Michael Verba (UNU-MERIT), “Modeling Knowledge Flow on the Global Innovation Network Reveals “Keystone Technologies””
In this paper we model technological knowledge as a dynamic network in order to identify technologies that are the drivers of innovation at the global level, which we term “keystone technologies.” The sphere of technologically relevant knowledge is conceptualized as a reflexive, directed, link- and node-weighted complex network, with distinct spheres of knowledge (or technology domains) representing network nodes and learning (or knowledge flows) across domains acting as inter-nodal links. The empirical knowledge network is constructed from a sweeping database, including records from 105 patent-granting authorities, and containing almost all patents granted anywhere in the World during the period of coverage. We instantiate the nodes of the knowledge network from patent categories of the International Patent Classification (IPC) system. Links between technology domains, representing knowledge transfer between fields of technology, are constructed from patent citations provided by inventors, aggregated at the patent subclass level. Modeling the evolving dynamics of knowledge flow on the World knowledge network allows us to identify the technologies that make the greatest contribution to the dynamics of technological progress and reveals trends in technology over the 22-year period spanning 1991-2012.