Xiaochi Zhou, Patrick Bulter, Changxuan Yang, Simon D. Rihm, Thitikarn Angkanaporn, Jethro Akroyd, Sebastian Mosbach, Markus Kraft
This paper presents a method to integrate formal domain knowledge into large language models by compiling ontological specifications into tools that enforce semantic constraints during knowledge generation.
The research introduces a novel approach to improve the accuracy and reliability of large language models (LLMs) by integrating them with formal domain knowledge. This is achieved by converting ontological specifications into tools that LLM agents use to ensure the knowledge they generate adheres to predefined semantic rules. This method, demonstrated using literature on metal-organic polyhedra synthesis, allows LLMs to interact with structured knowledge more effectively and reduces the need for manual adjustments. Ultimately, this approach aims to make LLMs more capable of generating precise and contextually appropriate information.
