Textual Case-based Reasoning (TCBR) is a powerful paradigm within CBR. Biologically inspired design – the invention of technological systems by analogy to biological systems – presents an opportunity for exploring cross-domain TCBR. Our in situ studies of the retrieval task in biologically inspired design identified findability and recognizability of biology articles on the Web relevant to a design problem as major challenges. To address these challenges, we have developed a technique for semantic tagging of biology articles based on Structure-Behavior-Function models of the biological systems described in the article. We have also implemented the technique in an interactive system called Biologue. Controlled experiments with Biologue indicate improvements in both findability and recognizability of useful biology articles. Our work suggests that task-specific but domain-general model-based tagging might be useful for TCBR in support of complex reasoning tasks engaging cross-domain analogies.
Posted in .