Categorizing biological information based on function-morphology for bioinspired conceptual design Academic Article uri icon

abstract

  • AbstractA function-based keyword search is a concept generation methodology studied in the bioinspired design area that conveys textual biological inspiration for engineering design. Current keyword search methods are inefficient primarily due to the knowledge gap between engineering and biology domains. To improve current keyword search methods, we propose an algorithm that extracts and organizes morphology-based solutions from biological text. WordNet is utilized to discover morphological solutions in biological text. The novel algorithm also adapts latent semantic analysis and the expectationmaximization algorithm to categorize morphological solutions and group biological text. We introduce a novel penalty function that reflects the distance between functions (problems) and morphologies (solutions). The penalty function allows the algorithm to extract morphological solutions directly related to a design problem. We compare the output of the algorithm to manually extracted solutions for validation. A case study is included to exemplify the utility of the developed algorithm. Upon implementation of the algorithm, engineering designers can discover innovative solutions in biological text in a straightforward, efficient manner.

published proceedings

  • AI EDAM-ARTIFICIAL INTELLIGENCE FOR ENGINEERING DESIGN ANALYSIS AND MANUFACTURING

author list (cited authors)

  • Lee, S., Mcadams, D. A., & Morris, E.

citation count

  • 3

complete list of authors

  • Lee, Sooyeon||Mcadams, Daniel A||Morris, Elissa

publication date

  • August 2017