Design representation as semantic networks

Design representation is a common task in the design process to facilitate learning, analysis, redesign, communication, and other activities. Traditional techniques rely on human expertize manual construction are difficult repeat scale. Here, we present methodology that utilizes readily available large-scale multidisciplinary knowledge base (KB) automatically generate as semantic network, i.e., network of entities relations, based descriptions textual form. The requires no ad hoc statistics, but KB. Thus, KB has an essential impact usefulness effectiveness methodology. Based participatory study, observe differences representations generated with alternative KBs. Specifically, trained engineering-related data, TechNet, provides more sensible engineering than commonsense KBs, WordNet ConceptNet, participants who engineers. We further discuss implications findings future research directions enhance networks.