Academic team formation as evolving hypergraphs

Abstract

International audienceThis paper quantitatively explores the social and socio-semantic patterns of constitution of academic collaboration teams. To this end, we broadly underline two critical features of social networks of knowledge-based collaboration: first, they essentially consist of group-level interactions which call for team-centered approaches. Formally, this induces the use of hypergraphs and n-adic interactions, rather than traditional dyadic frameworks of interaction such as graphs, binding only pairs of agents. Second, we advocate the joint consideration of structural and semantic features, as collaborations are allegedly constrained by both of them. Considering these provisions, we propose a framework which principally enables us to empirically test a series of hypotheses related to academic team formation patterns. In particular, we exhibit and characterize the influence of an implicit group structure driving recurrent team formation processes. On the whole, innovative production does not appear to be correlated with more original teams, while a polarization appears between groups composed of experts only or non-experts only, altogether corresponding to collectives with a high rate of repeated interactions

A few special cases: scientific creativity and network dynamics in the field of rare diseases

“…The probability of new co-authorships is the same as the probability of combinatorial creativity. It should be noted that, unlike the case illustrated in Taramasco et al (2010), this relationship holds in probability for a set of related papers and not for a single paper.…”

“…Apart from a few notable exceptions which reveal the emergence of novelty at various levels of social interaction, few papers try to combine complex phenomena like creativity along more than one dimension, by looking in particular at both the scientific community and the related semantic space (Orsenigo et al 2001;Taramasco et al 2010;Roth and Cointet 2010). This paper aims to define a link between the dynamics of creative thought and the social dynamics of the community.…”

“…A paper is a hyperlink in the co-authorship network and, at the same time, a hyperlink in the semantic network of concept co-occurrences (Taramasco et al 2010). More in general, it defines a set of relationships in the socio-semantic network of concepts and authors.…”

Proposing Ties in a Dense Hypergraph of Academics

“…teams) in academic social networks [36], by predicting the formation of links via new and old hyper-edges. Because academic teams tend to evince relationships in different and systematic ways, the hypergraph is a particularly robust datastructure for modeling team formation [40], recommending new collaborations [43] and as a model of how scientists actually seek out new projects [37]. In this paper, instead of assuming a set of edge-types, we show that a distributional measure of link similarity can help leverage the abundance of edge-types found in a text-mined hypergraph of NEs.…”

“…Although some work has explored predicting hyper-edges that connect more than two nodes [21,40], in our model, though it uses hyper-edges to define link similarity, it generates dyadic proposals more analogous to links in a unipartite graph. To compare our proposals to other link prediction methods, we simplify G into a unipartite graph, G, by treating links (sets of edges in G) as unique, unlabeled edges.…”