and visualize the querying authors? [3, 7, 1]; (d) how to incorporate users' preference into these complex queries [9, 10]. We will also introduce the computational challenges behind these querying tools and how to remedy them.