Proposal for a BioTRIZ-Based Idea Support System Incorporating LLM

Abstract

Translate Article

To support Japan’s Nature Positive goal by 2030, biomimetics plays a key role in eco-friendly technology development. The BioTRIZ data warehouse aids in resolving technical contradictions using biological functions, but requires expert knowledge. We propose a novel BioTRIZ system architecture incorporating a large language model (LLM) to allow non-experts to input problems in natural language. The LLM interprets the input, extracts technical ideas and contradictions, and interacts with the BioTRIZ database. It also outputs ELSI (ethical, legal, and social issues) considerations. This enhances accessibility and ethical awareness in biomimetic innovation.