Free-text inspiration search for systematic bio-inspiration support of engineering design

Abstract Current supportive bio-inspired design methods focus on handcrafting the inspiration engineers use to speed up design. However promising, such are not scalable as time investment is shifted an up-front investment. Furthermore, most proposed require engineer adopt a new process. The current study presents FISh, search method based standard engineering By leveraging machine translation between representative corpus of biological and texts, can start using terminology, which, behind scenes, automatically converted query. This conversion done language models trained patents publications for biology domains. Both aligned used English words. query retrieve documents that describe relevant functionality presented allows searching bio-inspiration free-text updating underlying datasets, organism aspects automated, allowing system stay date without requiring interactive effort. Finally, validated by comparing results existing designs with their inspiring organisms.