Recognizing the Structure of Biological Designs in Text Documents

Biologically inspired design entails analogical transfer of design knowledge from natural systems to technological systems. The goal of the Intelligent Biologically Inspired Design (IBID) project is to help engineers build a precise understanding of a biological system described in a text document and thereby acquire biological cases needed for biologically inspired design. In this paper, we describe a novel technique for recognizing the structural elements of a biological system in a text document. IBID's structure recognition technique allows the structural elements in a biological process to be mapped to domain-independent equivalent concepts useful for cross-domain analogical transfer. In a pilot study, we compare the performance of IBID's structure recognition technique with that of human subjects. Our findings suggest that IBID may be able to help non-experts in biology in recognizing structure.