Automatic property visualization for material survey support

Engineers in the semiconductor domain are spending significant time searching for materials with certain properties, such as carrier mobility and bandgap values within certain ranges, or to keep informed about newly discovered materials with desirable properties. Since these searches are not anchored in keywords, the usual information retrieval tools are of limited utility. In this paper, we propose a novel system that enables engineers to more efficiently conduct bibliographic surveys in the domain of semiconductor device materials. The system uses large-scale statistical learning and inference to extract relations between materials and their physical properties from text, and visualizes them.