In this paper, we examine how patterns of scientific collaboration contribute to knowledge creation. Recent studies have shown that scientists can benefit from their position within collaborative networks by being able to receive more information of better quality in a timely fashion, and by presiding over communication between collaborators. Here we focus on the tendency of scientists to cluster into tightly-knit communities, and discuss the implications of this tendency for scientific performance. We begin by reviewing a new method for finding communities, and we then assess its benefits in terms of computation time and accuracy. While communities often serve as a taxonomic scheme to map knowledge domains, they also affect how successfully scientists engage in the creation of new knowledge. By drawing on the longstanding debate on the relative benefits of social cohesion and brokerage, we discuss the conditions that facilitate collaborations among scientists within or across communities. We show that successful scientific production occurs within communities when scientists have cohesive collaborations with others from the same knowledge domain, and across communities when scientists intermediate among otherwise disconnected collaborators from different knowledge domains. We also discuss the implications of communities for information diffusion, and show how traditional epidemiological approaches need to be refined to take knowledge heterogeneity into account and preserve the system's ability to promote creative processes of novel recombinations of ideas
Diffusion-driven pattern formation in an opinion dynamical network model
Helmholtz Institute for Functional Marine Biodiversity (HIFMB)（赫尔姆霍兹功能海洋生物多样性研究所）；Carl-von-Ossietzky University, Institute for Chemistry and Biology of the Marine Environment（卡尔·冯·奥西茨基大学，海洋环境化学与生物学研究所）；Alfred-Wegener Institute (AWI), Helmholtz Center for Polar and Marine Research（阿尔弗雷德·魏格纳研究所（AWI），赫尔姆霍兹极地海洋研究中心）
A method for assessing the success and failure of community-level interventions in the presence of network diffusion, social reinforcement, and related social effects
