Retrieval Augmented Generation using Engineering Design Knowledge

TL;DR

This paper introduces a method to extract explicit engineering design facts from patent descriptions, creating a large knowledge base that enhances retrieval-augmented generation for technical design tasks.

Contribution

It develops a dataset and fine-tunes language models for relation extraction, enabling large-scale knowledge base creation to improve design-related language model responses.

Findings

Token classification achieves 99.7% accuracy

Knowledge base contains over 2.93 million facts

Enhanced LLM responses with explicit engineering knowledge

Abstract

Aiming to support Retrieval Augmented Generation (RAG) in the design process, we present a method to identify explicit, engineering design facts - {head entity :: relationship :: tail entity} from patented artefact descriptions. Given a sentence with a pair of entities (based on noun phrases) marked in a unique manner, our method extracts the relationship that is explicitly communicated in the sentence. For this task, we create a dataset of 375,084 examples and fine-tune language models for relation identification (token classification) and elicitation (sequence-to-sequence). The token classification approach achieves up to 99.7 % accuracy. Upon applying the method to a domain of 4,870 fan system patents, we populate a knowledge base of over 2.93 million facts. Using this knowledge base, we demonstrate how Large Language Models (LLMs) are guided by explicit facts to synthesise knowledge and generate technical and cohesive responses when sought out for knowledge retrieval tasks in the design process.